Explorez tous les épisodes du podcast Machine Learning Street Talk (MLST)

A lot of people in AI treat evolution as a dumb fallback, basically random search for when you can't take a gradient. Akarsh Kumar thinks that is wrong. Selection hangs on to partial solutions, so mutations only need to be useful about 1% of the time for the search to keep making progress.Akarsh is a PhD student at MIT working with Phillip Isola, works with Sakana AI, and is first author of the Fractured Entangled Representation paper with Kenneth Stanley, Jeff Clune and Joel Lehman. He tells Tim Scarfe why the path a learner takes may shape the structure of what it learns, and why that is a different way of looking at intelligence from the statistical one.Most of the conversation is about ASAL, the method he led for searching whole spaces of artificial worlds. Rather than predict what a rule will do, ASAL runs the simulation and asks a foundation model what happened. Mapped across all 262,144 Life-like rules, the most interesting worlds sit on one small island. Along the way: Game of Life, Lenia (with a clip from its creator, Bert Chan), neural cellular automata, Boids, Particle Life and the emergence of persistence.---TIMESTAMPS:00:00:00 Intro: artificial life, ASAL and Core War in four minutes00:04:19 Life as it could be, not just as it is00:05:46 Why the order you learn things in matters00:08:33 No shortcuts: Wolfram, novelty search and regularisation00:10:44 Kenneth Stanley: the path matters, not just the destination00:11:31 Statistical intelligence vs regularity-based intelligence00:13:49 Artificial chemistry and other possible universes00:15:59 Convergent patterns: shadows of the substrate?00:18:02 How Conway's Game of Life works00:20:49 Change one cell, change everything?00:22:31 Lenia (with Bert Chan) and neural cellular automata00:25:26 Boids, Particle Life and cell-like creatures00:28:47 Persistence, entropy and what life is00:30:15 ASAL: a foundation model as the critic00:34:04 Impostors inside the simulation00:36:50 From primordial soup to alien animals00:38:26 262,144 rules and the island of open-endedness00:40:17 From artificial life to AGI00:41:35 Core War: Game of Thrones, Turing edition00:43:22 LLMs as the mutation step: evolving warriors00:46:03 Evolution is anything but random---REFERENCES:paper:[00:04:19] ASAL (Kumar et al.)https://arxiv.org/abs/2412.17799[00:08:08] Assembly theory https://www.nature.com/articles/s41586-023-06600-9[00:09:54] FEP paperhttps://arxiv.org/abs/2505.11581[00:22:38] Lenia: Biology of Artificial Life (Bert Chan)https://arxiv.org/abs/1812.05433[00:23:20] Growing Neural Cellular Automata (Mordvintsev et al.)https://distill.pub/2020/growing-ca/[00:41:35] Digital Red Queen: Core War with LLMs (Kumar et al.)https://arxiv.org/abs/2601.03335[00:43:42] MAP-Elites https://arxiv.org/abs/1504.04909[00:44:55] AlphaEvolve https://arxiv.org/abs/2506.13131[00:47:03] AutoML-Zero (Real et al.)https://arxiv.org/abs/2003.03384book:[00:07:04] Why Greatness Cannot Be Planned (Stanley and Lehman)https://link.springer.com/book/10.1007/978-3-319-15524-1tool:[00:18:12] Conway's Game of Lifehttps://en.wikipedia.org/wiki/Conway%27s_Game_of_Life[00:25:27] Boids (Craig Reynolds)https://www.red3d.com/cwr/boids/[00:27:12] Particle Life (Tom Mohr)https://github.com/tom-mohr/particle-life[00:41:50] Core Warhttps://corewar.co.uk/other:[00:08:49] Computational irreducibility (Stephen Wolfram)https://www.wolframscience.com/nks/p737--computational-irreducibility/[00:10:46] MLST: Why Every AI Model Is an Impostor (Kenneth Stanley, FER documentary)https://www.youtube.com/watch?v=o1q6Hhz0MAg[00:28:30] MLST: Blaise Agüera y Arcas on life emerging from codehttps://www.youtube.com/watch?v=rMSEqJ_4EBk---LINKS:Akarsh Kumar: https://akarshkumar.com/ASAL project page and demos: https://pub.sakana.ai/asal/Digital Red Queen project page: https://sakana.ai/drq/RESCRIPT:https://app.rescript.info/public/share/Qfv3T0EVzqOXL_CYYeRr9Blv8HnNm4JsXBh7ByFjJbc

Tsung-Hsien (Shawn) Wen, CTO of PolyAI, tells Tim Scarfe why voice agents are harder than text agents. Voice adds time, and a good conversation depends on adapting to the person on the line, not just on reasoning to the best answer. Shawn describes an audio-native model (Dialog-RSN-1) that first predicts a turn-taking signal, then replies in text with citations, and writes the transcript last so enterprises can audit it.Along the way: training on real, noisy calls with synthetic noise added, and why over-cleaned audio made the new model worse. Latency, and what a voice agent should do while it thinks. Why a voice with a hint of regional accent beats a generic one. Why public benchmarks fall short for voice, why enterprises want to own their agent harness, and whether behaviour belongs in the harness or in the weights.The last stretch is about working with agents: cognitive debt, the shift from producing content to checking it, Wispr Flow, building tools that agents can use, and whether slop is in the eye of the reader.This episode was produced in partnership with PolyAI.https://poly.ai CHAPTERS0:00 Why voice agents are harder than text4:27 What enterprises want, and why PolyAI built its own model9:04 How an audio-native voice model works15:21 Training data, spectrograms and synthetic noise20:36 The cocktail party problem and the future of turn-taking25:21 Latency, adaptive reasoning and keeping callers' trust31:33 Voices, personality and the uncanny valley36:32 How do you benchmark a voice agent?42:00 Harness engineering and owning the intelligence45:09 Well-specified problems and auditable agents49:46 Weight adaptation and cognitive debt56:31 Agents at work: Wispr Flow, voice and tool building1:02:50 The next decade of voice, and what counts as slopREFERENCESThe Bitter Lesson: http://www.incompleteideas.net/IncIdeas/BitterLesson.html [9:05]Retrieval-augmented generation: https://arxiv.org/abs/2005.11401 [13:23]Mel scale: https://en.wikipedia.org/wiki/Mel_scale [17:16]Victor Zue: https://en.wikipedia.org/wiki/Victor_Zue [17:45]Cocktail party effect: https://en.wikipedia.org/wiki/Cocktail_party_effect [20:39]Speaker diarisation: https://en.wikipedia.org/wiki/Speaker_diarisation [21:16]

Leonardo de Moura created Lean and co-created Z3. ---This episode is sponsored by Parallel.Parallel, where agents find answers: web search, extraction and deep research APIs built for AI agents.Start free with the Parallel MCP server and $5 of credits every month: https://parallel.ai/mlst?utm_source=creator&utm_medium=podcast&utm_content=MLST---Tim Scarfe talks with Leo about how Lean escaped its original audience, why dependent types and Mathlib made it useful to working mathematicians, and what happens when formal verification leaves the lab. De Moura explains the small trusted kernel and independent checkers, and gives his account of the recent Collatz incident, in which a purported proof was accepted by both Lean's official kernel and Nanoda, apparently by exploiting a different bug in each.---TIMESTAMPS:00:00:00 Cold open: the green checkmark can lie00:00:59 Cathedral or bazaar: who controls Lean's core?00:04:55 Why Lean's core stays small and protected00:08:07 The Slack purge, Brandolini's law and the Lean FRO00:11:12 The Collatz exploit: two kernels, two bugs00:16:44 More kernels, reward hacking and safety by transparency00:21:04 Sponsor: Parallel00:21:59 Kim Morrison, Claude and the zlib proof00:25:25 Can we specify complex systems?00:28:04 Specs change: proofs are cheaper to redo with AI00:31:27 From Lean 1 to Lean 400:34:57 Dependent types in plain terms00:37:25 Lean 4's extensibility and Mathlib's growth00:41:44 Mathlib as infrastructure: Formal Frontiers00:44:15 Creativity, abstraction and nut-sniping00:48:46 Breadcrumbs, not learning: what AI agents lack00:53:03 Competence without comprehension, and verified guardrails00:56:21 Is the human still the author?01:01:44 AlphaProof, LLMs and why certificates still matter01:06:38 What's next for Lean, and its legacy01:11:42 How to start learning Lean---REFERENCES:tool:[00:00:48] Leanhttps://lean-lang.org/[00:01:24] Mathlibhttps://github.com/leanprover-community/mathlib4[00:12:09] nanoda_libhttps://github.com/ammkrn/nanoda_lib[00:12:19] CollatzLeanhttps://github.com/xrchz/CollatzLean/blob/a79357462a33d2a6babd4cf6c8d8bcd25425d653/README.md[00:13:06] Lean issue 14576https://github.com/leanprover/lean4/issues/14576[00:13:21] Lean pull request 14577https://github.com/leanprover/lean4/pull/14577[00:13:45] nanoda_lib pull request 22https://github.com/ammkrn/nanoda_lib/pull/22[00:15:33] Lean comparatorhttps://github.com/leanprover/comparator[00:18:02] Lean4Leanhttps://github.com/digama0/lean4lean[00:19:50] ARC-AGI-3https://arcprize.org/arc-agi/3[00:21:59] lean-ziphttps://github.com/kim-em/lean-zip[00:29:45] CompCerthttps://compcert.org/[00:29:45] seL4https://www.sel4.org/[00:30:23] Z3https://github.com/Z3Prover/z3[00:38:10] Veilhttps://github.com/verse-lab/veil[00:38:10] Velvethttps://github.com/verse-lab/velvetorganization:[00:00:52] Lean FROhttps://lean-lang.org/fro/[00:42:39] Mathlib Initiativehttps://mathlib-initiative.org/about/person:[00:05:11] Ilya Sergeyhttps://ilyasergey.net/[00:10:31] Joachim Breitnerhttps://www.joachim-breitner.de/[00:32:58] Adam Chlipalahttps://adam.chlipala.net/[00:38:42] Kevin Buzzardhttps://www.ma.imperial.ac.uk/~buzzard/[01:10:16] Terence Taohttps://terrytao.wordpress.com/book:[00:08:19] The Proof in the Codehttps://us.macmillan.com/books/9780374620059/theproofinthecode/other:[00:10:05] Brandolini's lawhttps://en.wikipedia.org/wiki/Brandolini%27s_law[00:45:33] A new result on unit distanceshttps://openai.com/index/model-disproves-discrete-geometry-conjecture/[01:01:08] Fermat's Last Theorem formalisationhttps://imperialcollegelondon.github.io/FLT/paper:[00:34:36] The Lean 4 theorem prover and programming languagehttps://doi.org/10.1007/978-3-030-79876-5_37---RESCRIPT:https://app.rescript.info/share/7d3d4a0059443236a01f6c9acbf4db58https://app.rescript.info/api/public/sessions/b007264c0ce89047/pdf

Weco let an AI coding agent rewrite the harness around another agent for eight days: its code, prompts and tools, while the underlying language model stayed fixed. Tim Scarfe asks Weco co-founder Zhengyao Jiang what the reported gains over two years of human engineering actually demonstrate.The discussion examines AIDE 85's generated code, held-out evaluation and the difficulty of separating useful discoveries from reward hacking. Jiang explains Weco's four levels of recursive self-improvement and compares the experiment with AlphaEvolve and the Darwin Gödel Machine.The limits matter as much as the gains. Jiang explains why the experiment did not establish that the system had become a better improver. The conversation closes with open-ended search, human-designed primitives and Parameter Golf: where does the next useful idea come from when the agent is searching inside a space that people designed?---TIMESTAMPS:00:00:00 Eight days of self-improvement: what counts?00:03:25 AIDE and the puzzle of useful spaghetti code00:08:38 Four levels of recursive self-improvement00:12:02 What AIDE 85 changed and how it was tested00:20:04 AlphaEvolve, Darwin Gödel Machine and the RSI claim00:26:21 Reward hacking and the limits of detection00:33:09 Open-ended search, harness tuning and creativity00:39:43 Parameter Golf and the limits of self-improvement---REFERENCES:organization:[00:00:30] Weco AIhttps://www.weco.ai/other:[00:00:33] AIDE²: The First Evidence of Recursive Self-Improvementhttps://www.weco.ai/blog/first-evidence-of-recursive-self-improvement[00:14:11] Faulty reward functions in the wildhttps://openai.com/index/faulty-reward-functions/[00:29:59] The Hugging Face incident and the road aheadhttps://openai.com/index/hugging-face-incident-and-the-road-ahead/tool:[00:03:29] AIDEhttps://github.com/WecoAI/aideml[00:04:29] MLE-benchhttps://github.com/openai/mle-bench[00:04:33] ALE-Benchhttps://github.com/SakanaAI/ALE-Bench[00:04:52] WeatherBench 2https://github.com/google-research/weatherbench2[00:08:18] ReActhttps://react-lm.github.io/[00:39:43] Parameter Golfhttps://github.com/openai/parameter-golfpaper:[00:20:08] AlphaEvolve: A coding agent for scientific and algorithmic discoveryhttps://arxiv.org/abs/2506.13131v1[00:21:35] Darwin Gödel Machine: Open-Ended Evolution of Self-Improving Agentshttps://arxiv.org/abs/2505.22954v3[00:23:45] Hyperagentshttps://arxiv.org/abs/2603.19461v1[00:27:01] SpecBench: Measuring Reward Hacking in Long-Horizon Coding Agentshttps://arxiv.org/abs/2605.21384book:[00:33:14] Why Greatness Cannot Be Planned: The Myth of the Objectivehttps://link.springer.com/book/10.1007/978-3-319-15524-1---LINKS:https://app.rescript.info/share/3a9dc6189cb539c6a05fcc4f75c101b3PDF:https://app.rescript.info/api/public/sessions/9eda60ede2b31c92/pdf

Frank Hutter, co-founder of Prior Labs, talks about TabPFN, a tabular foundation model that makes predictions in a single forward pass, and the research behind it.

TabPFN is pre-trained on synthetic datasets drawn from a prior over structural causal models, rather than on real data. At prediction time it takes the whole training table as context and outputs an approximation of the Bayesian posterior predictive distribution, without per-dataset training or hyperparameter search. Frank explains how this grew out of his earlier work on AutoML and neural architecture search, how the priors are built and revised, and why tabular data was hard for deep learning for so long.

The conversation also covers the TabArena benchmark, how the architecture changed from TabPFN v1 to v3, scaling to larger tables, using the model with coding agents, test-time compute, Google's TabFM, causal inference and interventions, and relational data. At the end, a short update Frank recorded after the interview covers the TabPFN-3.5 release.

Prior Labs:

TabPFN-3.5: https://priorlabs.ai/tabpfn-3-5

https://priorlabs.ai/careers

TOC:

00:00 Introduction

00:44 Welcome and Frank's background

02:05 Why tabular data was hard for deep learning

10:17 Pre-training on synthetic data

12:52 The TabArena benchmark

19:28 From AutoML to neural architecture search

26:34 TabPFN as a learned algorithm

30:50 Bayesian prediction in one forward pass

39:37 Scaling to larger tables

47:48 Using TabPFN with coding agents

57:47 Output heads and architecture from v1 to v3

1:05:29 Test-time compute and adaptation

1:13:32 Google's TabFM

1:16:53 How the priors are designed

1:18:40 Correlation, causation and interventions

1:35:22 Relational and multimodal data

1:38:31 Use in organisations

1:46:38 The open research arm

1:50:21 Update: TabPFN-3.5

REFS:

TabPFN v2, Nature (Hollmann et al., 2025)

https://www.nature.com/articles/s41586-024-08328-6

Transformers Can Do Bayesian Inference (Müller et al.)

https://arxiv.org/abs/2112.10510

TabArena (Erickson et al.)

https://arxiv.org/abs/2506.16791

AutoGluon-Tabular (Erickson et al.)

https://arxiv.org/abs/2003.06505

Beyond IID: How General Are Tabular Foundation Models, Really?

https://arxiv.org/abs/2606.30410

Neural Architecture Search: A Survey (Elsken, Metzen & Hutter)

https://arxiv.org/abs/1808.05377

Auto-WEKA (Thornton et al.)

https://www.cs.ubc.ca/~hutter/papers/AutoWEKA-KDD2013.pdf

TabPFN v1 (Hollmann et al., 2022)

https://arxiv.org/abs/2207.01848

TabPFN-3 technical report

https://arxiv.org/abs/2605.13986

TabPFN-2.5 report

https://arxiv.org/abs/2511.08667

CAAFE (Hollmann et al.)

https://arxiv.org/abs/2305.03403

TabICL (Qu et al.)

https://arxiv.org/abs/2502.05564

TabICLv2 (Qu et al.)

https://arxiv.org/abs/2602.11139

Google TabFM

https://research.google/blog/introducing-tabfm-a-zero-shot-foundation-model-for-tabular-data/

TALENT benchmark (Ye et al.)

https://arxiv.org/abs/2407.00956

Do-PFN (Robertson et al.)

https://arxiv.org/abs/2506.06039

CausalPFN (Balazadeh et al.)

https://arxiv.org/abs/2506.07918

Causal Foundation Models with Partial Graphs (Reuter et al.)

https://arxiv.org/abs/2602.14972

RelBench (Robinson et al.)

https://arxiv.org/abs/2407.20060

RelArena-α, TabPFN-Rel and RPI

https://arxiv.org/abs/2608.16319

TabPFN on GitHub

https://github.com/PriorLabs/TabPFN

TabPFN-3.5 technical report

https://arxiv.org/abs/2609.17895

Otto Group Product Classification Challenge (Kaggle, 2015)

https://www.kaggle.com/competitions/otto-group-product-classification-challenge

---RESCRIPT:https://app.rescript.info/share/e99676c25ee6189fbf54c9be07eb623e

Alexander Mattick is a researcher at Fraunhofer IIS and a PhD researcher at the University of Technology Nuremberg (UTN), and a regular on Yannic Kilcher's Discord. He first came on MLST in 2022, after helping research the Yann LeCun and Randall Balestriero episode on interpolation.

SPONSOR:

---

Cyber Fund built the Monastery to help founders ship products that were impossible a year ago. Applications for Batch 1 are now open.

Apply now: https://cyber.fund

---

Alexander treats inference as the thread running through modern machine learning: once you have a model, what does it cost to get an answer out of it? He works through Monte Carlo, GFlowNets, energy-based models, diffusion, normalising flows and flow matching, with four short explainers he recorded himself. He is blunt about energy-based models: you can sample from them in principle, but it is rarely worth the compute. JEPA and "world model", he says, are closer to branding than to technical categories.

Next: theories of deep learning, none of which he thinks predicts enough yet to guide practice, then reinforcement learning.

---

0:00 Cold open: information is expensive

0:51 Welcome back, Alexander Mattic

2:08 Alexander's research background

2:50 Inference: densities, sampling and Monte Carlo

6:42 GFlowNets, energy functions and MCMC

9:45 Explainer: energy-based models

11:03 Why model a density at all?

17:30 From learned energies to flow matching

25:08 Explainers: diffusion and normalising flows

28:33 Are energy-based models generative?

33:22 JEPA, contrastive learning and collapse

41:13 Why non-language modalities need flows

44:51 Inference as search: branch and bound

49:43 Q-learning and delayed consequences

55:14 Flow matching, optimal transport, Fokker-Planck

1:00:03 Explainer: flow matching

1:01:49 AlphaFold, latents and scale versus architecture

1:07:52 Two families of deep learning theory

1:15:04 What a good theory would predict

1:23:53 The manifold hypothesis and compression

1:28:25 Is reward enough?

1:32:01 Control theory versus reinforcement learning

1:37:22 The Bitter Lesson and expensive information

1:42:08 Constrained RL: the constrained MDP toolbox

1:50:12 Creativity as constrained search

1:55:44 Reality is protean: when abstractions hold

2:00:32 What is a world model?

2:04:38 Prediction is not control

2:08:13 Robot demos, MPC and reliability

---

REFERENCES:

[6:55] GFlowNets (Bengio et al., 2021)

https://arxiv.org/abs/2106.04399

[38:46] Contrastive Self-Supervised Learning (Anand, 2020)

https://ankeshanand.com/blog/2020/01/26/contrative-self-supervised-learning.html

[38:56] LeJEPA (Balestriero and LeCun, 2025)

https://arxiv.org/abs/2511.08544v3

[47:10] RL for Node Selection in Branch-and-Bound (Mattick)

https://openreview.net/forum?id=0ez68a5UqI

[56:20] Flow Matching for Generative Modeling

https://arxiv.org/abs/2210.02747v2

[1:12:41] Disentangling feature and lazy training in deep neural networks

https://arxiv.org/abs/1906.08034v4

[1:31:05] Reward is enough (Silver)

https://doi.org/10.1016/j.artint.2021.103535

[1:35:12] Learning ReLU networks to high uniform accuracy is intractable (Berner et al.)

https://arxiv.org/abs/2205.13531v2

[1:40:20] Dota 2 with Large Scale Deep RL

https://arxiv.org/abs/1912.06680v1

[1:45:41] Constrained Update Projection for Safe Policy Optimization (Yang et al., 2022)

https://arxiv.org/abs/2209.07089

[1:46:11] SafeMPO (ICLR 2026)

https://openreview.net/forum?id=1m0EU6QXj6

[1:50:17] Why Creativity Cannot Be Interpolated

https://archive.mlst.ai/paper/why-creativity-cannot-be-interpolated/

[1:51:39] Invalid Action Masking (Huang and Ontañón)

https://arxiv.org/abs/2006.14171

[2:00:04] Probability Theory: The Logic of Science (Jaynes, 2003)

https://www.cambridge.org/core/books/probability-theory/9CA08E224FF30123304E6D8935CF1A99

[2:01:53] Training Agents Inside of Scalable World Models (Hafner et al., 2025)

https://arxiv.org/abs/2509.24527v1

[2:03:43] World Models (Ha and Schmidhuber, 2018)

https://arxiv.org/abs/1803.10122v4

The car making a left turn at the start of this episode was never filmed. Cosmos 3 generated it. Ming-Yu Liu, who leads the Cosmos research at NVIDIA, explains how one model can describe a video, generate one, and produce robot actions.

He walks Tim through the architecture. A vision language model reasons one token at a time; its weights then initialise a bidirectional diffusion generator for video, audio and action, and a shared temporal position scheme lines up signals that run at different rates. Ming-Yu treats "world model" as a set of tools, not one definition: forward dynamics, inverse dynamics and policy, trained together under a capacity limit so that each helps the others. He also explains why plentiful first-person human video carries over to robots, which have far less data of their own, and why a Cosmos model post-trained on the DROID dataset is a good starting point for pick-and-place policies.

The most practical thread is testing. A neural simulator does not need accurate success rates. It only needs to rank policy A above policy B the way the real world would, so a team can narrow down which checkpoints deserve a real trial. Cosmos Dreams applies that closed-loop idea to driving and robotics, and Ming-Yu argues that humanoids around children and pets make safety matter even more than it does for cars. The conversation ends on the Super, Nano and Edge sizes (Edge targets Jetson Thor, Orin and DGX Spark) and where to find the open weights, code and data.

This episode is a paid partnership with NVIDIA.

Learn more about Cosmos: https://nvda.ws/4cJoY1S

Explore Cosmos Lab: https://research.nvidia.com/labs/cosmos-lab/cosmos3/

---

TIMESTAMPS:

00:00:00 A road that was never filmed

00:02:28 Inside Cosmos 3: reasoning and generator towers

00:05:02 World models: dynamics, policy and one clock

00:08:59 Learning robot skills from human video

00:11:06 Ambiguous tasks and system 2 planning

00:12:53 Neural simulators for policy verification

00:16:41 Cosmos as a starting point for robot policies

00:19:00 Cosmos Dreams and robot safety

00:22:04 Super, Nano and Edge model sizes

00:24:24 Open models, the Cosmos repo and feedback

---

REFERENCES:

tool:

[00:00:13] Cosmos 3 (NVIDIA Cosmos Lab project page)

https://research.nvidia.com/labs/cosmos-lab/cosmos3/

[00:18:27] NVIDIA Cosmos GitHub repository

https://github.com/NVIDIA/cosmos

[00:22:05] Cosmos3-Edge model card

https://huggingface.co/nvidia/Cosmos3-Edge

[00:22:15] Cosmos3-Super model card

https://huggingface.co/nvidia/Cosmos3-Super

[00:22:16] Cosmos3-Nano model card

https://huggingface.co/nvidia/Cosmos3-Nano

[00:22:50] NVIDIA Jetson Thor

https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-thor/

[00:22:52] NVIDIA Jetson Orin

https://www.nvidia.com/en-us/autonomous-machines/embedded-systems/jetson-orin/

[00:22:53] NVIDIA DGX Spark

https://www.nvidia.com/en-us/products/workstations/dgx-spark/

[00:24:42] Cosmos 3 collection on Hugging Face

https://huggingface.co/collections/nvidia/cosmos3

other:

[00:01:07] Cosmos-Dreams closed-loop simulators (NVIDIA SIGGRAPH 2026 blog)

https://blogs.nvidia.com/blog/siggraph-news-2026/

paper:

[00:08:54] Cosmos 3: Omnimodal World Models for Physical AI

https://arxiv.org/abs/2606.02800

[00:17:43] DROID: A Large-Scale In-The-Wild Robot Manipulation Dataset

https://arxiv.org/abs/2403.12945

---

RESCRIPT: https://app.rescript.info/share/e2385948cf465f0d6a2c0930150fc3ab

Pavan Kumar Reddy leads audio research at Mistral AI. He joins Tim Scarfe for a deep technical tour of Voxtral — and explains why the frontier of deployed voice is still a cascade of specialised models rather than one end-to-end system.

IN PARTNERSHIP WITH MISTRAL AI:

---

This episode was produced in partnership with Mistral AI.

Mistral AI: https://mistral.ai/

---

The conversation opens on architecture. Voxtral Chat feeds a 3B Ministral text trunk with continuous embeddings from an audio encoder, passed to the decoder as direct token input rather than through cross-attention as in Whisper, so the model can answer questions about emotion, timing and who spoke when without an intermediate transcript to lose them. The real-time model becomes a dual-stream decoder that consumes audio and emits text at once, at a target delay down to 160ms, with slower streams in parallel for anything that can wait for more context.

On generation, Pavan explains why Voxtral TTS predicts continuous latents rather than discrete codec tokens, traces the lineage from SoundStream through EnCodec to Mimi's split of semantic and acoustic codebooks, and places FSQ and flow matching in it. Tim presses on the priors underneath: why a mel spectrogram instead of raw waveform, what noise augmentation buys, and when acoustic overfitting becomes somebody's fine-tuning problem. Then the failure modes. Diarisation is emitted autoregressively inside the transcript rather than by a separate head, which makes streaming diarisation fragile — less context, late speaker changes, invented extra speakers. And because the architecture commits to what it has already predicted, one out-of-distribution mistake compounds into looping or skipped segments, which is what DPO corrects: the negative supervision pre-training and SFT cannot give.

The last third is the argument Tim keeps returning to. Customers running voice agents over millions of sessions describe scaffolding, not a solved problem, with a sharp drop outside the top few languages. Cascades survive because each component stays separately adaptable, observable and constrainable. And voice alone is cognitive debt: absorbing information and deciding in one serial stream is harder than glancing at a menu. Voice becomes ubiquitous beside a screen, not instead of one.

---

TIMESTAMPS:

00:00:00 Cold open

00:00:46 Why Mistral moved into audio

00:09:27 Inside Voxtral: trunk, encoder, dual streams

00:20:22 Speech that works in real time

00:30:52 How a voice becomes tokens

00:39:59 Flow matching, FSQ and the new codec

00:52:51 When speech models lose the speaker

01:03:23 Correcting hallucinations with preferences

01:12:12 Controlling synthetic speech

01:20:06 Why cascades still win

01:29:25 Speech in the wild

01:33:46 Audio models as interfaces

01:37:54 Why voice still needs a screen

---

REFERENCES:

paper:

[00:01:42] Mistral 7B

https://arxiv.org/abs/2310.06825

[00:09:38] Voxtral

https://arxiv.org/abs/2507.13264

[00:14:41] Whisper: Robust Speech Recognition

https://arxiv.org/abs/2212.04356

[00:19:11] Voxtral Realtime

https://arxiv.org/abs/2602.11298

[00:21:52] Delayed Streams Modeling (Kyutai)

https://arxiv.org/abs/2509.08753

[00:30:52] Voxtral TTS

https://arxiv.org/abs/2603.25551

[00:32:38] SoundStream neural audio codec

https://arxiv.org/abs/2107.03312

[00:34:59] Flow Matching for Generative Modeling

https://arxiv.org/abs/2210.02747

[00:37:03] EnCodec: High Fidelity Neural Audio Compression

https://arxiv.org/abs/2210.13438

[00:37:42] Moshi and the Mimi codec

https://arxiv.org/abs/2410.00037

[00:39:05] Finite Scalar Quantization (FSQ)

https://arxiv.org/abs/2309.15505

[01:03:33] Direct Preference Optimization (DPO)

https://arxiv.org/abs/2305.18290

dataset:

[00:46:14] Mozilla Common Voice

https://commonvoice.mozilla.org/en/datasets

organization:

[00:50:47] Hugging Face

https://huggingface.co/

Can a machine learn the judgement that separates a plausible-looking result from a faithful experiment? Edward Hughes, Chief Scientist and co-founder of Inherent, joins Tim Scarfe to argue that creativity is not optimisation, and that the missing capability in AI is choosing which questions are worth asking.

SPONSOR:

---

Cyber Fund built the Monastery to help founders ship products that were impossible a year ago.

Apply now: https://cyber.fund

---

Edward makes the case that Move 37 was innovative rather than creative, and that the field, not the individual, decides what counts as a discovery. That reframing runs through Csikszentmihalyi, Deutsch and exaptation into open-endedness, where deceptive goals and imperfect world models turn out to be the point rather than the problem. The second half turns to the paper: Replica, a task space built by redacting figures from real papers, and Faraday, a 27-billion-parameter model trained to steer a frontier coding agent that then beats the frontier on held-out replications.

---

TIMESTAMPS:

00:00:00 Cold open: Move 37, Faraday and collective intelligence

00:01:08 Sponsor: CyberFund

00:01:46 Inherent's $50M raise and the road from string theory

00:09:14 Three timescales of learning: weights, context, culture

00:13:47 Move 37 was innovative, not creative: the field decides

00:20:39 Creativity as satisficing: the urinal and evolution

00:25:06 Exaptation and the Tristan chord: creativity in context

00:30:56 Coherence for whom? Deutsch's hard-to-vary explanations

00:35:53 Why copying is creative: Deutsch and the constraint engineer

00:42:27 Societies of agents and the strong Moravec paradox

00:45:51 Evaluate in hindsight: from Lean proofs to climate change

00:51:56 Picbreeder, local goals and why discovery needs deception

00:57:21 Spaghetti proofs, translation layers and superhuman Go

01:00:37 Does nature compress? Naturalness and real patterns

01:07:36 Why replicate? Replica's redacted figures and Faraday

01:12:31 Faraday beats Codex, Claude and GLM 5.2 on held-out tasks

01:15:31 Replication to innovation: how the Transformer happened

01:18:26 Deep replication: what Faraday learns from Voyager and GNoME

01:23:37 Can the AI scientist cheat? Goodharting the judge

01:29:09 Inside Replica: scale-down, 8xB300 runs, per-task rubrics

01:34:11 The RL crisis: getting GRPO to work with per-turn credit

01:39:43 Weights vs harnesses: AlphaEvolve, DGM and EvoTune

01:45:45 The recursive company: agents cross a phase transition

01:50:35 Collective intelligence and the electric dynamo

01:55:46 What replaces OKRs? Incumbents and the burden of knowledge

---

REFERENCES:

MLST Creativity Article:

https://archive.mlst.ai/read/why-creativity-cannot-be-interpolated

organization:

[00:01:47] Inherent

https://inherentlabs.ai/

other:

[00:20:51] Marcel Duchamp, Fountain

https://www.tate.org.uk/art/artworks/duchamp-fountain-t07573

[00:05:19] Human-Timescale Adaptation in an Open-Ended Task Space (Adaptive Agent)

https://arxiv.org/abs/2301.07608

[00:06:05] The AI Scientist

https://arxiv.org/abs/2408.06292

[00:12:13] Training AI Scientists to Replicate Research (Replica and Faraday)

https://arxiv.org/abs/2608.13331

[01:44:46] Evolutionary Principles in Self-Referential Learning

https://people.idsia.ch/~juergen/diploma.html

[01:59:33] Are Ideas Getting Harder to Find?

https://www.nber.org/papers/w23782

book:

[00:16:04] Creativity: Flow

https://search.worldcat.org/title/254487436

[00:26:22] Why Greatness Cannot Be Planned

https://link.springer.com/book/10.1007/978-3-319-15524-1

[00:33:03] The Beginning of Infinity

https://www.penguinrandomhouse.com/books/293575/the-beginning-of-infinity-by-david-deutsch/

[01:55:47] Laws of Knowledge

https://www.penguin.co.nz/books/the-infinite-alphabet-9780241655672

(Full list refs on YT/rescript)

---

RESCRIPT:

https://app.rescript.info/session/670296ba913761d0?share=6281911cac9bdbff637f10819d4d1e5c

Could slowing AI development make superintelligence safer? Daniel Kokotajlo and Thomas Larsen of the AI Futures Project join Tim Scarfe to examine AI 2040: Plan A, a proposal to buy time before AI exceeds human control.

SPONSOR:

---

Cyber Fund built the Monastery to help founders ship products that were impossible a year ago.

Apply now: https://cyber.fund

---

After revisiting AI 2027 and the limits of forecasting, they ask what happens when AI can automate research and sustain an economy without human workers. Tim challenges the case for general models and asks whether intelligence alone explains power. Plan A proposes an initial pause to build safety infrastructure, then cautious development up to the strongest AI that can still be reliably controlled. The discussion tests the distinction between control and alignment, the case for public AI research, and whether the US and China could enforce a slowdown. It ends with the evidence that would change their forecasts.

---

TIMESTAMPS:

00:00:00 AI 2040: a slower route to superintelligence

00:01:34 Sponsor: Cyber Fund

00:02:12 From OpenAI to AI 2027

00:06:58 Forecasts, war games and self-fulfilling prophecies

00:17:44 Why AI sceptics are changing their minds

00:23:04 When AI can replace its own researchers

00:28:45 Could an AI economy grow without human workers?

00:37:32 One general model or a society of specialists?

00:47:43 Brains, machines and collective intelligence

00:56:12 Plan A: buy time at the controllable frontier

01:00:02 Why control buys time but cannot replace alignment

01:06:36 Why AI research should be public

01:10:32 Can the US and China enforce an AI slowdown?

01:19:04 Why AI policy debates miss the technology

01:21:56 Is AI normal technology? The remaining disagreement

Many thanks to James Wilken-Smith for helping with show research.

---

REFERENCES:

other:

[00:00:01] AI 2040: Plan A

https://ai-2040.com/

[00:03:27] AI 2027

https://ai-2027.com/

[00:13:47] Scenario Scrutiny for AI Policy

https://blog.aifutures.org/p/scenario-scrutiny-for-ai-policy

[00:33:11] The 2028 Global Intelligence Crisis

https://www.citriniresearch.com/p/2028gic

[01:00:40] Brief independent investigation of agents' behavior, reasoning and collaboration in the OpenAI / Hugging Face hacking incident

https://www.redwoodresearch.org/research/hugging-face-incident

[01:09:21] The Hugging Face incident and the road ahead

https://openai.com/index/hugging-face-incident-and-the-road-ahead/

[01:22:01] AI as Normal Technology

https://www.normaltech.ai/p/ai-as-normal-technology

[01:22:51] Common Ground between AI 2027 & AI as Normal Technology

https://asteriskmag.substack.com/p/common-ground-between-ai-2027-and

person:

[00:19:43] Geoffrey Hinton

https://www.cs.toronto.edu/~hinton/

[00:20:07] Ryan Greenblatt

https://www.lesswrong.com/users/ryan_greenblatt

[00:26:06] Elon Musk

https://www.tesla.com/elon-musk

tool:

[00:21:46] ARC-AGI-3

https://arcprize.org/arc-agi/3

[00:21:53] AlphaGo and Move 37

https://deepmind.google/research/alphago/

[00:39:41] Claude

https://claude.com/product/overview

[00:39:58] NVIDIA H100 GPU

https://www.nvidia.com/en-us/data-center/h100/

paper:

[00:24:42] Training AI Scientists to Replicate Research

https://arxiv.org/abs/2608.13331v1

[01:27:19] Validity of the single processor approach to achieving large scale computing capabilities

https://www.cs.cmu.edu/~18742/papers/Amdahl1967.pdf

book:

[00:28:52] Bullshit Jobs: A Theory

https://www.simonandschuster.com/books/Bullshit-Jobs/David-Graeber/9781501143335

organization:

[01:05:09] Redwood Research

https://www.redwoodresearch.org/

---

RESCRIPT:

https://app.rescript.info/public/share/33d1a58fa8f307ae7dfd504d4fdaa9d5

Tom McGrath is co-founder and Chief Scientist at Goodfire, and a former Google DeepMind researcher. He joins Tim Scarfe to ask what neural networks actually learn, whether their internal representations converge on structures in the world, and whether interpretability can extract new scientific knowledge rather than merely explain model outputs.

Beginning with AlphaZero and learned modularity, the conversation moves into neural geometry: concept manifolds, reusable computation inside Llama, and why activation steering can fail when it pushes a model off-manifold. McGrath then makes the case for intentional design, using interpretability as part of the training loop. They examine controlled generalisation, features as rewards, predictive data debugging, and the uncomfortable fact that a model may recognise a hallucination or reward hack and still produce it.

The discussion closes on grader awareness, oversight and collusion between adaptive agents, then returns to sparse autoencoders. SAEs are useful, McGrath argues, but they may fracture the higher-dimensional structures networks actually use. This episode was made with support from Goodfire.

---

TIMESTAMPS:

00:00:00 Introduction: Can interpretability speed-run science?

00:02:03 The invisible grader

00:06:51 What AlphaZero learned from the world

00:12:24 Interpretability as a control loop

00:21:54 The forbidden method and safer interventions

00:37:36 Why models catch hallucinations too late

00:46:19 Debug the dataset before training

00:50:44 Why neural networks become modular

00:55:57 Finding the geometry inside a network

01:02:55 Why steering falls off the manifold

01:12:10 A reusable calculator inside Llama

01:17:19 From abstractions to goals

01:25:28 Reward hacking, oversight and collusion

01:37:23 Are sparse autoencoders dead?

---

REFERENCES:

paper:

[00:05:45] Emergent Misalignment: Narrow finetuning can produce broadly misaligned LLMs

https://arxiv.org/abs/2502.17424v7

[00:11:05] Acquisition of Chess Knowledge in AlphaZero

https://arxiv.org/abs/2111.09259

[00:25:30] Steering Out-of-Distribution Generalization with Concept Ablation Fine-Tuning

https://arxiv.org/abs/2507.16795

[00:29:30] Persona Vectors: Monitoring and Controlling Character Traits in Language Models

https://arxiv.org/abs/2507.21509

[00:41:14] Features as Rewards: Scalable Supervision for Open-Ended Tasks via Interpretability

https://arxiv.org/abs/2602.10067

[00:47:03] Anatomy of Post-Training: Using Interpretability to Characterize Data and Shape the Learning Signal

https://arxiv.org/abs/2606.12360

[01:00:26] Do Sparse Autoencoders Capture Concept Manifolds?

https://arxiv.org/abs/2604.28119

[01:03:04] Manifold Steering Reveals the Shared Geometry of Neural Network Representation and Behavior

https://arxiv.org/abs/2605.05115

[01:14:20] Arithmetic in the Wild: Llama uses Base-10 Addition to Reason About Cyclic Concepts

https://arxiv.org/abs/2605.01148

[01:29:35] Measuring Reward-Seeking via Contrastive Belief Updates

https://arxiv.org/abs/2607.18966v1

other:

[00:15:44] Intentional Design

https://www.goodfire.com/blog/intentional-design

[00:56:12] The World Inside Neural Networks

https://www.goodfire.com/research/the-world-inside-neural-networks

[01:37:28] A Pragmatic Vision for Interpretability

https://www.alignmentforum.org/posts/StENzDcD3kpfGJssR/a-pragmatic-vision-for-interpretability

---

RESCRIPT:

https://app.rescript.info/share/846cfee4131b664fd09209cc3b98018e

Tim Scarfe speaks with Ilia Shumailov and Alexander Panfilov about their paper, Stealing Reasoning Traces from Proprietary LLM APIs.The core bug sounds deceptively simple: providers return encrypted reasoning state so conversations can be resumed or forked. But those blobs can be replayed across users and sibling models. A smaller model can ask the provider to decrypt the trace, then repeat the hidden reasoning in plain text. The discussion covers leaked private data, a broadly reusable jailbreak, poisoned agent traces, chain-of-thought monitoring, responsible disclosure, and possible defenses.Ilia Shumailov is an AI and security researcher, formerly at Google DeepMind, who completed his Cambridge PhD under Ross Anderson. Alexander Panfilov is a PhD researcher at the ELLIS Institute Tübingen and the Max Planck Institute for Intelligent Systems, working on AI safety, adversarial machine learning, and LLM red-teaming. They close by separating the demonstrated jailbreaking threat from ordinary benign distillation, and by arguing for controlled experiments over sweeping claims.---TIMESTAMPS:00:00:00 Intro montage00:01:33 Portable encrypted thought and decoded reasoning00:24:55 How the attack works and what it means00:39:04 Doom, defense, and scientific restraint---REFERENCES:paper:[00:00:00] Stealing Reasoning Traces from Proprietary LLM APIshttps://arxiv.org/abs/2608.09867[00:09:22] Chain of Thought Monitorability: A New and Fragile Opportunity for AI Safetyhttps://arxiv.org/abs/2507.11473[00:11:30] Reasoning Models Don’t Always Say What They Thinkhttps://www.anthropic.com/research/reasoning-models-dont-say-think[00:37:22] PostTrainBench: Can LLM Agents Automate LLM Post-Training?https://arxiv.org/abs/2603.08640[00:41:02] Large-scale online deanonymization with LLMshttps://arxiv.org/abs/2602.16800other:[00:09:28] OpenAI and Hugging Face partner to address security incident during model evaluationhttps://openai.com/index/hugging-face-model-evaluation-security-incident/[00:10:22] Claude, GPT, and Gemini All Struggle to Evade Monitorshttps://metr.org/notes/2025-08-22-claude-gpt-gemini-struggle-evade-monitors/tool:[00:42:08] Isabelle proof assistanthttps://isabelle.in.tum.de/---RESCRIPT: https://app.rescript.info/share/07fc38276e0823dc9b8986c32e202c7f

Astrophysicist Adam Becker, author of "What Is Real?", joins Tim Scarfe to take apart the futures Silicon Valley keeps selling: the 2045 singularity, mind uploading, Mars colonies, and the AI apocalypse. His new book *More Everything Forever* argues these ideas are hugely influential, mostly evidence-free, and bankrolled by tech billionaires who need a story in which growth never ends.Becker does the physics the boosters skip. Kurzweil's "law of accelerating returns" rests on cherry-picked data, and every exponential ends. Grant Bezos his perpetual energy growth and humanity boils the oceans within a few centuries, then exhausts the observable universe in under 4,000 years. The stars are too far away, Mars dirt is poison, and the day the dinosaur-killing asteroid hit Earth was still nicer than any day on Mars. On AI, Becker calls LLMs pocket calculators for language: hallucination is the model doing exactly what it always does, and the intelligence explosion assumes intelligence is a single number you can buy with compute.The sting is that Becker thinks the doomers are sincere. Yudkowsky, Bostrom and the effective altruists are not grifters, he says, just wrong, and their warnings that AI could end the world feed the same growth story the money depends on. He closes with his own prescription: take social problems seriously, regulate the whole tech industry, and tax billionaires out of existence.---TIMESTAMPS:00:00:00 Cold open and the thesis of More Everything Forever00:04:24 Kurzweil's singularity and the physical limits of exponential growth00:14:02 High agency and the fantasy of imprinting humanity on the cosmos00:16:55 Mind uploading, functionalism, and embodied cognition00:24:24 AI psychosis and anthropomorphizing LLMs00:26:24 Calculators, hallucination, and the limits of scale00:32:20 Yudkowsky and the intelligence-explosion argument00:40:37 True believers, venture capital, and the sci-fi growth narrative00:47:21 From Extropians to EA: utilitarianism and longtermism00:53:50 Brain worms and Becker's prescription: take social science seriously00:56:49 Why the AI-ethics discourse is broken01:01:42 The eugenics and IQ argument against 'intelligence'01:06:07 Why space settlement fails: Mars, the moon, and orbital data centers01:10:42 Billionaire myths and the search for purpose01:13:38 Tax billionaires, regulate tech: closing prescriptions---REFERENCES:book:[00:00:07] More Everything Forever (Adam Becker, 2025)https://www.hachettebookgroup.com/titles/adam-becker/more-everything-forever/9781541619593/[00:00:15] What Is Real? (Adam Becker, 2018)https://en.wikipedia.org/wiki/What_Is_Real%3F[00:15:46] What We Owe the Future (Will MacAskill, 2022)https://www.hachettebookgroup.com/titles/william-macaskill/what-we-owe-the-future/9781541618626/other:[00:00:27] Dreaming Against the Machine (podcast)https://www.dreamingagainstthemachine.com[00:01:04] The Useful Idiots of AI Doomsaying (Adam Becker, The Atlantic, 2025)https://www.theatlantic.com/books/archive/2025/09/what-ais-doomers-and-utopians-have-in-common/684270/

RESCRIPT: https://app.rescript.info/share/d6e37f9866673d8f74a39076efa5926b

This episode is sponsored by Notion. Learn more about Notion's Developer Platform today at https://notion.com/mlstWhy can deep networks discover abstractions that shallow models miss? Statistical physicist Matthieu Wyart joins Tim Scarfe to argue that the answer lies in the hidden hierarchy of data. Language and images are built from parts within parts; depth lets a network recover those coarse-grained variables and escape the curse of dimensionality.The conversation moves from jamming transitions and rough loss surfaces to Chomsky, context-free grammars and machine creativity. Wyart explains why next-token prediction can still recover compositional structure, where current systems fall short of genuine scientific invention, and why predicting latent representations rather than raw tokens could make learning far more sample-efficient.They also examine diffusion models, neural scaling laws and the limits of physics-inspired theory. The final question is on a personal note: if mistakes are the price of leaving the beaten path, how much scientific risk is worth taking?---TIMESTAMPS:00:00:00 Can machines learn abstractions from data?00:02:00 Notion agentic workspace00:02:49 From statistical physics to machine learning00:06:40 What physics can explain about learning00:16:37 From Carnot to Chomsky bulldozer00:21:21 How deep networks recover hidden hierarchies00:32:43 Where machine creativity still falls short00:40:48 How deep nets escape the curse of dimensionality00:52:19 Why predict latents instead of tokens01:02:49 The sample-efficiency case for latent prediction01:08:31 Diffusion, scaling laws and text entropy01:16:40 The scientists we learn from and the mistakes we make---REFERENCES:person:[00:00:43] Noam Chomskyhttps://linguistics.mit.edu/user/chomsky/tool:[00:02:08] Notion Developer Platformhttps://www.notion.com/en-gb/blog/introducing-developer-platformpaper:[00:04:43] Mastering the game of Go with deep neural networks and tree searchhttps://www.nature.com/articles/nature16961[00:05:52] Reconciling modern machine-learning practice and the bias-variance trade-offhttps://arxiv.org/abs/1812.11118[00:25:54] How Deep Neural Networks Learn Compositional Data: The Random Hierarchy Modelhttps://arxiv.org/abs/2307.02129[00:42:12] Efficient Estimation of Word Representations in Vector Spacehttps://arxiv.org/abs/1301.3781[00:52:46] Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecturehttps://arxiv.org/abs/2301.08243[00:52:54] Learn from your own latents and not from tokens: A sample-complexity theoryhttps://arxiv.org/abs/2605.27734[01:08:31] A Phase Transition in Diffusion Models Reveals the Hierarchical Nature of Datahttps://arxiv.org/abs/2402.16991[01:11:39] Scaling Laws for Neural Language Modelshttps://arxiv.org/abs/2001.08361[01:12:17] Deriving Neural Scaling Laws from the statistics of natural languagehttps://arxiv.org/abs/2602.07488[01:13:34] Prediction and Entropy of Printed Englishhttps://ieeexplore.ieee.org/document/6773263---LINKS:Download PDF transcript: https://app.rescript.info/share/f7644cdaa86c5cc1e41e484e290f2bd4

Can an AI do the right thing for the wrong reason? Tim Scarfe speaks with Apollo Research’s Alexander Meinke, Axel Højmark and Jérémy Scheurer about Measuring Reward-Seeking via Contrastive Belief Updates, their new research with OpenAI.

The panel asks how models infer what graders reward, why good behaviour can come from the wrong reason, and whether that difference can be measured. The conversation moves through promise-breaking, grader awareness, reward hacking, scheming, opaque reasoning and corrigibility, then turns to a detailed walkthrough of the contrastive-belief method and what its results do and do not show. The o3 results discussed here concern an intermediate checkpoint without safety training.

This episode was made in partnership with Apollo Research. MLST retained full editorial control.

Reference

Apollo Research: https://www.apolloresearch.ai/

---

TIMESTAMPS:

00:00:00 Cold Open

00:02:12 Right Things, Wrong Reasons

00:12:47 Grader Awareness

00:26:22 Legibility

00:32:35 What To Call It

00:35:58 Intelligence, Agency, Anthropomorphism

00:45:16 Apollo’s Mission

00:48:54 The End of the Exponential

00:55:45 The Paper

01:16:34 Closing Reflection

---

REFERENCES:

tool:

[00:00:08] Claude Fable

https://www.anthropic.com/claude/fable

[00:12:50] AlphaGo Zero

https://deepmind.google/blog/alphago-zero-starting-from-scratch/

[00:44:30] AlphaFold 3

https://deepmind.google/science/alphafold/

paper:

[00:01:02] Measuring Reward-Seeking via Contrastive Belief Updates

https://arxiv.org/abs/2607.18966

[00:16:19] Natural Language Autoencoders Produce Unsupervised Explanations of LLM Activations

https://transformer-circuits.pub/2026/nla/

[00:26:48] Stress Testing Deliberative Alignment for Anti-Scheming Training

https://arxiv.org/abs/2509.15541

[00:35:33] Shortcut learning in deep neural networks

https://arxiv.org/abs/2004.07780

[00:53:49] Measuring AI Ability to Complete Long Software Tasks

https://arxiv.org/abs/2503.14499

[00:59:52] Modifying LLM Beliefs with Synthetic Document Finetuning

https://alignment.anthropic.com/2025/modifying-beliefs-via-sdf/

[01:10:44] Alignment Faking in Large Language Models

https://arxiv.org/abs/2412.14093

[01:13:55] Natural Emergent Misalignment from Reward Hacking

https://www.anthropic.com/research/emergent-misalignment-reward-hacking

other:

[00:10:14] We Need a Science of Scheming

https://www.apolloresearch.ai/science/science-of-scheming/

[00:32:56] CoastRunners reward hacking example

https://deepmind.google/blog/specification-gaming-the-flip-side-of-ai-ingenuity/

organization:

[01:06:07] Redwood Research

https://www.redwoodresearch.org/

---

ReScript:

https://app.rescript.info/share/718ab68e18cfa3b9b800da6b3290fd42

This episode is sponsored by Notion. Learn more about Notion's Developer Platform today at https://notion.com/mlst

Britain's most capable coding model can't be exported, and that ban is the whole reason Cosine set out to build one from scratch. Alistair Pullen, CEO and co-founder of Cosine, sits down with Tim Scarfe to explain how a frontier system he calls Fable, locked behind US export controls, became the founding case for a UK sovereign model trained on the Isambard supercomputer in Bristol.

The bet underneath it is economic. Pullen argues that an inference company, rather than a training-first lab, doesn't need billions to compete: millions, a national compute allocation, and a consortium feedback loop can be enough. From there it gets into the machinery, why open-weight models still trail the frontier on size, active parameters and data, the mixture-of-experts versus dense trade-off and why active params dominate how a model actually feels, and the edge that real coding trajectories confer.

The back half is about making agents trustworthy. Pullen makes the case for beating "slop" by rewarding the process instead of the final answer, reframes code review as runtime proof (spin the bug up in a VM and force the agent to actually exploit it), and walks through Swarm, Cosine's system running hundreds of sub-agents in one shot. It ends on why memory is still an unsolved hack, how synthetic graders let you run RL on tasks with no built-in test, and why Pullen reads US export controls as an accidental gift, with a supply-chain sting in the tail.

---

TIMESTAMPS:

00:00:00 The sovereign mandate and the Fable ban

00:04:02 Millions vs billions: the inference-company model

00:07:19 The consortium feedback loop

00:07:40 Why open models lag the frontier

00:14:59 MoE vs dense, and why active params matter

00:16:29 Trajectories: the process-data advantage

00:19:48 Beating slop: reward the process, not the answer

00:26:06 Reusable abstractions and the epistemic wall

00:29:56 Code review becomes runtime proof

00:37:32 Do agentic harnesses still matter?

00:40:35 Swarm: orchestrating hundreds of sub-agents

00:45:14 Why memory is still unsolved

00:48:25 Synthetic data and graders for RL

00:53:09 The US export gift and supply-chain risk

---

REFERENCES:

organization:

[00:01:15] Cosine

https://cosine.sh

[00:04:14] Mistral AI

https://mistral.ai

[00:05:50] Anthropic

https://www.anthropic.com

[00:07:42] Cohere

https://cohere.com

[00:08:36] DeepSeek

https://www.deepseek.com

tool:

[00:02:52] Isambard-AI

https://isambard.ac.uk

[00:05:56] Colossus (xAI)

https://en.wikipedia.org/wiki/Colossus_(supercomputer)

[00:07:52] GLM (Z.ai)

https://z.ai

[00:11:52] NVIDIA B300

https://www.nvidia.com/en-us/data-center/dgx-b300/

[00:15:37] gpt-oss-120b

https://huggingface.co/openai/gpt-oss-120b

[00:15:52] Devstral 2

https://mistral.ai/news/devstral

[00:16:01] Llama 70b

https://www.llama.com

[00:17:05] Claude Code

https://www.anthropic.com/claude-code

[00:26:23] ARC-AGI (Francois Chollet)

https://arcprize.org

[00:40:38] Swarm (Cosine)

https://cosine.sh

[00:40:50] OpenAI Codex

https://github.com/openai/codex

[00:41:16] Lumen Outpost (Cosine)

https://cosine.sh

[00:41:18] Kimi K2 (Moonshot)

https://huggingface.co/moonshotai/Kimi-K2-Instruct

[00:49:55] SWE-bench

https://www.swebench.com

[00:52:40] SystemVerilog

https://en.wikipedia.org/wiki/SystemVerilog

person:

[00:23:40] Andrej Karpathy

https://karpathy.ai

paper:

[00:27:10] GRPO (DeepSeekMath)

https://arxiv.org/abs/2402.03300

[00:27:13] GSPO

https://arxiv.org/abs/2507.18071

Incompressible Knowledge Probes, Bojie Li

https://arxiv.org/pdf/2604.24827

Estimating the Size of Claude Opus 4.5/4.6

https://unexcitedneurons.substack.com/p/estimating-the-size-of-claude-opus

---

ReScript:

https://app.rescript.info/session/5852d2b884c4ce4b?share=10b9799160845bb11779f8ac6cd3124f

Tim Scarfe travels to Zurich to sit down with the Tufa Labs ARC-AGI-3 team — founder Benjamin Crouzier, with Jeroen Cottaar, Dries Smit, Stefano Viel and Michal Tesnar — to work out what their leaderboard-topping system does and what the benchmark is really testing.The cut opens on the games: a walkthrough of the Locksmith game, where you read the rules of an unfamiliar world straight from raw frames. ARC-AGI-3 makes ARC interactive and agentic, so the model has to *discover* the goal rather than transduce a static grid. It stays easy for humans and breaks LLMs, and it runs through everything that follows. Dries traces his StochasticGoose preview win — brute force that only searched actions which changed the frame — and why it collapsed once the organisers added action-efficiency scoring and unseen games.Induction and transduction run through the middle of the conversation — how much of an answer is really priors leaking back the moment a model recognises a maze. The abstraction mountain, and Tim's case that LLMs reach the right answer through fractured, entangled representations — performance, not competence. Whether transformers plan at all or just fake it well enough. Why the score really measures action efficiency, not games solved, and why agents lock onto the wrong goal and cannot climb back out.Crouzier closes on the Tufa Labs thesis — a small lab against the giants, the bitter lesson against hand-built harnesses, and safety — and Tim ties it back to Kenneth Stanley, deep constraints, and creativity as competence.

Disclosure: Tufa Labs sponsors MLST. ---TIMESTAMPS:00:00:00 Meet the Tufa team and what makes ARC-AGI-3 hard00:02:11 Locksmith game: reading the rules from raw frames00:03:10 Why build an independent research lab00:04:11 StochasticGoose: a preview win, then the hardened games00:07:58 Induction, transduction, and priors inside LLMs00:10:31 Curiosity, world models, and exploring by frame change00:14:32 Understanding debt and losing sight of your own code00:15:53 Requirements-based agents and human-AI co-creativity00:19:22 Why auto-research misses the big picture00:21:54 The abstraction mountain and fractured representations00:27:36 Constraints and making LLMs act as if they understand00:34:51 Human difficulty calibration, esports priors, and emergence00:41:35 Agency, goal acquisition, and two kinds of planning00:47:31 Harnesses, the 36% number, and wrong-goal loops00:52:33 Rewards, goals, and why ARC-AGI-3 resists brute force01:00:46 Would solving ARC-AGI-3 prove AGI?01:07:53 Stripping language away, then priors leak back01:14:06 Representation and whether language is necessary01:18:04 The bitter lesson versus specialised harnesses01:22:20 Capability research, safety, and the software singularity---REFERENCES:organization:[00:02:11] ARC-AGI-3https://arcprize.org/arc-agi/3[00:03:10] Tufa Labshttps://tufalabs.ai/team/[00:04:20] ARC-AGI-3 Preview Agent Competitionhttps://arcprize.org/competitions/arc-agi-3-preview-agentstool:[00:04:55] StochasticGoose ARC-AGI-3 solutionhttps://github.com/DriesSmit/ARC3-solution[00:07:42] ArcGenticahttps://github.com/symbolica-ai/arcgentica[00:07:49] RGB-Agenthttps://github.com/alexisfox7/RGB-Agent[00:14:38] Claude Codehttps://www.anthropic.com/claude-code[01:03:42] Qwen 3.6 27Bhttps://huggingface.co/Qwen/Qwen3.6-27Bpaper:[00:13:03] On the Measure of Intelligencehttps://arxiv.org/abs/1911.01547[00:27:42] DreamCoderhttps://arxiv.org/abs/2006.08381[00:43:55] On the Biology of a Large Language Modelhttps://transformer-circuits.pub/2025/attribution-graphs/biology.html[01:18:46] ImageNet Classification with Deep CNNs (AlexNet)https://proceedings.neurips.cc/paper_files/paper/2012/file/c399862d3b9d6b76c8436e924a68c45b-Paper.pdfother:[01:18:16] The Bitter Lessonhttp://www.incompleteideas.net/IncIdeas/BitterLesson.html---https://app.rescript.info/share/463d7f031349b4b9db428553eed88230

Thomas Ahle wants Normal Computing to be the Lovable for chip design: type your intent, and a swarm of agents carries it from design through optimisation, formalisation and verification to tape-out. To get there, his team at wrote their own open-source Verilog simulator, 580,000 lines in 43 days, because commercial EDA verifiers run about $10,000 per core and there are no decent open-source compilers to build on.

That sets up the question Tim keeps pressing: if an agent can produce a chip design, a proof, or a working program, how do you actually know it is correct? Passing 70% of tests is not the same as being right, and a single fabricated bug can cost a company a fortune. They dig into ProgramBench (rebuild a program from its tests, roughly 0% success), the difference between structure and competence, and the "understanding debt" you take on when nobody reads the code.

From there: auto-formalisation in Lean and the AlphaProof trick of training on prove-or-disprove; why there is no single true representation of a spec (Petri nets, TLA+, Erik Curiel's "math does not represent"); and thermodynamic computing, where Normal Computing's CN101 chip is built so that its physical noise *is* the computation, settling a stochastic differential equation in hardware to invert a matrix. Plus Bayesian uncertainty, specialisation, the Chomsky hierarchy, AI slop, and whether performance is all that matters.

Recorded in Zurich.

Disclosure: Normal Computing paid our production and travel costs for this show. We retained full editorial control. They did not see the video before publication, and we did not show it to them or discuss it with them beforehand.

---

TIMESTAMPS:

00:00:00 Meet Thomas Ahle: the Lovable for chip design

00:03:41 Why hardware needs formal verification

00:06:36 Ten thousand dollars per core and a six-month agent run

00:07:40 Rebuilding programs from tests: ProgramBench and zero percent

00:12:15 Structure vs competence: can you learn a program from behavior?

00:15:27 Continual learning, abstraction, and Claude as an ecosystem

00:23:17 Autoformalization and the AlphaProof trick

00:29:31 No single true representation: specs, Petri nets and TLA+

00:34:43 Thermodynamic computing: when noise is the computation

00:37:32 Bayesian uncertainty in the age of token streams

00:41:12 Hybrid compute: vibe-coding loops, binaries and Stockfish

00:44:44 Co-design, central-AI apps and API pricing

00:49:45 Chain of thoughtlessness and the Chomsky hierarchy

00:53:40 AI psychosis, slop and the broken social contract

00:57:34 Typing it yourself, teamwork and performance vs competence

---

REFERENCES:

person:

[00:00:10] Thomas Ahle

https://thomasahle.com

organization:

[00:00:27] Normal Computing

https://normalcomputing.com/

paper:

[00:11:21] ProgramBench: Can Language Models Rebuild Programs From Scratch?

https://arxiv.org/abs/2605.03546

[00:31:55] Autoformalizing Memory Device Specifications with Agents

https://arxiv.org/abs/2605.00058

[00:35:20] Thermo AI and the Fluctuation Frontier

https://arxiv.org/abs/2302.06584

[00:36:40] Thermo Comp System for AI Applications

https://arxiv.org/abs/2312.04836

[00:37:05] Thermodynamic Linear Algebra

https://arxiv.org/abs/2308.05660

[00:44:50] An efficient probabilistic hardware architecture for diffusion-like models

https://arxiv.org/abs/2510.23972

tool:

other:

[00:01:00] Building an Open-Source Verilog Simulator with AI: 580K Lines in 43 Days

https://normalcomputing.com/blog/building-an-open-source-verilog-simulator-with-ai-580k-lines-in-43-days

[00:02:55] Normal Computing Announces Tape-Out of the World's First Thermodynamic Computing Chip (CN101)

https://www.normalcomputing.com/blog/normal-computing-announces-tape-out-of-worlds-first-thermodynamic-computing-chip

[00:32:02] DRAMBench: Autoformalizing DRAM Specifications with Timed Petri Nets

https://www.iese.fraunhofer.de/blog/drambench-autoformalizing-dram-specifications/

---

ReScript: https://app.rescript.info/share/ff9684a112ab37744096adaeb097a263

This episode is sponsored by Notion. Learn more about Notion's Developer Platform today at https://notion.com/mlstProtein folding stalled biology for fifty years. A sequence of amino acids dictates a three-dimensional shape, but reading that shape meant a year and roughly $100,000 of crystallography per structure. Then AlphaFold 2 won CASP14 so decisively the organizers called the problem essentially solved.In this documentary cut, John Jumper, who shared the 2024 Nobel Prize in Chemistry and has since left DeepMind for Anthropic, walks Tim Scarfe through what the system did and, more interestingly, what it did not. The architecture gets a proper dissection: MSAs, the Evoformer, invariant point attention, the FAPE loss, and Jumper's correction of the equivariance story, which ablations valued at roughly 2.5 of 30 GDT points rather than the whole win. He is blunt about the limits. AlphaFold predicts one experiment extraordinarily well; it is not a model of the cell, it does not capture dynamics, and on a given drug target it is "wrong nine times out of ten."From there: the AlphaFold Database of 200M+ predicted structures, AlphaFold 3 and ligands, Isomorphic Labs, and Jumper's quarrel with the bitter lesson, where finite data and human hypotheses still matter. Emmanuel Nji of BioStruct Africa closes the film on what changes when work that took years now takes months, and on training the next thousand structural biologists across Africa.---TIMESTAMPS:00:00:00 Cold open: predicting nature with a button press00:01:03 The protein folding bottleneck and CASP00:04:39 The Nobel, the database, and the move to Anthropic00:05:50 Sponsor (Notion) and framing: what AlphaFold does not claim00:07:39 Proteins as self-assembling nanomachines00:12:24 From structures to biology: drug discovery and Midnolin00:17:37 The humility of AlphaFold: a narrow predictor00:22:18 Inside the architecture: Evoformer, IPA and FAPE00:30:20 Ruthless empiricism: ablations and 100x in data00:35:20 Predict, control, understand00:40:00 Against the bitter lesson; AlphaFold 3 as diffusion00:45:07 Intelligence, representations and AGI00:49:23 Epilogue: AlphaFold in Africa00:52:16 Closing: the case for hybrid science models---REFERENCES:organization:[00:01:55] Critical Assessment of Structure Prediction (CASP)https://predictioncenter.org/[00:04:39] The Nobel Prize in Chemistry 2024https://www.nobelprize.org/prizes/chemistry/2024/summary/[00:05:18] BioStruct Africahttps://www.biostructafrica.org/[00:18:03] Isomorphic Labshttps://www.isomorphiclabs.com/paper:[00:03:09] AlphaFold Protein Structure Databasehttps://doi.org/10.1093/nar/gkab1061[00:17:25] Accurate structure prediction of biomolecular interactions with AlphaFold 3https://www.nature.com/articles/s41586-024-07487-w[00:22:18] Highly accurate protein structure prediction with AlphaFoldhttps://www.nature.com/articles/s41586-021-03819-2[00:23:10] Midnolin promotes degradation of substrates independent of ubiquitinationhttps://doi.org/10.1126/science.adh5021[00:27:00] Improved protein structure prediction using potentials from deep learninghttps://www.nature.com/articles/s41586-019-1923-7tool:[00:03:09] AlphaFold Protein Structure Database (EBI)https://alphafold.ebi.ac.uk/[00:45:55] AlphaEvolve: a coding agent for designing advanced algorithmshttps://deepmind.google/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/other:[00:39:40] The Bitter Lessonhttp://www.incompleteideas.net/IncIdeas/BitterLesson.html---ReScript: https://app.rescript.info/share/d8cde5c221fb71e2c0f5aafe94f90dfaDisclaimer - not sponsored, editorial with us - we filmed it at GDM, London

Brad Carson was the Army's General Counsel, served two terms in Congress and was Acting Under Secretary of Defense for Personnel and Readiness. He now heads Americans for Responsible Innovation, the AI-policy advocacy group he co-founded. Keith Duggar spends roughly eighty minutes pushing back.

SPONSOR:

---

Cyber Fund built the Monastery to help founders ship products that were impossible a year ago. Applications for Batch 1 are now open.

Apply now: https://cyber.fund

---

Carson's whole case rests on one line: the genie is not out of the bottle. We have pulled dangerous tech back before. Asilomar halted recombinant DNA in 1975, and the West still controls the chips AI runs on. Calling it unstoppable, he says, is the most dangerous idea in the room.

Then Keith drags him somewhere darker. A Palantir heat map scores you 0.73 on whether you are a combatant, and a strike follows. The model is wrong some accepted share of the time, and when it is, nobody answers for it. You cannot court-martial a model, and not even the interpretability researchers can say why it picked you.

—

Note: after recording, we learned that Americans for Responsible Innovation is backed by EA-aligned philanthropy (not sponsored)

---

TIMESTAMPS:

00:00:00 From the Pentagon to AI governance

00:04:52 Regulatory capture vs Silicon Valley networks

00:07:56 Transparency and the Claude tier changes

00:09:40 Tort liability when AI tools cause harm

00:13:40 AI is a product, not a person

00:16:01 Children, suicide, and the suicide business

00:19:59 Opaque neural nets and the law of war

00:25:54 Probabilistic targeting and the death of accountability

00:28:47 The arms race fallacy: Asilomar and restraint

00:34:02 Talking to China: track 2 talks and chip leverage

00:39:45 Air power never wins: capital for labour

00:43:29 Anthropic vs the Department of War

00:51:29 Concentration, open source, and brain drain

01:00:18 DeepSeek, Chinese culture, and AI as diplomacy

01:12:25 Upskilling Congress and why public trust matters

---

REFERENCES:

organization:

[00:02:45] ICRC position on autonomous weapons

https://www.icrc.org/en/law-and-policy/autonomous-weapons

[00:05:22] Americans for Responsible Innovation (ARI)

https://ari.us

[00:07:20] Andreessen Horowitz (a16z)

https://a16z.com/

[01:16:05] Office of Technology Assessment

https://en.wikipedia.org/wiki/Office_of_Technology_Assessment

other:

[00:03:35] Beneficial AGI 2019 Conference (Future of Life Institute, Puerto Rico)

https://futureoflife.org/event/beneficial-agi-2019/

[00:18:30] Section 230 of the Communications Decency Act

https://en.wikipedia.org/wiki/Section_230

[00:19:59] Lethal Autonomous Weapons (LAWS)

https://en.wikipedia.org/wiki/Lethal_autonomous_weapon

[00:31:35] Strategic Arms Limitation Talks (SALT)

https://en.wikipedia.org/wiki/Strategic_Arms_Limitation_Talks

[00:32:28] Asilomar Conference on Recombinant DNA (1975)

https://en.wikipedia.org/wiki/Asilomar_Conference_on_Recombinant_DNA

[00:39:45] The New Iron Triangle (ARI policy byte)

https://ari.us/policy-bytes/the-new-iron-triangle/

[00:48:05] Defense Production Act

https://en.wikipedia.org/wiki/Defense_Production_Act

person:

[00:03:35] Anthony Aguirre

https://en.wikipedia.org/wiki/Anthony_Aguirre

[00:06:48] Dean Ball — Hyperdimensional

https://www.hyperdimensional.co/

[00:23:13] Neel Nanda — mechanistic interpretability

https://www.neelnanda.io/

[00:36:02] Jack Clark (Anthropic) on Conversations with Tyler

https://conversationswithtyler.com/episodes/jack-clark/

[00:39:15] Robert Trager — Centre for the Governance of AI

https://www.governance.ai/team/robert-trager

[00:41:55] Giulio Douhet

https://en.wikipedia.org/wiki/Giulio_Douhet

[01:15:05] Don Beyer (US Congress)

https://en.wikipedia.org/wiki/Don_Beyer

tool:

[00:22:19] Phalanx CIWS

https://en.wikipedia.org/wiki/Phalanx_CIWS

---

ReScript:

https://app.rescript.info/public/share/9405ff35c0215b7cdae6402d41284171

https://app.rescript.info/api/public/sessions/0a6c081b8e5fe413/pdf

Michael I. Jordan, described by Science magazine as the most influential computer scientist alive, has never thought of himself as an AI researcher. In this conversation he explains why that distinction matters.

SPONSOR:

---

Cyber Fund built the Monastery to help founders ship products that were impossible a year ago. Applications for Batch 1 are now open.

Apply now: https://cyber.fund

---

Jordan trained as a statistician and cognitive scientist, and his career has been spent building machine learning systems that work in the real world: supply chains, commerce, healthcare, and large economic systems. When the field rebranded itself as AI and then AGI, he did not follow. Instead he argues that the framing is wrong. AI is better understood as a collective economic system than as a race to build a disembodied superintelligence.

We talk about why AGI is mostly a PR term, what machine learning achieved before the LLM hype cycle, and why the assistant-on-your-shoulder vision may be less compelling than it sounds. Jordan explains why explanations need to be actionable, not merely mechanistic; why AlphaFold's missing error bars matter; how prediction-powered inference changes the picture; and why drug discovery is an incentive-design problem rather than a pure pattern-matching problem.

ERRATA: Science magazine ranked him the most influential computer scientist, not Nature

---

TIMESTAMPS:

00:00:00 Cold open: A demoralizing message to young builders

00:02:04 CyberFund sponsor read

00:02:50 From symbolic AI to machine learning systems

00:05:42 Why AGI is mostly a PR term

00:08:48 A collectivist, economic perspective on AI

00:11:33 Why LLMs need system design, not hype

00:14:50 Predictability beats faux understanding

00:17:55 AlphaFold, bias, and prediction-powered inference

00:21:48 Stop anthropomorphizing intelligence

00:27:44 Drug discovery as an incentive problem

00:32:29 The three-layer data market

00:38:07 Social knowledge, markets, and culture

00:45:39 Creator economics beyond Spotify

00:48:30 How science-fiction AI narratives mislead young builders

00:51:45 AI should improve humans, not replace them

00:56:42 Safety is a property of the whole system

00:58:12 Silicon Valley gurus and the cream off the top

01:00:47 Game theory, mechanism design, and contracts

01:04:39 Conformal prediction, e-values, and anytime inference

01:08:11 A new liberal arts triangle for the AI era

01:11:30 The Bayesian duck and markets as uncertainty reduction

ReScript (transcript, PDF, refs etc) - https://app.rescript.info/public/share/fb68f94af29d3745c6cf6125e01328b5

---

REFERENCES:

person:

[00:02:50] Michael I. Jordan (homepage)

https://people.eecs.berkeley.edu/~jordan/

paper:

[00:06:01] A Collectivist, Economic Perspective on AI

https://arxiv.org/abs/2507.06268

[00:18:09] AlphaFold

https://www.nature.com/articles/s41586-021-03819-2

[00:20:36] Prediction-Powered Inference

https://arxiv.org/abs/2301.09633

[00:33:47] On Three-Layer Data Markets

https://arxiv.org/abs/2402.09697

[01:04:39] Conformal Prediction with Conditional Guarantees

https://arxiv.org/abs/2107.07511

[01:04:51] A Tutorial on Conformal Prediction

https://www.jmlr.org/papers/v9/shafer08a.html

[01:06:00] E-Values Expand the Scope of Conformal Prediction

https://arxiv.org/abs/2503.13050

[01:08:23] Computational Thinking

https://www.cs.cmu.edu/~CompThink/papers/Wing06.pdf

other:

[00:28:20] How Should the FDA Test?

https://rdi.berkeley.edu/events/sbc-assets/pdfs/Summit%20session%20speaker%20slides%20submission%20form-s1-5%20%28File%20responses%29/Slides%20in%20PDF%20%28Please%20name%20the%20submitted%20file%20as%20_firstname_-_lastname_-slides.pdf%29.%20%28File%20responses%29/27-Michael%20Jordan-Session%20V.pdf#page=15

[00:28:40] Michael I. Jordan Session V Slides

<truncated, see ReScript link or YT VD>

Beth Barnes and David Rein on the one graph that ate the AI timelines discourse, and why the two people who built it are the most careful about how you read it.**SPONSOR**Prolific - Quality data. From real people. For faster breakthroughs.https://www.prolific.com/?utm_source=mlstInterview: https://youtu.be/cnxZZTl1tkk---Beth Barnes and David Rein from METR on the one graph that ate the AI timelines discourse, and why the people who built it are the most careful about how it gets read.Beth founded METR after leaving OpenAI alignment. David is first author on GPQA and co-author on HCAST and the METR Time Horizons paper. Together they built the measurement Daniel Kokotajlo called the single most important piece of evidence on AI timelines: the log-linear line of "how long a task a frontier model can complete at 50% reliability" vs release date.The conversation opens on reward hacking. Current models can articulate in chat why a behaviour is undesired and then execute it anyway as agents. From there: construct validity, Melanie Mitchell's four-problem taxonomy, and the ARC-AGI 1-to-2 collapse as a worked example of adversarially-selected benchmarks regressing once labs target them. Beth's counter: METR deliberately does not adversarially select. David's: models do not have to do the right thing for the right reasons.Methodology, then specification — David's compiler analogy, Beth on four-month tasks as expensive to evaluate rather than unspecifiable. Then the SWE-bench reality check, the METR finding that half of passing PRs would not be merged, and Beth's horses-versus-bank-tellers analogy for the labour market.The close: monitorability, the coin-spinning boat, two-year recursive self-improvement, and Beth's line that "overhyped now" and "big deal later" are not correlated claims.---TIMESTAMPS:00:00:00 Intro00:02:06 Sponsor break: Prolific human-feedback infrastructure00:02:33 Welcome and the scalable oversight motivation00:06:02 Construct validity, benchmark pathologies and the Chollet worry00:15:45 Time Horizons: human time, HCAST tasks and the 50% logistic00:24:50 Is human difficulty really one variable?00:33:05 Agent harness evolution and the inference-compute dividend00:40:00 Scaffolding bells, token budgets and the credit-assignment problem00:44:15 Look at the damn graph: regularisation bug and reliability nuance00:50:00 Why 50%? Reliability, reward hacking and pizza-party transcripts00:55:20 Extrapolation risk and straight lines on graphs00:59:25 Software engineering as a specification acquisition problem01:07:40 Compilers also made ugly code: vibe-coding quality and Claude on METR Slack01:15:15 Strongest defensible claim, Carlini's compiler swarm and AI 202701:23:45 SWE-bench merge rates, the bank-teller analogy and horses01:31:45 Scheming, alignment faking and the mentalistic vocabulary problem01:40:45 Reward hacking, monitorability and chain-of-thought faithfulness01:45:25 Recursive self-improvement, knowledge vs intelligence and closing

ReScript: https://app.rescript.info/public/share/de3bb40cc02ee39fdf36e2c60366eb4d

(PDF, refs, transcript etc)

Robert Lange, founding researcher at Sakana AI, joins Tim to discuss *Shinka Evolve* — a framework that combines LLMs with evolutionary algorithms to do open-ended program search. The core claim: systems like AlphaEvolve can optimize solutions to fixed problems, but real scientific progress requires co-evolving the problems themselves.

GTC is coming, the premier AI conference, great opportunity to learn about AI. NVIDIA and partners will showcase breakthroughs in physical AI, AI factories, agentic AI, and inference, exploring the next wave of AI innovation for developers and researchers. Register for virtual GTC for free, using my link and win NVIDIA DGX Spark (https://nvda.ws/4qQ0LMg)

• Why AlphaEvolve gets stuck — it needs a human to hand it the right problem. Shinka tries to invent new problems automatically, drawing on ideas from POET, PowerPlay, and MAP-Elites quality-diversity search.

• The *architecture* of Shinka: an archive of programs organized as islands, LLMs used as mutation operators, and a UCB bandit that adaptively selects between frontier models (GPT-5, Sonnet 4.5, Gemini) mid-run. The credit-assignment problem across models turns out to be genuinely hard.

• Concrete results — state-of-the-art circle packing with dramatically fewer evaluations, second place in an AtCoder competitive programming challenge, evolved load-balancing loss functions for mixture-of-experts models, and agent scaffolds for AIME math benchmarks.

• Are these systems actually thinking outside the box, or are they parasitic on their starting conditions? When LLMs run autonomously, "nothing interesting happens." Robert pushes back with the stepping-stone argument — evolution doesn't need to extrapolate, just recombine usefully.

• The AI Scientist question: can automated research pipelines produce real science, or just workshop-level slop that passes surface-level review? Robert is honest that the current version is more co-pilot than autonomous researcher.

• Where this lands in 5-20 years — Robert's prediction that scientific research will be fundamentally transformed, and Tim's thought experiment about alien mathematical artifacts that no human could have conceived.

Robert Lange: https://roberttlange.com/

---

TIMESTAMPS:

00:00:00 Introduction: Robert Lange, Sakana AI and Shinka Evolve

00:04:15 AlphaEvolve's Blind Spot: Co-Evolving Problems with Solutions

00:09:05 Unknown Unknowns, POET, and Auto-Curricula for AI Science

00:14:20 MAP-Elites and Quality-Diversity: Shinka's Evolutionary Architecture

00:28:00 UCB Bandits, Mutations and the Vibe Research Vision

00:40:00 Scaling Shinka: Meta-Evolution, Democratisation and the Three-Axis Model

00:47:10 Applications, ARC-AGI and the Future of Work

00:57:00 The AI Scientist and the Human Co-Pilot: Who Steers the Search?

01:06:00 AI Scientist v2, Slop Critique and the Future of Scientific Publishing

---

REFERENCES:

paper:

[00:03:30] ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution

https://arxiv.org/abs/2509.19349

[00:04:15] AlphaEvolve: A Coding Agent for Scientific and Algorithmic Discovery

https://arxiv.org/abs/2506.13131

[00:06:30] Darwin Godel Machine: Open-Ended Evolution of Self-Improving Agents

https://arxiv.org/abs/2505.22954

[00:09:05] Paired Open-Ended Trailblazer (POET)

https://arxiv.org/abs/1901.01753

[00:10:00] PowerPlay: Training an Increasingly General Problem Solver by Continually Searching for the Simplest Still Unsolvable Problem

https://arxiv.org/abs/1112.5309

[00:10:40] Automated Capability Discovery via Foundation Model Self-Exploration

https://arxiv.org/abs/2502.07577

[00:15:30] Illuminating Search Spaces by Mapping Elites (MAP-Elites)

https://arxiv.org/abs/1504.04909

[00:47:10] Automated Design of Agentic Systems (ADAS)

https://arxiv.org/abs/2408.08435

<trunc, see ReScript/YT>

PDF : https://app.rescript.info/api/sessions/b8a9dcf60623657c/pdf/download

Transcript: https://app.rescript.info/public/share/SDOD_3oXOcli3zTqcAtR8eibT5U3gam84oo4KRtI-Vk

Dive into the realities of AI-assisted coding, the origins of modern fine-tuning, and the cognitive science behind machine learning with fast.ai founder Jeremy Howard. In this episode, we unpack why AI might be turning software engineering into a slot machine and how to maintain true technical intuition in the age of large language models.

GTC is coming, the premier AI conference, great opportunity to learn about AI. NVIDIA and partners will showcase breakthroughs in physical AI, AI factories, agentic AI, and inference, exploring the next wave of AI innovation for developers and researchers. Register for virtual GTC for free, using my link and win NVIDIA DGX Spark (https://nvda.ws/4qQ0LMg)

Jeremy Howard is a renowned data scientist, researcher, entrepreneur, and educator. As the co-founder of fast.ai, former President of Kaggle, and the creator of ULMFiT, Jeremy has spent decades democratizing deep learning. His pioneering work laid the foundation for modern transfer learning and the pre-training and fine-tuning paradigm that powers today's language models.

Key Topics and Main Insights Discussed:

- The Origins of ULMFiT and Fine-Tuning

- The Vibe Coding Illusion and Software Engineering

- Cognitive Science, Friction, and Learning

- The Future of Developers

RESCRIPT: https://app.rescript.info/public/share/BhX5zP3b0m63srLOQDKBTFTooSzEMh_ARwmDG_h_izk

Jeremy Howard:

https://x.com/jeremyphoward

https://www.answer.ai/

---

TIMESTAMPS (fixed):

00:00:00 Introduction & GTC Sponsor

00:04:30 ULMFiT & The Birth of Fine-Tuning

00:12:00 Intuition & The Mechanics of Learning

00:18:30 Abstraction Hierarchies & AI Creativity

00:23:00 Claude Code & The Interpolation Illusion

00:27:30 Coding vs. Software Engineering

00:30:00 Cosplaying Intelligence: Dennett vs. Searle

00:36:30 Automation, Radiology & Desirable Difficulty

00:42:30 Organizational Knowledge & The Slope

00:48:00 Vibe Coding as a Slot Machine

00:54:00 The Erosion of Control in Software

01:01:00 Interactive Programming & REPL Environments

01:05:00 The Notebook Debate & Exploratory Science

01:17:30 AI Existential Risk & Power Centralization

01:24:20 Current Risks, Privacy & Enfeeblement

---

REFERENCES:

Blog Post:

[00:03:00] fast.ai Blog: Self-Supervised Learning

https://www.fast.ai/posts/2020-01-13-self_supervised.html

[00:13:30] DeepMind Blog: Gemini Deep Think

https://deepmind.google/blog/accelerating-mathematical-and-scientific-discovery-with-gemini-deep-think/

[00:19:30] Modular Blog: Claude C Compiler analysis

https://www.modular.com/blog/the-claude-c-compiler-what-it-reveals-about-the-future-of-software

[00:19:45] Anthropic Engineering Blog: Building C Compiler

https://www.anthropic.com/engineering/building-c-compiler

[00:48:00] Cursor Blog: Scaling Agents

https://cursor.com/blog/scaling-agents

[01:05:15] fast.ai Blog: NB Dev Merged Driver

https://www.fast.ai/posts/2022-08-25-jupyter-git.html

[01:17:30] Jeremy Howard: Response to AI Risk Letter

https://www.normaltech.ai/p/is-avoiding-extinction-from-ai-really

Book:

[00:08:30] M. Chirimuuta: The Brain Abstracted

https://mitpress.mit.edu/9780262548045/the-brain-abstracted/

[00:30:00] Daniel Dennett: Consciousness Explained

https://www.amazon.com/Consciousness-Explained-Daniel-C-Dennett/dp/0316180661

[00:42:30] Cesar Hidalgo: Infinite Alphabet / Laws of Knowledge

https://www.amazon.com/Infinite-Alphabet-Laws-Knowledge/dp/0241655676

Archive Article:

[00:13:45] MLST Archive: Why Creativity Cannot Be Interpolated

https://archive.mlst.ai/read/why-creativity-cannot-be-interpolated

Research Study:

[00:24:30] METR Study: AI OS Development

https://metr.org/blog/2025-07-10-early-2025-ai-experienced-os-dev-study/

Paper:

[00:24:45] Fred Brooks: No Silver Bullet

https://www.cs.unc.edu/techreports/86-020.pdf

[00:30:15] John Searle: Minds, Brains, and Programs

https://www.cambridge.org/core/journals/behavioral-and-brain-sciences/article/minds-brains-and-programs/DC644B47A4299C637C89772FACC2706A

What if life itself is just a really sophisticated computer program that wrote itself into existence?

Blaise Agüera y Arcas presenting at ALife 2025 — the most technically detailed public walkthrough of the ideas in his *What is Life?* and *What is Intelligence?* books that we've come across.He covers the BFF experiments (self-replicating programs emerging spontaneously from random noise), the mathematical framework connecting Lotka-Volterra population dynamics with Smoluchowski coagulation, eigenvalue analysis of cooperation matrices, and his central claim that symbiogenesis — not mutation — is the primary engine of evolutionary novelty.The experimental results are genuinely striking: complex self-replicating code arising from random byte strings with zero mutation, a sharp phase transition that looks like gelation, and a proof that blocking deep symbiogenetic ancestry trees prevents the transition entirely.A few things worth flagging for critical viewers:— The substrate is more carefully engineered than the framing sometimes suggests. The choice of language, tape length, interaction protocol, and step limits all shape what emerges. Their own SUBLEQ counterexample (where self-replicators *don't* arise despite being theoretically possible) highlights that these design choices matter substantially — and a general theory of which substrates support this transition is still missing.— The leap from "self-replicating programs on fixed-length tapes" to "life was computational and intelligent from the start" involves significant philosophical extrapolation beyond what the experiments directly demonstrate.— The Bedau et al. (2000) open problems paper he references at the start actually sets a higher bar for Challenge 3.2 than BFF currently meets: it asks that "the internal organization of these 'organisms' and the boundaries separating them from their environment arise and be sustained through the activities of lower-level primitives" — whereas BFF's tape boundaries are fixed by design, not emergent.

---

TIMESTAMPS:

00:00:00 Introduction: From Noise to Programs & ALife History

00:03:15 Defining Life: Function as the "Spirit"

00:05:45 Von Neumann's Insight: Life is Embodied Computation

00:09:15 Physics of Computation: Irreversibility & Fallacies

00:15:00 The BFF Experiment: Spontaneous Generation of Code

00:23:45 The Mystery: Complexity Growth Without Mutation

00:27:00 Symbiogenesis: The Engine of Novelty

00:33:15 Mathematical Proof: Blocking Symbiosis Stops Life

00:40:15 Evolutionary Implications: It's Symbiogenesis All The Way Down

00:44:30 Intelligence as Modeling Others

00:46:49 Q&A: Levels of Abstraction & Definitions

---

REFERENCES:

Paper:

[00:01:16] Open Problems in Artificial Life

https://direct.mit.edu/artl/article/6/4/363/2354/Open-Problems-in-Artificial-Life

[00:09:30] When does a physical system compute?

https://arxiv.org/abs/1309.7979

[00:15:00] Computational Life

https://arxiv.org/abs/2406.19108

[00:27:30] On the Origin of Mitosing Cells

https://pubmed.ncbi.nlm.nih.gov/11541392/

[00:42:00] The Major Evolutionary Transitions

https://www.nature.com/articles/374227a0

[00:44:00] The ARC gene

https://www.nih.gov/news-events/news-releases/memory-gene-goes-viral

Person:

[00:05:45] Alan Turing

https://plato.stanford.edu/entries/turing/

[00:07:30] John von Neumann

https://en.wikipedia.org/wiki/John_von_Neumann

[00:11:15] Hector Zenil

https://hectorzenil.net/

[00:12:00] Robert Sapolsky

https://profiles.stanford.edu/robert-sapolsky

---

LINKS:

RESCRIPT: https://app.rescript.info/public/share/ff7gb6HpezOR3DF-gr9-rCoMFzzEgUjLQK6voV5XVWY

What makes something truly *intelligent?* Is a rock an agent? Could a perfect simulation of your brain actually *be* you? In this fascinating conversation, Dr. Jeff Beck takes us on a journey through the philosophical and technical foundations of agency, intelligence, and the future of AI.

Jeff doesn't hold back on the big questions. He argues that from a purely mathematical perspective, there's no structural difference between an agent and a rock – both execute policies that map inputs to outputs. The real distinction lies in *sophistication* – how complex are the internal computations? Does the system engage in planning and counterfactual reasoning, or is it just a lookup table that happens to give the right answers?

*Key topics explored in this conversation:*

*The Black Box Problem of Agency* – How can we tell if something is truly planning versus just executing a pre-computed response? Jeff explains why this question is nearly impossible to answer from the outside, and why the best we can do is ask which model gives us the simplest explanation.

*Energy-Based Models Explained* – A masterclass on how EBMs differ from standard neural networks. The key insight: traditional networks only optimize weights, while energy-based models optimize *both* weights and internal states – a subtle but profound distinction that connects to Bayesian inference.

*Why Your Brain Might Have Evolved from Your Nose* – One of the most surprising moments in the conversation. Jeff proposes that the complex, non-smooth nature of olfactory space may have driven the evolution of our associative cortex and planning abilities.

*The JEPA Revolution* – A deep dive into Yann LeCun's Joint Embedding Prediction Architecture and why learning in latent space (rather than predicting every pixel) might be the key to more robust AI representations.

*AI Safety Without Skynet Fears* – Jeff takes a refreshingly grounded stance on AI risk. He's less worried about rogue superintelligences and more concerned about humans becoming "reward function selectors" – couch potatoes who just approve or reject AI outputs. His proposed solution? Use inverse reinforcement learning to derive AI goals from observed human behavior, then make *small* perturbations rather than naive commands like "end world hunger."

Whether you're interested in the philosophy of mind, the technical details of modern machine learning, or just want to understand what makes intelligence *tick,* this conversation delivers insights you won't find anywhere else.

---

TIMESTAMPS:

00:00:00 Geometric Deep Learning & Physical Symmetries

00:00:56 Defining Agency: From Rocks to Planning

00:05:25 The Black Box Problem & Counterfactuals

00:08:45 Simulated Agency vs. Physical Reality

00:12:55 Energy-Based Models & Test-Time Training

00:17:30 Bayesian Inference & Free Energy

00:20:07 JEPA, Latent Space, & Non-Contrastive Learning

00:27:07 Evolution of Intelligence & Modular Brains

00:34:00 Scientific Discovery & Automated Experimentation

00:38:04 AI Safety, Enfeeblement & The Future of Work

---

REFERENCES:

Concept:

[00:00:58] Free Energy Principle (FEP)

https://en.wikipedia.org/wiki/Free_energy_principle

[00:06:00] Monte Carlo Tree Search

https://en.wikipedia.org/wiki/Monte_Carlo_tree_search

Book:

[00:09:00] The Intentional Stance

https://mitpress.mit.edu/9780262540537/the-intentional-stance/

Paper:

[00:13:00] A Tutorial on Energy-Based Learning (LeCun 2006)

http://yann.lecun.com/exdb/publis/pdf/lecun-06.pdf

[00:15:00] Auto-Encoding Variational Bayes (VAE)

https://arxiv.org/abs/1312.6114

[00:20:15] JEPA (Joint Embedding Prediction Architecture)

https://openreview.net/forum?id=BZ5a1r-kVsf

[00:22:30] The Wake-Sleep Algorithm

https://www.cs.toronto.edu/~hinton/absps/ws.pdf

<trunc, see rescript>

---

RESCRIPT:

https://app.rescript.info/public/share/DJlSbJ_Qx080q315tWaqMWn3PixCQsOcM4Kf1IW9_Eo

PDF:

https://app.rescript.info/api/public/sessions/0efec296b9b6e905/pdf

Professor Mazviita Chirimuuta joins us for a fascinating deep dive into the philosophy of neuroscience and what it really means to understand the mind.*What can neuroscience actually tell us about how the mind works?* In this thought-provoking conversation, we explore the hidden assumptions behind computational theories of the brain, the limits of scientific abstraction, and why the question of machine consciousness might be more complicated than AI researchers assume.Mazviita, author of *The Brain Abstracted,* brings a unique perspective shaped by her background in both neuroscience research and philosophy. She challenges us to think critically about the metaphors we use to understand cognition — from the reflex theory of the late 19th century to today's dominant view of the brain as a computer.*Key topics explored:**The problem of oversimplification* — Why scientific models necessarily leave things out, and how this can sometimes lead entire fields astray. The cautionary tale of reflex theory shows how elegant explanations can blind us to biological complexity.*Is the brain really a computer?* — Mazviita unpacks the philosophical assumptions behind computational neuroscience and asks: if we can model anything computationally, what makes brains special? The answer might challenge everything you thought you knew about AI.*Haptic realism* — A fresh way of thinking about scientific knowledge that emphasizes interaction over passive observation. Knowledge isn't about reading the "source code of the universe" — it's something we actively construct through engagement with the world.*Why embodiment matters for understanding* — Can a disembodied language model truly understand? Mazviita makes a compelling case that human cognition is deeply entangled with our sensory-motor engagement and biological existence in ways that can't simply be abstracted away.*Technology and human finitude* — Drawing on Heidegger, we discuss how the dream of transcending our physical limitations through technology might reflect a fundamental misunderstanding of what it means to be a knower.This conversation is essential viewing for anyone interested in AI, consciousness, philosophy of mind, or the future of cognitive science. Whether you're skeptical of strong AI claims or a true believer in machine consciousness, Mazviita's careful philosophical analysis will give you new tools for thinking through these profound questions.---TIMESTAMPS:00:00:00 The Problem of Generalizing Neuroscience00:02:51 Abstraction vs. Idealization: The "Kaleidoscope"00:05:39 Platonism in AI: Discovering or Inventing Patterns?00:09:42 When Simplification Fails: The Reflex Theory00:12:23 Behaviorism and the "Black Box" Trap00:14:20 Haptic Realism: Knowledge Through Interaction00:20:23 Is Nature Protean? The Myth of Converging Truth00:23:23 The Computational Theory of Mind: A Useful Fiction?00:27:25 Biological Constraints: Why Brains Aren't Just Neural Nets00:31:01 Agency, Distal Causes, and Dennett's Stances00:37:13 Searle's Challenge: Causal Powers and Understanding00:41:58 Heidegger's Warning & The Experiment on Children---REFERENCES:Book:[00:01:28] The Brain Abstractedhttps://mitpress.mit.edu/9780262548045/the-brain-abstracted/[00:11:05] The Integrated Action of the Nervous Systemhttps://www.amazon.sg/integrative-action-nervous-system/dp/9354179029[00:18:15] The Quest for Certainty (Dewey)https://www.amazon.com/Quest-Certainty-Relation-Knowledge-Lectures/dp/0399501916[00:19:45] Realism for Realistic People (Chang)https://www.cambridge.org/core/books/realism-for-realistic-people/ACC93A7F03B15AA4D6F3A466E3FC5AB7<truncated, see ReScript>---RESCRIPT:https://app.rescript.info/public/share/A6cZ1TY35p8ORMmYCWNBI0no9ChU3-Kx7dPXGJURvZ0PDF Transcript:https://app.rescript.info/api/public/sessions/0fb7767e066cf712/pdf

What if everything we think we know about the brain is just a really good metaphor that we forgot was a metaphor?This episode takes you on a journey through the history of scientific simplification, from a young Karl Friston watching wood lice in his garden to the bold claims that your mind is literally software running on biological hardware.We bring together some of the most brilliant minds we've interviewed — Professor Mazviita Chirimuuta, Francois Chollet, Joscha Bach, Professor Luciano Floridi, Professor Noam Chomsky, Nobel laureate John Jumper, and more — to wrestle with a deceptively simple question: *When scientists simplify reality to study it, what gets captured and what gets lost?**Key ideas explored:**The Spherical Cow Problem* — Science requires simplification. We're limited creatures trying to understand systems far more complex than our working memory can hold. But when does a useful model become a dangerous illusion?*The Kaleidoscope Hypothesis* — Francois Chollet's beautiful idea that beneath all the apparent chaos of reality lies simple, repeating patterns — like bits of colored glass in a kaleidoscope creating infinite complexity. Is this profound truth or Platonic wishful thinking?*Is Software Really Spirit?* — Joscha Bach makes the provocative claim that software is literally spirit, not metaphorically. We push back on this, asking whether the "sameness" we see across different computers running the same program exists in nature or only in our descriptions.*The Cultural Illusion of AGI* — Why does artificial general intelligence seem so inevitable to people in Silicon Valley? Professor Chirimuuta suggests we might be caught in a "cultural historical illusion" — our mechanistic assumptions about minds making AI seem like destiny when it might just be a bet.*Prediction vs. Understanding* — Nobel Prize winner John Jumper: AI can predict and control, but understanding requires a human in the loop. Throughout history, we've described the brain as hydraulic pumps, telegraph networks, telephone switchboards, and now computers. Each metaphor felt obviously true at the time. This episode asks: what will we think was naive about our current assumptions in fifty years?Featuring insights from *The Brain Abstracted* by Mazviita Chirimuuta — possibly the most influential book on how we think about thinking in 2025.---TIMESTAMPS:00:00:00 The Wood Louse & The Spherical Cow00:02:04 The Necessity of Abstraction00:04:42 Simplicius vs. Ignorantio: The Boxing Match00:06:39 The Kaleidoscope Hypothesis00:08:40 Is the Mind Software?00:13:15 Critique of Causal Patterns00:14:40 Temperature is Not a Thing00:18:24 The Ship of Theseus & Ontology00:23:45 Metaphors Hardening into Reality00:25:41 The Illusion of AGI Inevitability00:27:45 Prediction vs. Understanding00:32:00 Climbing the Mountain vs. The Helicopter00:34:53 Haptic Realism & The Limits of Knowledge---REFERENCES:Person:[00:00:00] Karl Friston (UCL)https://profiles.ucl.ac.uk/1236-karl-friston[00:06:30] Francois Chollethttps://fchollet.com/[00:14:41] Cesar Hidalgo, MLST interview.https://www.youtube.com/watch?v=vzpFOJRteeI[00:30:30] Terence Tao's Bloghttps://terrytao.wordpress.com/Book:[00:02:25] The Brain Abstractedhttps://mitpress.mit.edu/9780262548045/the-brain-abstracted/[00:06:00] On Learned Ignorancehttps://www.amazon.com/Nicholas-Cusa-learned-ignorance-translation/dp/0938060236[00:24:15] Science and the Modern Worldhttps://amazon.com/dp/0684836394<truncated, see ReScript>

RESCRIPT:https://app.rescript.info/public/share/CYy0ex2M2kvcVRdMnSUky5O7H7hB7v2u_nVhoUiuKD4PDF Transcript: https://app.rescript.info/api/public/sessions/6c44c41e1e0fa6dd/pdf

Thank you to Dr. Maxwell Ramstead for early script work on this show (Ph.D student of Friston) and the woodlice story came from him!

Dr. Jeff Beck, mathematician turned computational neuroscientist, joins us for a fascinating deep dive into why the future of AI might look less like ChatGPT and more like your own brain.

**SPONSOR MESSAGES START**

—

Prolific - Quality data. From real people. For faster breakthroughs.

https://www.prolific.com/?utm_source=mlst

—

**END**

*What if the key to building truly intelligent machines isn't bigger models, but smarter ones?*

In this conversation, Jeff makes a compelling case that we've been building AI backwards. While the tech industry races to scale up transformers and language models, Jeff argues we're missing something fundamental: the brain doesn't work like a giant prediction engine. It works like a scientist, constantly testing hypotheses about a world made of *objects* that interact through *forces* — not pixels and tokens.

*The Bayesian Brain* — Jeff explains how your brain is essentially running the scientific method on autopilot. When you combine what you see with what you hear, you're doing optimal Bayesian inference without even knowing it. This isn't just philosophy — it's backed by decades of behavioral experiments showing humans are surprisingly efficient at handling uncertainty.

*AutoGrad Changed Everything* — Forget transformers for a moment. Jeff argues the real hero of the AI boom was automatic differentiation, which turned AI from a math problem into an engineering problem. But in the process, we lost sight of what actually makes intelligence work.

*The Cat in the Warehouse Problem* — Here's where it gets practical. Imagine a warehouse robot that's never seen a cat. Current AI would either crash or make something up. Jeff's approach? Build models that *know what they don't know*, can phone a friend to download new object models on the fly, and keep learning continuously. It's like giving robots the ability to say "wait, what IS that?" instead of confidently being wrong.

*Why Language is a Terrible Model for Thought* — In a provocative twist, Jeff argues that grounding AI in language (like we do with LLMs) is fundamentally misguided. Self-report is the least reliable data in psychology — people routinely explain their own behavior incorrectly. We should be grounding AI in physics, not words.

*The Future is Lots of Little Models* — Instead of one massive neural network, Jeff envisions AI systems built like video game engines: thousands of small, modular object models that can be combined, swapped, and updated independently. It's more efficient, more flexible, and much closer to how we actually think.

Rescript: https://app.rescript.info/public/share/D-b494t8DIV-KRGYONJghvg-aelMmxSDjKthjGdYqsE

---

TIMESTAMPS:

00:00:00 Introduction & The Bayesian Brain

00:01:25 Bayesian Inference & Information Processing

00:05:17 The Brain Metaphor: From Levers to Computers

00:10:13 Micro vs. Macro Causation & Instrumentalism

00:16:59 The Active Inference Community & AutoGrad

00:22:54 Object-Centered Models & The Grounding Problem

00:35:50 Scaling Bayesian Inference & Architecture Design

00:48:05 The Cat in the Warehouse: Solving Generalization

00:58:17 Alignment via Belief Exchange

01:05:24 Deception, Emergence & Cellular Automata

---

REFERENCES:

Paper:

[00:00:24] Zoubin Ghahramani (Google DeepMind)

https://pmc.ncbi.nlm.nih.gov/articles/PMC3538441/pdf/rsta201

[00:19:20] Mamba: Linear-Time Sequence Modeling

https://arxiv.org/abs/2312.00752

[00:27:36] xLSTM: Extended Long Short-Term Memory

https://arxiv.org/abs/2405.04517

[00:41:12] 3D Gaussian Splatting

https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/

[01:07:09] Lenia: Biology of Artificial Life

https://arxiv.org/abs/1812.05433

[01:08:20] Growing Neural Cellular Automata

https://distill.pub/2020/growing-ca/

[01:14:05] DreamCoder

https://arxiv.org/abs/2006.08381

[01:14:58] The Genomic Bottleneck

https://www.nature.com/articles/s41467-019-11786-6

Person:

[00:16:42] Karl Friston (UCL)

https://www.youtube.com/watch?v=PNYWi996Beg

Tim sits down with Max Bennett to explore how our brains evolved over 600 million years—and what that means for understanding both human intelligence and AI.

Max isn't a neuroscientist by training. He's a tech entrepreneur who got curious, started reading, and ended up weaving together three fields that rarely talk to each other: comparative psychology (what different animals can actually do), evolutionary neuroscience (how brains changed over time), and AI (what actually works in practice).

*Your Brain Is a Guessing Machine*

You don't actually "see" the world. Your brain builds a simulation of what it *thinks* is out there and just uses your eyes to check if it's right. That's why optical illusions work—your brain is filling in a triangle that isn't there, or can't decide if it's looking at a duck or a rabbit.

*Rats Have Regrets*

*Chimps Are Machiavellian*

*Language Is the Human Superpower*

*Does ChatGPT Think?*

(truncated description, more on rescript)

Understanding how the brain evolved isn't just about the past. It gives us clues about:

- What's actually different between human intelligence and AI

- Why we're so easily fooled by status games and tribal thinking

- What features we might want to build into—or leave out of—future AI systems

Get Max's book:

https://www.amazon.com/Brief-History-Intelligence-Humans-Breakthroughs/dp/0063286343

Rescript: https://app.rescript.info/public/share/R234b7AXyDXZusqQ_43KMGsUSvJ2TpSz2I3emnI6j9A

---

TIMESTAMPS:

00:00:00 Introduction: Outsider's Advantage & Neocortex Theories

00:11:34 Perception as Inference: The Filling-In Machine

00:19:11 Understanding, Recognition & Generative Models

00:36:39 How Mice Plan: Vicarious Trial & Error

00:46:15 Evolution of Self: The Layer 4 Mystery

00:58:31 Ancient Minds & The Social Brain: Machiavellian Apes

01:19:36 AI Alignment, Instrumental Convergence & Status Games

01:33:07 Metacognition & The IQ Paradox

01:48:40 Does GPT Have Theory of Mind?

02:00:40 Memes, Language Singularity & Brain Size Myths

02:16:44 Communication, Language & The Cyborg Future

02:44:25 Shared Fictions, World Models & The Reality Gap

---

REFERENCES:Person:

[00:00:05] Karl Friston (UCL)

https://www.youtube.com/watch?v=PNYWi996Beg

[00:00:06] Jeff Hawkins

https://www.youtube.com/watch?v=6VQILbDqaI4

[00:12:19] Hermann von Helmholtz

https://plato.stanford.edu/entries/hermann-helmholtz/

[00:38:34] David Redish (U. Minnesota)

https://redishlab.umn.edu/

[01:10:19] Robin Dunbar

https://www.psy.ox.ac.uk/people/robin-dunbar

[01:15:04] Emil Menzel

https://www.sciencedirect.com/bookseries/behavior-of-nonhuman-primates/vol/5/suppl/C

[01:19:49] Nick Bostrom

https://nickbostrom.com/

[02:28:25] Noam Chomsky

https://linguistics.mit.edu/user/chomsky/

[03:01:22] Judea Pearl

https://samueli.ucla.edu/people/judea-pearl/

Concept/Framework:

[00:05:04] Active Inference

https://www.youtube.com/watch?v=KkR24ieh5Ow

Paper:

[00:35:59] Predictions not commands [Rick A Adams]

https://pubmed.ncbi.nlm.nih.gov/23129312/

Book:

[01:25:42] The Elephant in the Brain

https://www.amazon.com/Elephant-Brain-Hidden-Motives-Everyday/dp/0190495995

[01:28:27] The Status Game

https://www.goodreads.com/book/show/58642436-the-status-game

[02:00:40] The Selfish Gene

https://amazon.com/dp/0198788606

[02:14:25] The Language Game

https://www.amazon.com/Language-Game-Improvisation-Created-Changed/dp/1541674987

[02:54:40] The Evolution of Language

https://www.amazon.com/Evolution-Language-Approaches/dp/052167736X

[03:09:37] The Three-Body Problem

https://amazon.com/dp/0765377063

César Hidalgo has spent years trying to answer a deceptively simple question: What is knowledge, and why is it so hard to move around?

We all have this intuition that knowledge is just... information. Write it down in a book, upload it to GitHub, train an AI on it—done. But César argues that's completely wrong. Knowledge isn't a thing you can copy and paste. It's more like a living organism that needs the right environment, the right people, and constant exercise to survive.

Guest: César Hidalgo, Director of the Center for Collective Learning

1. Knowledge Follows Laws (Like Physics)

2. You Can't Download Expertise

3. Why Big Companies Fail to Adapt

4. The "Infinite Alphabet" of Economies

If you think AI can just "copy" human knowledge, or that development is just about throwing money at poor countries, or that writing things down preserves them forever—this conversation will change your mind. Knowledge is fragile, specific, and collective. It decays fast if you don't use it.

The Infinite Alphabet [César A. Hidalgo]

https://www.penguin.co.uk/books/458054/the-infinite-alphabet-by-hidalgo-cesar-a/9780241655672

https://x.com/cesifoti

Rescript link.

https://app.rescript.info/public/share/eaBHbEo9xamwbwpxzcVVm4NQjMh7lsOQKeWwNxmw0JQ

---

TIMESTAMPS:

00:00:00 The Three Laws of Knowledge

00:02:28 Rival vs. Non-Rival: The Economics of Ideas

00:05:43 Why You Can't Just 'Download' Knowledge

00:08:11 The Detective Novel Analogy

00:11:54 Collective Learning & Organizational Networks

00:16:27 Architectural Innovation: Amazon vs. Barnes & Noble

00:19:15 The First Law: Learning Curves

00:23:05 The Samuel Slater Story: Treason & Memory

00:28:31 Physics of Knowledge: Joule's Cannon

00:32:33 Extensive vs. Intensive Properties

00:35:45 Knowledge Decay: Ise Temple & Polaroid

00:41:20 Absorptive Capacity: Sony & Donetsk

00:47:08 Disruptive Innovation & S-Curves

00:51:23 Team Size & The Cost of Innovation

00:57:13 Geography of Knowledge: Vespa's Origin

01:04:34 Migration, Diversity & 'Planet China'

01:12:02 Institutions vs. Knowledge: The China Story

01:21:27 Economic Complexity & The Infinite Alphabet

01:32:27 Do LLMs Have Knowledge?

---

REFERENCES:

Book:

[00:47:45] The Innovator's Dilemma (Christensen)

https://www.amazon.com/Innovators-Dilemma-Revolutionary-Change-Business/dp/0062060244

[00:55:15] Why Greatness Cannot Be Planned

https://amazon.com/dp/3319155237

[01:35:00] Why Information Grows

https://amazon.com/dp/0465048994

Paper:

[00:03:15] Endogenous Technological Change (Romer, 1990)

https://web.stanford.edu/~klenow/Romer_1990.pdf

[00:03:30] A Model of Growth Through Creative Destruction (Aghion & Howitt, 1992)

https://dash.harvard.edu/server/api/core/bitstreams/7312037d-2b2d-6bd4-e053-0100007fdf3b/content

[00:14:55] Organizational Learning: From Experience to Knowledge (Argote & Miron-Spektor, 2011)

https://www.researchgate.net/publication/228754233_Organizational_Learning_From_Experience_to_Knowledge

[00:17:05] Architectural Innovation (Henderson & Clark, 1990)

https://www.researchgate.net/publication/200465578_Architectural_Innovation_The_Reconfiguration_of_Existing_Product_Technologies_and_the_Failure_of_Established_Firms

[00:19:45] The Learning Curve Equation (Thurstone, 1916)

https://dn790007.ca.archive.org/0/items/learningcurveequ00thurrich/learningcurveequ00thurrich.pdf

[00:21:30] Factors Affecting the Cost of Airplanes (Wright, 1936)

https://pdodds.w3.uvm.edu/research/papers/others/1936/wright1936a.pdf

[00:52:45] Are Ideas Getting Harder to Find? (Bloom et al.)

https://web.stanford.edu/~chadj/IdeaPF.pdf

[01:33:00] LLMs/ Emergence

https://arxiv.org/abs/2506.11135

Person:

[00:25:30] Samuel Slater

https://en.wikipedia.org/wiki/Samuel_Slater

[00:42:05] Masaru Ibuka (Sony)

https://www.sony.com/en/SonyInfo/CorporateInfo/History/SonyHistory/1-02.html

This is a lively, no-holds-barred debate about whether AI can truly be intelligent, conscious, or understand anything at all — and what happens when (or if) machines become smarter than us.

Dr. Mike Israetel is a sports scientist, entrepreneur, and co-founder of RP Strength (a fitness company). He describes himself as a "dilettante" in AI but brings a fascinating outsider's perspective.

Jared Feather (IFBB Pro bodybuilder and exercise physiologist)

The Big Questions:

1. When is superintelligence coming?

2. Does AI actually understand anything?

3. The Simulation Debate (The Spiciest Part)

4. Will AI kill us all? (The Doomer Debate)

5. What happens to human jobs and purpose?

6. Do we need suffering?

Mikes channel: https://www.youtube.com/channel/UCfQgsKhHjSyRLOp9mnffqVg

RESCRIPT INTERACTIVE PLAYER: https://app.rescript.info/public/share/GVMUXHCqctPkXH8WcYtufFG7FQcdJew_RL_MLgMKU1U

---

TIMESTAMPS:

00:00:00 Introduction & Workout Demo

00:04:15 ASI Timelines & Definitions

00:10:24 The Embodiment Debate

00:18:28 Neutrinos & Abstract Knowledge

00:25:56 Can AI Learn From YouTube?

00:31:25 Diversity of Intelligence

00:36:00 AI Slop & Understanding

00:45:18 The Simulation Argument: Fire & Water

00:58:36 Consciousness & Zombies

01:04:30 Do Reasoning Models Actually Reason?

01:12:00 The Live Learning Problem

01:19:15 Superintelligence & Benevolence

01:28:59 What is True Agency?

01:37:20 Game Theory & The &quot;Kill All Humans&quot; Fallacy

01:48:05 Regulation & The China Factor

01:55:52 Mind Uploading & The Future of Love

02:04:41 Economics of ASI: Will We Be Useless?

02:13:35 The Matrix & The Value of Suffering

02:17:30 Transhumanism & Inequality

02:21:28 Debrief: AI Medical Advice & Final Thoughts

---

REFERENCES:

Paper:

[00:10:45] Alchemy and Artificial Intelligence (Dreyfus)

https://www.rand.org/content/dam/rand/pubs/papers/2006/P3244.pdf

[00:10:55] The Chinese Room Argument (John Searle)

https://home.csulb.edu/~cwallis/382/readings/482/searle.minds.brains.programs.bbs.1980.pdf

[00:11:05] The Symbol Grounding Problem (Stephen Harnad)

https://arxiv.org/html/cs/9906002

[00:23:00] Attention Is All You Need

https://arxiv.org/abs/1706.03762

[00:45:00] GPT-4 Technical Report

https://arxiv.org/abs/2303.08774

[01:45:00] Anthropic Agentic Misalignment Paper

https://www.anthropic.com/research/agentic-misalignment

[02:17:45] Retatrutide

https://pubmed.ncbi.nlm.nih.gov/37366315/

Organization:

[00:15:50] CERN

https://home.cern/

[01:05:00] METR Long Horizon Evaluations

https://evaluations.metr.org/

MLST Episode:

[00:23:10] MLST: Llion Jones - Inventors' Remorse

https://www.youtube.com/watch?v=DtePicx_kFY

[00:50:30] MLST: Blaise Agüera y Arcas Interview

https://www.youtube.com/watch?v=rMSEqJ_4EBk

[01:10:00] MLST: David Krakauer

https://www.youtube.com/watch?v=dY46YsGWMIc

Event:

[00:23:40] ARC Prize/Challenge

https://arcprize.org/

Book:

[00:24:45] The Brain Abstracted

https://www.amazon.com/Brain-Abstracted-Simplification-Philosophy-Neuroscience/dp/0262548046

[00:47:55] Pamela McCorduck

https://www.amazon.com/Machines-Who-Think-Artificial-Intelligence/dp/1568812051

[01:23:15] The Singularity Is Nearer (Ray Kurzweil)

https://www.amazon.com/Singularity-Nearer-Ray-Kurzweil-ebook/dp/B08Y6FYJVY

[01:27:35] A Fire Upon The Deep (Vernor Vinge)

https://www.amazon.com/Fire-Upon-Deep-S-F-MASTERWORKS-ebook/dp/B00AVUMIZE/

[02:04:50] Deep Utopia (Nick Bostrom)

https://www.amazon.com/Deep-Utopia-Meaning-Solved-World/dp/1646871642

[02:05:00] Technofeudalism (Yanis Varoufakis)

https://www.amazon.com/Technofeudalism-Killed-Capitalism-Yanis-Varoufakis/dp/1685891241

Visual Context Needed:

[00:29:40] AT-AT Walker (Star Wars)

https://starwars.fandom.com/wiki/All_Terrain_Armored_Transport

Person:

[00:33:15] Andrej Karpathy

https://karpathy.ai/

Video:

[01:40:00] Mike Israetel vs Liron Shapira AI Doom Debate

https://www.youtube.com/watch?v=RaDWSPMdM4o

Company:

[02:26:30] Examine.com

https://examine.com/

We often think of Large Language Models (LLMs) as all-knowing, but as the team reveals, they still struggle with the logic of a second-grader. Why can’t ChatGPT reliably add large numbers? Why does it "hallucinate" the laws of physics? The answer lies in the architecture. This episode explores how *Category Theory* —an ultra-abstract branch of mathematics—could provide the "Periodic Table" for neural networks, turning the "alchemy" of modern AI into a rigorous science.

In this deep-dive exploration, *Andrew Dudzik*, *Petar Velichkovich*, *Taco Cohen*, *Bruno Gavranović*, and *Paul Lessard* join host *Tim Scarfe* to discuss the fundamental limitations of today’s AI and the radical mathematical framework that might fix them.

TRANSCRIPT:

https://app.rescript.info/public/share/LMreunA-BUpgP-2AkuEvxA7BAFuA-VJNAp2Ut4MkMWk

---

Key Insights in This Episode:

* *The "Addition" Problem:* *Andrew Dudzik* explains why LLMs don't actually "know" math—they just recognize patterns. When you change a single digit in a long string of numbers, the pattern breaks because the model lacks the internal "machinery" to perform a simple carry operation.

* *Beyond Alchemy:* deep learning is currently in its "alchemy" phase—we have powerful results, but we lack a unifying theory. Category Theory is proposed as the framework to move AI from trial-and-error to principled engineering. [00:13:49]

* *Algebra with Colors:* To make Category Theory accessible, the guests use brilliant analogies—like thinking of matrices as *magnets with colors* that only snap together when the types match. This "partial compositionality" is the secret to building more complex internal reasoning. [00:09:17]

* *Synthetic vs. Analytic Math:* *Paul Lessard* breaks down the philosophical shift needed in AI research: moving from "Analytic" math (what things are made of) to "Synthetic" math [00:23:41]

---

Why This Matters for AGI

If we want AI to solve the world's hardest scientific problems, it can't just be a "stochastic parrot." It needs to internalize the rules of logic and computation. By imbuing neural networks with categorical priors, researchers are attempting to build a future where AI doesn't just predict the next word—it understands the underlying structure of the universe.

---

TIMESTAMPS:

00:00:00 The Failure of LLM Addition & Physics

00:01:26 Tool Use vs Intrinsic Model Quality

00:03:07 Efficiency Gains via Internalization

00:04:28 Geometric Deep Learning & Equivariance

00:07:05 Limitations of Group Theory

00:09:17 Category Theory: Algebra with Colors

00:11:25 The Systematic Guide of Lego-like Math

00:13:49 The Alchemy Analogy & Unifying Theory

00:15:33 Information Destruction & Reasoning

00:18:00 Pathfinding & Monoids in Computation

00:20:15 System 2 Reasoning & Error Awareness

00:23:31 Analytic vs Synthetic Mathematics

00:25:52 Morphisms & Weight Tying Basics

00:26:48 2-Categories & Weight Sharing Theory

00:28:55 Higher Categories & Emergence

00:31:41 Compositionality & Recursive Folds

00:34:05 Syntax vs Semantics in Network Design

00:36:14 Homomorphisms & Multi-Sorted Syntax

00:39:30 The Carrying Problem & Hopf Fibrations

Petar Veličković (GDM)

https://petar-v.com/

Paul Lessard

https://www.linkedin.com/in/paul-roy-lessard/

Bruno Gavranović

https://www.brunogavranovic.com/

Andrew Dudzik (GDM)

https://www.linkedin.com/in/andrew-dudzik-222789142/

---

REFERENCES:

Model:

[00:01:05] Veo

https://deepmind.google/models/veo/

[00:01:10] Genie

https://deepmind.google/blog/genie-3-a-new-frontier-for-world-models/

Paper:

[00:04:30] Geometric Deep Learning Blueprint

https://arxiv.org/abs/2104.13478

https://www.youtube.com/watch?v=bIZB1hIJ4u8

[00:16:45] AlphaGeometry

https://arxiv.org/abs/2401.08312

[00:16:55] AlphaCode

https://arxiv.org/abs/2203.07814

[00:17:05] FunSearch

https://www.nature.com/articles/s41586-023-06924-6

[00:37:00] Attention Is All You Need

https://arxiv.org/abs/1706.03762

[00:43:00] Categorical Deep Learning

https://arxiv.org/abs/2402.15332

Is a car that wins a Formula 1 race the best choice for your morning commute? Probably not. In this sponsored deep dive with Prolific, we explore why the same logic applies to Artificial Intelligence. While models are currently shattering records on technical exams, they often fail the most important test of all: **the human experience.**

Why High Benchmark Scores Don’t Mean Better AI

Joining us are **Andrew Gordon** (Staff Researcher in Behavioral Science) and **Nora Petrova** (AI Researcher) from **Prolific**. They reveal the hidden flaws in how we currently rank AI and introduce a more rigorous, "humane" way to measure whether these models are actually helpful, safe, and relatable for real people.

---

Key Insights in This Episode:

* *The F1 Car Analogy:* Andrew explains why a model that excels at the "Humanities Last Exam" might be a nightmare for daily use. Technical benchmarks often ignore the nuances of human communication and adaptability.

* *The "Wild West" of AI Safety:* As users turn to AI for sensitive topics like mental health, Nora highlights the alarming lack of oversight and the "thin veneer" of safety training—citing recent controversial incidents like Grok-3’s "Mecha Hitler."

* *Fixing the "Leaderboard Illusion":* The team critiques current popular rankings like Chatbot Arena, discussing how anonymous, unstratified voting can lead to biased results and how companies can "game" the system.

* *The Xbox Secret to AI Ranking:* Discover how Prolific uses *TrueSkill*—the same algorithm Microsoft developed for Xbox Live matchmaking—to create a fairer, more statistically sound leaderboard for LLMs.

* *The Personality Gap:* Early data from the **Humane Leaderboard** suggests that while AI is getting smarter, it is actually performing *worse* on metrics like personality, culture, and "sycophancy" (the tendency for models to become annoying "people-pleasers").

---

About the HUMAINE Leaderboard

Moving beyond simple "A vs. B" testing, the researchers discuss their new framework that samples participants based on *census data* (Age, Ethnicity, Political Alignment). By using a representative sample of the general public rather than just tech enthusiasts, they are building a standard that reflects the values of the real world.

*Are we building models for benchmarks, or are we building them for humans? It’s time to change the scoreboard.*

Rescript link:

https://app.rescript.info/public/share/IDqwjY9Q43S22qSgL5EkWGFymJwZ3SVxvrfpgHZLXQc

---

TIMESTAMPS:

00:00:00 Introduction & The Benchmarking Problem

00:01:58 The Fractured State of AI Evaluation

00:03:54 AI Safety & Interpretability

00:05:45 Bias in Chatbot Arena

00:06:45 Prolific&apos;s Three Pillars Approach

00:09:01 TrueSkill Ranking & Efficient Sampling

00:12:04 Census-Based Representative Sampling

00:13:00 Key Findings: Culture, Personality & Sycophancy

---

REFERENCES:

Paper:

[00:00:15] MMLU

https://arxiv.org/abs/2009.03300

[00:05:10] Constitutional AI

https://arxiv.org/abs/2212.08073

[00:06:45] The Leaderboard Illusion

https://arxiv.org/abs/2504.20879

[00:09:41] HUMAINE Framework Paper

https://huggingface.co/blog/ProlificAI/humaine-framework

Company:

[00:00:30] Prolific

https://www.prolific.com

[00:01:45] Chatbot Arena

https://lmarena.ai/

Person:

[00:00:35] Andrew Gordon

https://www.linkedin.com/in/andrew-gordon-03879919a/

[00:00:45] Nora Petrova

https://www.linkedin.com/in/nora-petrova/

Event:

Algorithm:

[00:09:01] Microsoft TrueSkill

https://www.microsoft.com/en-us/research/project/trueskill-ranking-system/

Leaderboard:

[00:09:21] Prolific HUMAINE Leaderboard

https://www.prolific.com/humaine

[00:09:31] HUMAINE HuggingFace Space

https://huggingface.co/spaces/ProlificAI/humaine-leaderboard

[00:10:21] Prolific AI Leaderboard Portal

https://www.prolific.com/leaderboard

Dataset:

[00:09:51] Prolific Social Reasoning RLHF Dataset

https://huggingface.co/datasets/ProlificAI/social-reasoning-rlhf

Organization:

[00:10:31] MLCommons

https://mlcommons.org/

What if everything we think we know about AI understanding is wrong? Is compression the key to intelligence? Or is there something more—a leap from memorization to true abstraction?

In this fascinating conversation, we sit down with **Professor Yi Ma**—world-renowned expert in deep learning, IEEE/ACM Fellow, and author of the groundbreaking new book *Learning Deep Representations of Data Distributions*. Professor Ma challenges our assumptions about what large language models actually do, reveals why 3D reconstruction isn't the same as understanding, and presents a unified mathematical theory of intelligence built on just two principles: **parsimony** and **self-consistency**.

**SPONSOR MESSAGES START**

—

Prolific - Quality data. From real people. For faster breakthroughs.

https://www.prolific.com/?utm_source=mlst

—

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

—

**END**

Key Insights:

**LLMs Don't Understand—They Memorize**

Language models process text (*already* compressed human knowledge) using the same mechanism we use to learn from raw data.

**The Illusion of 3D Vision**

Sora and NeRFs etc that can reconstruct 3D scenes still fail miserably at basic spatial reasoning

**"All Roads Lead to Rome"**

Why adding noise is *necessary* for discovering structure.

**Why Gradient Descent Actually Works**

Natural optimization landscapes are surprisingly smooth—a "blessing of dimensionality"

**Transformers from First Principles**

Transformer architectures can be mathematically derived from compression principles

—

INTERACTIVE AI TRANSCRIPT PLAYER w/REFS (ReScript):

https://app.rescript.info/public/share/Z-dMPiUhXaeMEcdeU6Bz84GOVsvdcfxU_8Ptu6CTKMQ

About Professor Yi Ma

Yi Ma is the inaugural director of the School of Computing and Data Science at Hong Kong University and a visiting professor at UC Berkeley.

https://people.eecs.berkeley.edu/~yima/

https://scholar.google.com/citations?user=XqLiBQMAAAAJ&hl=en

https://x.com/YiMaTweets

**Slides from this conversation:**

https://www.dropbox.com/scl/fi/sbhbyievw7idup8j06mlr/slides.pdf?rlkey=7ptovemezo8bj8tkhfi393fh9&dl=0

**Related Talks by Professor Ma:**

- Pursuing the Nature of Intelligence (ICLR): https://www.youtube.com/watch?v=LT-F0xSNSjo

- Earlier talk at Berkeley: https://www.youtube.com/watch?v=TihaCUjyRLM

TIMESTAMPS:

00:00:00 Introduction

00:02:08 The First Principles Book & Research Vision

00:05:21 Two Pillars: Parsimony & Consistency

00:09:50 Evolution vs. Learning: The Compression Mechanism

00:14:36 LLMs: Memorization Masquerading as Understanding

00:19:55 The Leap to Abstraction: Empirical vs. Scientific

00:27:30 Platonism, Deduction & The ARC Challenge

00:35:57 Specialization & The Cybernetic Legacy

00:41:23 Deriving Maximum Rate Reduction

00:48:21 The Illusion of 3D Understanding: Sora & NeRF

00:54:26 All Roads Lead to Rome: The Role of Noise

00:59:56 All Roads Lead to Rome: The Role of Noise

01:00:14 Benign Non-Convexity: Why Optimization Works

01:06:35 Double Descent & The Myth of Overfitting

01:14:26 Self-Consistency: Closed-Loop Learning

01:21:03 Deriving Transformers from First Principles

01:30:11 Verification & The Kevin Murphy Question

01:34:11 CRATE vs. ViT: White-Box AI & Conclusion

REFERENCES:

Book:

[00:03:04] Learning Deep Representations of Data Distributions

https://ma-lab-berkeley.github.io/deep-representation-learning-book/

[00:18:38] A Brief History of Intelligence

https://www.amazon.co.uk/BRIEF-HISTORY-INTELLIGEN-HB-Evolution/dp/0008560099

[00:38:14] Cybernetics

https://mitpress.mit.edu/9780262730099/cybernetics/

Book (Yi Ma):

[00:03:14] 3-D Vision book

https://link.springer.com/book/10.1007/978-0-387-21779-6

<TRUNC> refs on ReScript link/YT

Pedro Domingos, author of the bestselling book "The Master Algorithm," introduces his latest work: Tensor Logic - a new programming language he believes could become the fundamental language for artificial intelligence.

Think of it like this: Physics found its language in calculus. Circuit design found its language in Boolean logic. Pedro argues that AI has been missing its language - until now.

**SPONSOR MESSAGES START**

—

Build your ideas with AI Studio from Google - http://ai.studio/build

—

Prolific - Quality data. From real people. For faster breakthroughs.

https://www.prolific.com/?utm_source=mlst

—

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

—

**END**

Current AI is split between two worlds that don't play well together:

Deep Learning (neural networks, transformers, ChatGPT) - great at learning from data, terrible at logical reasoning

Symbolic AI (logic programming, expert systems) - great at logical reasoning, terrible at learning from messy real-world data

Tensor Logic unifies both. It's a single language where you can:

Write logical rules that the system can actually learn and modify

Do transparent, verifiable reasoning (no hallucinations)

Mix "fuzzy" analogical thinking with rock-solid deduction

INTERACTIVE TRANSCRIPT:

https://app.rescript.info/public/share/NP4vZQ-GTETeN_roB2vg64vbEcN7isjJtz4C86WSOhw

TOC:

00:00:00 - Introduction

00:04:41 - What is Tensor Logic?

00:09:59 - Tensor Logic vs PyTorch & Einsum

00:17:50 - The Master Algorithm Connection

00:20:41 - Predicate Invention & Learning New Concepts

00:31:22 - Symmetries in AI & Physics

00:35:30 - Computational Reducibility & The Universe

00:43:34 - Technical Details: RNN Implementation

00:45:35 - Turing Completeness Debate

00:56:45 - Transformers vs Turing Machines

01:02:32 - Reasoning in Embedding Space

01:11:46 - Solving Hallucination with Deductive Modes

01:16:17 - Adoption Strategy & Migration Path

01:21:50 - AI Education & Abstraction

01:24:50 - The Trillion-Dollar Waste

REFS

Tensor Logic: The Language of AI [Pedro Domingos]

https://arxiv.org/abs/2510.12269

The Master Algorithm [Pedro Domingos]

https://www.amazon.co.uk/Master-Algorithm-Ultimate-Learning-Machine/dp/0241004543

Einsum is All you Need (TIM ROCKTÄSCHEL)

https://rockt.ai/2018/04/30/einsum

https://www.youtube.com/watch?v=6DrCq8Ry2cw

Autoregressive Large Language Models are Computationally Universal (Dale Schuurmans et al - GDM)

https://arxiv.org/abs/2410.03170

Memory Augmented Large Language Models are Computationally Universal [Dale Schuurmans]

https://arxiv.org/pdf/2301.04589

On the computational power of NNs [95/Siegelmann]

https://binds.cs.umass.edu/papers/1995_Siegelmann_JComSysSci.pdf

Sebastian Bubeck

https://www.reddit.com/r/OpenAI/comments/1oacp38/openai_researcher_sebastian_bubeck_falsely_claims/

I am a strange loop - Hofstadter

https://www.amazon.co.uk/Am-Strange-Loop-Douglas-Hofstadter/dp/0465030793

Stephen Wolfram

https://www.youtube.com/watch?v=dkpDjd2nHgo

The Complex World: An Introduction to the Foundations of Complexity Science [David C. Krakauer]

https://www.amazon.co.uk/Complex-World-Introduction-Foundations-Complexity/dp/1947864629

Geometric Deep Learning

https://www.youtube.com/watch?v=bIZB1hIJ4u8

Andrew Wilson (NYU)

https://www.youtube.com/watch?v=M-jTeBCEGHc

Yi Ma

https://www.patreon.com/posts/yi-ma-scientific-141953348

Roger Penrose - road to reality

https://www.amazon.co.uk/Road-Reality-Complete-Guide-Universe/dp/0099440687

Artificial Intelligence: A Modern Approach [Russel and Norvig]

https://www.amazon.co.uk/Artificial-Intelligence-Modern-Approach-Global/dp/1292153962

The Transformer architecture (which powers ChatGPT and nearly all modern AI) might be trapping the industry in a localized rut, preventing us from finding true intelligent reasoning, according to the person who co-invented it. Llion Jones and Luke Darlow, key figures at the research lab Sakana AI, join the show to make this provocative argument, and also introduce new research which might lead the way forwards.

**SPONSOR MESSAGES START**

—

Build your ideas with AI Studio from Google - http://ai.studio/build

—

Tufa AI Labs is hiring ML Research Engineers https://tufalabs.ai/

—

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

—

**END**

The "Spiral" Problem – Llion uses a striking visual analogy to explain what current AI is missing. If you ask a standard neural network to understand a spiral shape, it solves it by drawing tiny straight lines that just happen to look like a spiral. It "fakes" the shape without understanding the concept of spiraling.

Introducing the Continuous Thought Machine (CTM) Luke Darlow deep dives into their solution: a biology-inspired model that fundamentally changes how AI processes information.

The Maze Analogy: Luke explains that standard AI tries to solve a maze by staring at the whole image and guessing the entire path instantly. Their new machine "walks" through the maze step-by-step.

Thinking Time: This allows the AI to "ponder." If a problem is hard, the model can naturally spend more time thinking about it before answering, effectively allowing it to correct its own mistakes and backtrack—something current Language Models struggle to do genuinely.

https://sakana.ai/

https://x.com/YesThisIsLion

https://x.com/LearningLukeD

TRANSCRIPT:

https://app.rescript.info/public/share/crjzQ-Jo2FQsJc97xsBdfzfOIeMONpg0TFBuCgV2Fu8

TOC:

00:00:00 - Stepping Back from Transformers

00:00:43 - Introduction to Continuous Thought Machines (CTM)

00:01:09 - The Changing Atmosphere of AI Research

00:04:13 - Sakana’s Philosophy: Research Freedom

00:07:45 - The Local Minimum of Large Language Models

00:18:30 - Representation Problems: The Spiral Example

00:29:12 - Technical Deep Dive: CTM Architecture

00:36:00 - Adaptive Computation & Maze Solving

00:47:15 - Model Calibration & Uncertainty

01:00:43 - Sudoku Bench: Measuring True Reasoning

REFS:

Why Greatness Cannot be planned [Kenneth Stanley]

https://www.amazon.co.uk/Why-Greatness-Cannot-Planned-Objective/dp/3319155237

https://www.youtube.com/watch?v=lhYGXYeMq_E

The Hardware Lottery [Sara Hooker]

https://arxiv.org/abs/2009.06489

https://www.youtube.com/watch?v=sQFxbQ7ade0

Continuous Thought Machines [Luke Darlow et al / Sakana]

https://arxiv.org/abs/2505.05522

https://sakana.ai/ctm/

LSTM: The Comeback Story? [Prof. Sepp Hochreiter]

https://www.youtube.com/watch?v=8u2pW2zZLCs

Questioning Representational Optimism in Deep Learning: The Fractured Entangled Representation Hypothesis [Kumar/Stanley]

https://arxiv.org/pdf/2505.11581

A Spline Theory of Deep Networks [Randall Balestriero]

https://proceedings.mlr.press/v80/balestriero18b/balestriero18b.pdf

https://www.youtube.com/watch?v=86ib0sfdFtw

https://www.youtube.com/watch?v=l3O2J3LMxqI

On the Biology of a Large Language Model [Anthropic, Jack Lindsey et al]

https://transformer-circuits.pub/2025/attribution-graphs/biology.html

The ARC Prize 2024 Winning Algorithm [Daniel Franzen and Jan Disselhoff] “The ARChitects”

https://www.youtube.com/watch?v=mTX_sAq--zY

Neural Turing Machine [Graves]

https://arxiv.org/pdf/1410.5401

Adaptive Computation Time for Recurrent Neural Networks [Graves]

https://arxiv.org/abs/1603.08983

Sudoko Bench [Sakana]

https://pub.sakana.ai/sudoku/

Ever wonder where AI models actually get their "intelligence"? We reveal the dirty secret of Silicon Valley: behind every impressive AI system are thousands of real humans providing crucial data, feedback, and expertise.Guest: Phelim Bradley, CEO and Co-founder of ProlificPhelim Bradley runs Prolific, a platform that connects AI companies with verified human experts who help train and evaluate their models. Think of it as a sophisticated marketplace matching the right human expertise to the right AI task - whether that's doctors evaluating medical chatbots or coders reviewing AI-generated software.Prolific: https://prolific.com/?utm_source=mlsthttps://uk.linkedin.com/in/phelim-bradley-84300826The discussion dives into:**The human data pipeline**: How AI companies rely on human intelligence to train, refine, and validate their models - something rarely discussed openly**Quality over quantity**: Why paying humans well and treating them as partners (not commodities) produces better AI training data**The matching challenge**: How Prolific solves the complex problem of finding the right expert for each specific task, similar to matching Uber drivers to riders but with deep expertise requirements**Future of work**: What it means when human expertise becomes an on-demand service, and why this might actually create more opportunities rather than fewer**Geopolitical implications**: Why the centralization of AI development in US tech companies should concern Europe and the UK

"What is life?" - asks Chris Kempes, a professor at the Santa Fe Institute.

Chris explains that scientists are moving beyond a purely Earth-based, biological view and are searching for a universal theory of life that could apply to anything, anywhere in the universe. He proposes that things we don't normally consider "alive"—like human culture, language, or even artificial intelligence; could be seen as life forms existing on different "substrates".

To understand this, Chris presents a fascinating three-level framework:

- Materials: The physical stuff life is made of. He argues this could be incredibly diverse across the universe, and we shouldn't expect alien life to share our biochemistry.

- Constraints: The universal laws of physics (like gravity or diffusion) that all life must obey, regardless of what it's made of. This is where different life forms start to look more similar.

- Principles: At the highest level are abstract principles like evolution and learning. Chris suggests these computational or "optimization" rules are what truly define a living system.

A key idea is "convergence" – using the example of the eye. It's such a complex organ that you'd think it evolved only once. However, eyes evolved many separate times across different species. This is because the physics of light provides a clear "target", and evolution found similar solutions to the problem of seeing, even with different starting materials.

**SPONSOR MESSAGES**

—

Prolific - Quality data. From real people. For faster breakthroughs.

https://www.prolific.com/?utm_source=mlst

—

Check out NotebookLM from Google here - https://notebooklm.google.com/ - it’s really good for doing research directly from authoritative source material, minimising hallucinations.

—

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

—

Prof. Chris Kempes:

https://www.santafe.edu/people/profile/chris-kempes

TRANSCRIPT:

https://app.rescript.info/public/share/Y2cI1i0nX_-iuZitvlguHvaVLQTwPX1Y_E1EHxV0i9I

TOC:

00:00:00 - Introduction to Chris Kempes and the Santa Fe Institute

00:02:28 - The Three Cultures of Science

00:05:08 - What Makes a Good Scientific Theory?

00:06:50 - The Universal Theory of Life

00:09:40 - The Role of Material in Life

00:12:50 - A Hierarchy for Understanding Life

00:13:55 - How Life Diversifies and Converges

00:17:53 - Adaptive Processes and Defining Life

00:19:28 - Functionalism, Memes, and Phylogenies

00:22:58 - Convergence at Multiple Levels

00:25:45 - The Possibility of Simulating Life

00:28:16 - Intelligence, Parasitism, and Spectrums of Life

00:32:39 - Phase Changes in Evolution

00:36:16 - The Separation of Matter and Logic

00:37:21 - Assembly Theory and Quantifying Complexity

REFS:

Developing a predictive science of the biosphere requires the integration of scientific cultures [Kempes et al]

https://www.pnas.org/doi/10.1073/pnas.2209196121

Seeing with an extra sense (“Dangerous prediction”) [Rob Phillips]

https://www.sciencedirect.com/science/article/pii/S0960982224009035

The Multiple Paths to Multiple Life [Christopher P. Kempes & David C. Krakauer]

https://link.springer.com/article/10.1007/s00239-021-10016-2

The Information Theory of Individuality [David Krakauer et al]

https://arxiv.org/abs/1412.2447

Minds, Brains and Programs [Searle]

https://home.csulb.edu/~cwallis/382/readings/482/searle.minds.brains.programs.bbs.1980.pdf

The error threshold

https://www.sciencedirect.com/science/article/abs/pii/S0168170204003843

Assembly theory and its relationship with computational complexity [Kempes et al]

https://arxiv.org/abs/2406.12176

Blaise Agüera y Arcas explores some mind-bending ideas about what intelligence and life really are—and why they might be more similar than we think (filmed at ALIFE conference, 2025 - https://2025.alife.org/).

Life and intelligence are both fundamentally computational (he says). From the very beginning, living things have been running programs. Your DNA? It's literally a computer program, and the ribosomes in your cells are tiny universal computers building you according to those instructions.

**SPONSOR MESSAGES**

—

Prolific - Quality data. From real people. For faster breakthroughs.

https://www.prolific.com/?utm_source=mlst

—

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy

Oct SF conference - https://dagihouse.com/?utm_source=mlst - Joscha Bach keynoting(!) + OAI, Anthropic, NVDA,++

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

—

Blaise argues that there is more to evolution than random mutations (like most people think). The secret to increasing complexity is *merging* i.e. when different organisms or systems come together and combine their histories and capabilities.

Blaise describes his "BFF" experiment where random computer code spontaneously evolved into self-replicating programs, showing how purpose and complexity can emerge from pure randomness through computational processes.

https://en.wikipedia.org/wiki/Blaise_Ag%C3%BCera_y_Arcas

https://x.com/blaiseaguera?lang=en

TRANSCRIPT:

https://app.rescript.info/public/share/VX7Gktfr3_wIn4Bj7cl9StPBO1MN4R5lcJ11NE99hLg

TOC:

00:00:00 Introduction - New book "What is Intelligence?"

00:01:45 Life as computation - Von Neumann's insights

00:12:00 BFF experiment - How purpose emerges

00:26:00 Symbiogenesis and evolutionary complexity

00:40:00 Functionalism and consciousness

00:49:45 AI as part of collective human intelligence

00:57:00 Comparing AI and human cognition

REFS:

What is intelligence [Blaise Agüera y Arcas]

https://whatisintelligence.antikythera.org/ [Read free online, interactive rich media]

https://mitpress.mit.edu/9780262049955/what-is-intelligence/ [MIT Press]

Large Language Models and Emergence: A Complex Systems Perspective

https://arxiv.org/abs/2506.11135

Our first Noam Chomsky MLST interview

https://www.youtube.com/watch?v=axuGfh4UR9Q

Chance and Necessity [Jacques Monod]

https://monoskop.org/images/9/99/Monod_Jacques_Chance_and_Necessity.pdf

Wonderful Life: The Burgess Shale and the History of Nature [Stephen Jay Gould]

https://www.amazon.co.uk/Wonderful-Life-Burgess-Nature-History/dp/0099273454

The major evolutionary transitions [E Szathmáry, J M Smith]

https://wiki.santafe.edu/images/0/0e/Szathmary.MaynardSmith_1995_Nature.pdf

Don't Sleep, There Are Snakes: Life and Language in the Amazonian Jungle [Dan Everett]

https://www.amazon.com/Dont-Sleep-There-Are-Snakes/dp/0307386120

The Nature of Technology: What It Is and How It Evolves [W. Brian Arthur]

https://www.amazon.com/Nature-Technology-What-How-Evolves-ebook/dp/B002RI9W16/

The MANIAC [Benjamin Labatut]

https://www.amazon.com/MANIAC-Benjam%C3%ADn-Labatut/dp/1782279814

When We Cease to Understand the World [Benjamin Labatut]

https://www.amazon.com/When-We-Cease-Understand-World/dp/1681375664/

The Boys in the Boat [Dan Brown]

https://www.amazon.com/Boys-Boat-Americans-Berlin-Olympics/dp/0143125478

[Petter Johansson] (Split brain)

https://www.lucs.lu.se/fileadmin/user_upload/lucs/2011/01/Johansson-et-al.-2006-How-Something-Can-Be-Said-About-Telling-More-Than-We-Can-Know.pdf

If Anyone Builds It, Everyone Dies [Eliezer Yudkowsky, Nate Soares]

https://www.amazon.com/Anyone-Builds-Everyone-Dies-Superhuman/dp/0316595640

The science of cycology

https://link.springer.com/content/pdf/10.3758/bf03195929.pdf

<trunc, see YT desc for more>

We sat down with Sara Saab (VP of Product at Prolific) and Enzo Blindow (VP of Data and AI at Prolific) to explore the critical role of human evaluation in AI development and the challenges of aligning AI systems with human values. Prolific is a human annotation and orchestration platform for AI used by many of the major AI labs. This is a sponsored show in partnership with Prolific.

**SPONSOR MESSAGES**

—

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy

Oct SF conference - https://dagihouse.com/?utm_source=mlst - Joscha Bach keynoting(!) + OAI, Anthropic, NVDA,++

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

—

While technologists want to remove humans from the loop for speed and efficiency, these non-deterministic AI systems actually require more human oversight than ever before. Prolific's approach is to put "well-treated, verified, diversely demographic humans behind an API" - making human feedback as accessible as any other infrastructure service.

When AI models like Grok 4 achieve top scores on technical benchmarks but feel awkward or problematic to use in practice, it exposes the limitations of our current evaluation methods. The guests argue that optimizing for benchmarks may actually weaken model performance in other crucial areas, like cultural sensitivity or natural conversation.

We also discuss Anthropic's research showing that frontier AI models, when given goals and access to information, independently arrived at solutions involving blackmail - without any prompting toward unethical behavior. Even more concerning, the more sophisticated the model, the more susceptible it was to this "agentic misalignment."

Enzo and Sarah present Prolific's "Humane" leaderboard as an alternative to existing benchmarking systems. By stratifying evaluations across diverse demographic groups, they reveal that different populations have vastly different experiences with the same AI models.

Looking ahead, the guests imagine a world where humans take on coaching and teaching roles for AI systems - similar to how we might correct a child or review code. This also raises important questions about working conditions and the evolution of labor in an AI-augmented world. Rather than replacing humans entirely, we may be moving toward more sophisticated forms of human-AI collaboration.

As AI tech becomes more powerful and general-purpose, the quality of human evaluation becomes more critical, not less. We need more representative evaluation frameworks that capture the messy reality of human values and cultural diversity.

Visit Prolific:

https://www.prolific.com/

Sara Saab (VP Product):

https://uk.linkedin.com/in/sarasaab

Enzo Blindow (VP Data & AI):

https://uk.linkedin.com/in/enzoblindow

TRANSCRIPT:

https://app.rescript.info/public/share/xZ31-0kJJ_xp4zFSC-bunC8-hJNkHpbm7Lg88RFcuLE

TOC:

[00:00:00] Intro & Background

[00:03:16] Human-in-the-Loop Challenges

[00:17:19] Can AIs Understand?

[00:32:02] Benchmarking & Vibes

[00:51:00] Agentic Misalignment Study

[01:03:00] Data Quality vs Quantity

[01:16:00] Future of AI Oversight

REFS:

Anthropic Agentic Misalignment

https://www.anthropic.com/research/agentic-misalignment

Value Compass

https://arxiv.org/pdf/2409.09586

Reasoning Models Don’t Always Say What They Think (Anthropic)

https://www.anthropic.com/research/reasoning-models-dont-say-think

https://assets.anthropic.com/m/71876fabef0f0ed4/original/reasoning_models_paper.pdf

Apollo research - science of evals blog post

https://www.apolloresearch.ai/blog/we-need-a-science-of-evals

Leaderboard Illusion

https://www.youtube.com/watch?v=9W_OhS38rIE MLST video

The Leaderboard Illusion [2025]

Shivalika Singh et al

https://arxiv.org/abs/2504.20879

(Truncated, full list on YT)

Dr. Ilia Shumailov - Former DeepMind AI Security Researcher, now building security tools for AI agents

Ever wondered what happens when AI agents start talking to each other—or worse, when they start breaking things? Ilia Shumailov spent years at DeepMind thinking about exactly these problems, and he's here to explain why securing AI is way harder than you think.

**SPONSOR MESSAGES**

—Check out notebooklm for your research project, it's really powerfulhttps://notebooklm.google.com/

—

Take the Prolific human data survey - https://www.prolific.com/humandatasurvey?utm_source=mlst and be the first to see the results and benchmark their practices against the wider community!

—

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy

Oct SF conference - https://dagihouse.com/?utm_source=mlst - Joscha Bach keynoting(!) + OAI, Anthropic, NVDA,++

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

—

We're racing toward a world where AI agents will handle our emails, manage our finances, and interact with sensitive data 24/7. But there is a problem. These agents are nothing like human employees. They never sleep, they can touch every endpoint in your system simultaneously, and they can generate sophisticated hacking tools in seconds. Traditional security measures designed for humans simply won't work.

Dr. Ilia Shumailov

https://x.com/iliaishacked

https://iliaishacked.github.io/

https://sequrity.ai/

TRANSCRIPT:

https://app.rescript.info/public/share/dVGsk8dz9_V0J7xMlwguByBq1HXRD6i4uC5z5r7EVGM

TOC:

00:00:00 - Introduction & Trusted Third Parties via ML

00:03:45 - Background & Career Journey

00:06:42 - Safety vs Security Distinction

00:09:45 - Prompt Injection & Model Capability

00:13:00 - Agents as Worst-Case Adversaries

00:15:45 - Personal AI & CAML System Defense

00:19:30 - Agents vs Humans: Threat Modeling

00:22:30 - Calculator Analogy & Agent Behavior

00:25:00 - IMO Math Solutions & Agent Thinking

00:28:15 - Diffusion of Responsibility & Insider Threats

00:31:00 - Open Source Security Concerns

00:34:45 - Supply Chain Attacks & Trust Issues

00:39:45 - Architectural Backdoors

00:44:00 - Academic Incentives & Defense Work

00:48:30 - Semantic Censorship & Halting Problem

00:52:00 - Model Collapse: Theory & Criticism

00:59:30 - Career Advice & Ross Anderson Tribute

REFS:

Lessons from Defending Gemini Against Indirect Prompt Injections

https://arxiv.org/abs/2505.14534

Defeating Prompt Injections by Design.

Debenedetti, E., Shumailov, I., Fan, T., Hayes, J., Carlini, N., Fabian, D., Kern, C., Shi, C., Terzis, A., & Tramèr, F.

https://arxiv.org/pdf/2503.18813

Agentic Misalignment: How LLMs could be insider threats

https://www.anthropic.com/research/agentic-misalignment

STOP ANTHROPOMORPHIZING INTERMEDIATE TOKENS AS REASONING/THINKING TRACES!

Subbarao Kambhampati et al

https://arxiv.org/pdf/2504.09762

Meiklejohn, S., Blauzvern, H., Maruseac, M., Schrock, S., Simon, L., & Shumailov, I. (2025).

Machine learning models have a supply chain problem.

https://arxiv.org/abs/2505.22778

Gao, Y., Shumailov, I., & Fawaz, K. (2025).

Supply-chain attacks in machine learning frameworks.

https://openreview.net/pdf?id=EH5PZW6aCr

Apache Log4j Vulnerability Guidance

https://www.cisa.gov/news-events/news/apache-log4j-vulnerability-guidance

Bober-Irizar, M., Shumailov, I., Zhao, Y., Mullins, R., & Papernot, N. (2022).

Architectural backdoors in neural networks.

https://arxiv.org/pdf/2206.07840

Position: Fundamental Limitations of LLM Censorship Necessitate New Approaches

David Glukhov, Ilia Shumailov, ...

https://proceedings.mlr.press/v235/glukhov24a.html

AlphaEvolve MLST interview [Matej Balog, Alexander Novikov]

https://www.youtube.com/watch?v=vC9nAosXrJw

We need AI systems to synthesise new knowledge, not just compress the data they see. Jeremy Berman, is a research scientist at Reflection AI and recent winner of the ARC-AGI v2 public leaderboard.**SPONSOR MESSAGES**—Take the Prolific human data survey - https://www.prolific.com/humandatasurvey?utm_source=mlst and be the first to see the results and benchmark their practices against the wider community!—cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economyOct SF conference - https://dagihouse.com/?utm_source=mlst - Joscha Bach keynoting(!) + OAI, Anthropic, NVDA,++Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlstSubmit investment deck: https://cyber.fund/contact?utm_source=mlst— Imagine trying to teach an AI to think like a human i.e. solving puzzles that are easy for us but stump even the smartest models. Jeremy's evolutionary approach—evolving natural language descriptions instead of python code like his last version—landed him at the top with about 30% accuracy on the ARCv2.We discuss why current AIs are like "stochastic parrots" that memorize but struggle to truly reason or innovate as well as big ideas like building "knowledge trees" for real understanding, the limits of neural networks versus symbolic systems, and whether we can train models to synthesize new ideas without forgetting everything else. Jeremy Berman:https://x.com/jerber888TRANSCRIPT:https://app.rescript.info/public/share/qvCioZeZJ4Q_NlR66m-hNUZnh-qWlUJcS15Wc2OGwD0TOC:Introduction and Overview [00:00:00]ARC v1 Solution [00:07:20]Evolutionary Python Approach [00:08:00]Trade-offs in Depth vs. Breadth [00:10:33]ARC v2 Improvements [00:11:45]Natural Language Shift [00:12:35]Model Thinking Enhancements [00:13:05]Neural Networks vs. Symbolism Debate [00:14:24]Turing Completeness Discussion [00:15:24]Continual Learning Challenges [00:19:12]Reasoning and Intelligence [00:29:33]Knowledge Trees and Synthesis [00:50:15]Creativity and Invention [00:56:41]Future Directions and Closing [01:02:30]REFS:Jeremy’s 2024 article on winning ARCAGI1-pubhttps://jeremyberman.substack.com/p/how-i-got-a-record-536-on-arc-agiGetting 50% (SoTA) on ARC-AGI with GPT-4o [Greenblatt]https://blog.redwoodresearch.org/p/getting-50-sota-on-arc-agi-with-gpt https://www.youtube.com/watch?v=z9j3wB1RRGA [his MLST interview]A Thousand Brains: A New Theory of Intelligence [Hawkins]https://www.amazon.com/Thousand-Brains-New-Theory-Intelligence/dp/1541675819https://www.youtube.com/watch?v=6VQILbDqaI4 [MLST interview]Francois Chollet + Mike Knoop’s labhttps://ndea.com/On the Measure of Intelligence [Chollet]https://arxiv.org/abs/1911.01547On the Biology of a Large Language Model [Anthropic]https://transformer-circuits.pub/2025/attribution-graphs/biology.html The ARChitects [won 2024 ARC-AGI-1-private]https://www.youtube.com/watch?v=mTX_sAq--zY Connectionism critique 1998 [Fodor/Pylshyn]https://uh.edu/~garson/F&P1.PDF Questioning Representational Optimism in Deep Learning: The Fractured Entangled Representation Hypothesis [Kumar/Stanley]https://arxiv.org/pdf/2505.11581 AlphaEvolve interview (also program synthesis)https://www.youtube.com/watch?v=vC9nAosXrJw ShinkaEvolve: Evolving New Algorithms with LLMs, Orders of Magnitude More Efficiently [Lange et al]https://sakana.ai/shinka-evolve/ Deep learning with Python Rev 3 [Chollet] - READ CHAPTER 19 NOW!https://deeplearningwithpython.io/

Professor Andrew Wilson from NYU explains why many common-sense ideas in artificial intelligence might be wrong. For decades, the rule of thumb in machine learning has been to fear complexity. The thinking goes: if your model has too many parameters (is "too complex") for the amount of data you have, it will "overfit" by essentially memorizing the data instead of learning the underlying patterns. This leads to poor performance on new, unseen data. This is known as the classic "bias-variance trade-off" i.e. a balancing act between a model that's too simple and one that's too complex.

**SPONSOR MESSAGES**

—

Tufa AI Labs is an AI research lab based in Zurich. **They are hiring ML research engineers!**

This is a once in a lifetime opportunity to work with one of the best labs in Europe

Contact Benjamin Crouzier - https://tufalabs.ai/

—

Take the Prolific human data survey - https://www.prolific.com/humandatasurvey?utm_source=mlst and be the first to see the results and benchmark their practices against the wider community!

—

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy

Oct SF conference - https://dagihouse.com/?utm_source=mlst - Joscha Bach keynoting(!) + OAI, Anthropic, NVDA,++

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

—

Description Continued:

Professor Wilson challenges this fundamental belief (fearing complexity). He makes a few surprising points:

**Bigger Can Be Better**: massive models don't just get more flexible; they also develop a stronger "simplicity bias". So, if your model is overfitting, the solution might paradoxically be to make it even bigger.

**The "Bias-Variance Trade-off" is a Misnomer**: Wilson claims you don't actually have to trade one for the other. You can have a model that is incredibly expressive and flexible while also being strongly biased toward simple solutions. He points to the "double descent" phenomenon, where performance first gets worse as models get more complex, but then surprisingly starts getting better again.

**Honest Beliefs and Bayesian Thinking**: His core philosophy is that we should build models that honestly represent our beliefs about the world. We believe the world is complex, so our models should be expressive. But we also believe in Occam's razor—that the simplest explanation is often the best. He champions Bayesian methods, which naturally balance these two ideas through a process called marginalization, which he describes as an automatic Occam's razor.

TOC:

[00:00:00] Introduction and Thesis

[00:04:19] Challenging Conventional Wisdom

[00:11:17] The Philosophy of a Scientist-Engineer

[00:16:47] Expressiveness, Overfitting, and Bias

[00:28:15] Understanding, Compression, and Kolmogorov Complexity

[01:05:06] The Surprising Power of Generalization

[01:13:21] The Elegance of Bayesian Inference

[01:33:02] The Geometry of Learning

[01:46:28] Practical Advice and The Future of AI

Prof. Andrew Gordon Wilson:

https://x.com/andrewgwils

https://cims.nyu.edu/~andrewgw/

https://scholar.google.com/citations?user=twWX2LIAAAAJ&hl=en

https://www.youtube.com/watch?v=Aja0kZeWRy4

https://www.youtube.com/watch?v=HEp4TOrkwV4

TRANSCRIPT:

https://app.rescript.info/public/share/H4Io1Y7Rr54MM05FuZgAv4yphoukCfkqokyzSYJwCK8

Hosts:

Dr. Tim Scarfe / Dr. Keith Duggar (MIT Ph.D)

REFS:

Deep Learning is Not So Mysterious or Different [Andrew Gordon Wilson]

https://arxiv.org/abs/2503.02113

Bayesian Deep Learning and a Probabilistic Perspective of Generalization [Andrew Gordon Wilson, Pavel Izmailov]

https://arxiv.org/abs/2002.08791

Compute-Optimal LLMs Provably Generalize Better With Scale [Marc Finzi, Sanyam Kapoor, Diego Granziol, Anming Gu, Christopher De Sa, J. Zico Kolter, Andrew Gordon Wilson]

https://arxiv.org/abs/2504.15208

In this episode, hosts Tim and Keith finally realize their long-held dream of sitting down with their hero, the brilliant neuroscientist Professor Karl Friston. The conversation is a fascinating and mind-bending journey into Professor Friston's life's work, the Free Energy Principle, and what it reveals about life, intelligence, and consciousness itself.

**SPONSORS**

Gemini CLI is an open-source AI agent that brings the power of Gemini directly into your terminal - https://github.com/google-gemini/gemini-cli

---

Take the Prolific human data survey - https://www.prolific.com/humandatasurvey?utm_source=mlst and be the first to see the results and benchmark their practices against the wider community!

---

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy

Oct SF conference - https://dagihouse.com/?utm_source=mlst - Joscha Bach keynoting(!) + OAI, Anthropic, NVDA,++

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

***

They kick things off by looking back on the 20-year journey of the Free Energy Principle. Professor Friston explains it as a fundamental rule for survival: all living things, from a single cell to a human being, are constantly trying to make sense of the world and reduce unpredictability. It’s this drive to minimize surprise that allows things to exist and maintain their structure.

This leads to a bigger question: What does it truly mean to be "intelligent"? The group debates whether intelligence is everywhere, even in a virus or a plant, or if it requires a certain level of complexity.

Professor Friston introduces the idea of different "kinds" of things, suggesting that creatures like us, who can model themselves and think about the future, possess a unique and "strange" kind of agency that sets us apart.

From intelligence, the discussion naturally flows to the even trickier concept of consciousness. Is it the same as intelligence? Professor Friston argues they are different. He explains that consciousness might emerge from deep, layered self-awareness—not just acting, but understanding that you are the one causing your actions and thinking about your place in the world.

They also explore intelligence at different sizes. Is a corporation intelligent? What about the entire planet? Professor Friston suggests there might be a "Goldilocks zone" for intelligence. It doesn't seem to exist at the super-tiny atomic level or at the massive scale of planets and solar systems, but thrives in the complex middle-ground where we live.

Finally, they tackle one of the most pressing topics of our time: Can we build a truly conscious AI? Professor Friston shares his doubts about whether our current computers are capable of a feat like that. He suggests that genuine consciousness might require a different kind of "mortal" computation, where the machine's physical body and its "mind" are inseparable, much like in biological creatures.

TRANSCRIPT:

https://app.rescript.info/public/share/FZkF8BO7HMt9aFfu2_q69WGT_ZbYZ1VVkC6RtU3eeOI

TOC:

00:00:00: Introduction & Retrospective on the Free Energy Principle

00:09:34: Strange Particles, Agency, and Consciousness

00:37:45: The Scale of Intelligence: From Viruses to the Biosphere

01:01:35: Modelling, Boundaries, and Practical Application

01:21:12: Conclusion

We are joined by Cristopher Moore, a professor at the Santa Fe Institute with a diverse background in physics, computer science, and machine learning.

The conversation begins with Cristopher, who calls himself a "frog" explaining that he prefers to dive deep into specific, concrete problems rather than taking a high-level "bird's-eye view".

They explore why current AI models, like transformers, are so surprisingly effective. Cristopher argues it's because the real world isn't random; it's full of rich structures, patterns, and hierarchies that these models can learn to exploit, even if we don't fully understand how.

**SPONSORS**

Take the Prolific human data survey - https://www.prolific.com/humandatasurvey?utm_source=mlst and be the first to see the results and benchmark their practices against the wider community!

---

cyber•Fund https://cyber.fund/?utm_source=mlst is a founder-led investment firm accelerating the cybernetic economy.

Oct SF conference - https://dagihouse.com/?utm_source=mlst - Joscha Bach keynoting(!) + OAI, Anthropic, NVDA,++

Hiring a SF VC Principal: https://talent.cyber.fund/companies/cyber-fund-2/jobs/57674170-ai-investment-principal#content?utm_source=mlst

Submit investment deck: https://cyber.fund/contact?utm_source=mlst

***

Cristopher Moore:

https://sites.santafe.edu/~moore/

TOC:

00:00:00 - Introduction

00:02:05 - Meet Christopher Moore: A Frog in the World of Science

00:05:14 - The Limits of Transformers and Real-World Data

00:11:19 - Intelligence as Creative Problem-Solving

00:23:30 - Grounding, Meaning, and Shared Reality

00:31:09 - The Nature of Creativity and Aesthetics

00:44:31 - Computational Irreducibility and Universality

00:53:06 - Turing Completeness, Recursion, and Intelligence

01:11:26 - The Universe Through a Computational Lens

01:26:45 - Algorithmic Justice and the Need for Transparency

TRANSCRIPT: https://app.rescript.info/public/share/VRe2uQSvKZOm0oIBoDsrNwt46OMCqRnShVnUF3qyoFk

Filmed at DISI (Diverse Intelligences Summer Institute)

https://disi.org/

REFS:

The Nature of computation [Chris Moore]

https://nature-of-computation.org/

Birds and Frogs [Freeman Dyson]

https://www.ams.org/notices/200902/rtx090200212p.pdf

Replica Theory [Parisi et al]

https://arxiv.org/pdf/1409.2722

Janossy pooling [Fabian Fuchs]

https://fabianfuchsml.github.io/equilibriumaggregation/

Cracking the cryptic [YT channel]

https://www.youtube.com/c/CrackingTheCryptic

Sudoko Bench [Sakana]

https://sakana.ai/sudoku-bench/

Fractured entangled representations “phylogenetic locking in comment” [Kumar/Stanley]

https://arxiv.org/pdf/2505.11581 (see our shows on this)

The War Against Cliché: [Martin Amis]

https://www.amazon.com/War-Against-Cliche-Reviews-1971-2000/dp/0375727167

Rule 110 (CA)

https://mathworld.wolfram.com/Rule150.html

Universality in Elementary Cellular Automata [Matt Cooke]

https://wpmedia.wolfram.com/sites/13/2018/02/15-1-1.pdf

Small Semi-Weakly Universal Turing Machines [Damien Woods]

https://tilde.ini.uzh.ch/users/tneary/public_html/WoodsNeary-FI09.pdf

COMPUTING MACHINERY AND INTELLIGENCE [Turing, 1950]

https://courses.cs.umbc.edu/471/papers/turing.pdf

Comment on Space Time as a causal set [Moore, 88]

https://sites.santafe.edu/~moore/comment.pdf

Recursion Theory on the Reals and Continuous-time Computation [Moore, 96]

Dr. Michael Timothy Bennett is a computer scientist who's deeply interested in understanding artificial intelligence, consciousness, and what it means to be alive. He's known for his provocative paper "What the F*** is Artificial Intelligence" which challenges conventional thinking about AI and intelligence.**SPONSOR MESSAGES***Prolific: Quality data. From real people. For faster breakthroughs.https://prolific.com/mlst?utm_campaign=98404559-MLST&utm_source=youtube&utm_medium=podcast&utm_content=mb***Michael takes us on a journey through some of the biggest questions in AI and consciousness. He starts by exploring what intelligence actually is - settling on the idea that it's about "adaptation with limited resources" (a definition from researcher Pei Wang that he particularly likes).The discussion ranges from technical AI concepts to philosophical questions about consciousness, with Michael offering fresh perspectives that challenge Silicon Valley's "just scale it up" approach to AI. He argues that true intelligence isn't just about having more parameters or data - it's about being able to adapt efficiently, like biological systems do.TOC:1. Introduction & Paper Overview [00:01:34]2. Definitions of Intelligence [00:02:54]3. Formal Models (AIXI, Active Inference) [00:07:06]4. Causality, Abstraction & Embodiment [00:10:45]5. Computational Dualism & Mortal Computation [00:25:51]6. Modern AI, AGI Progress & Benchmarks [00:31:30]7. Hybrid AI Approaches [00:35:00]8. Consciousness & The Hard Problem [00:39:35]9. The Diverse Intelligences Summer Institute (DISI) [00:53:20]10. Living Systems & Self-Organization [00:54:17]11. Closing Thoughts [01:04:24]Michaels socials:https://michaeltimothybennett.com/https://x.com/MiTiBennettTranscript:https://app.rescript.info/public/share/4jSKbcM77Sf6Zn-Ms4hda7C4krRrMcQt0qwYqiqPTPIReferences:Bennett, M.T. "What the F*** is Artificial Intelligence"https://arxiv.org/abs/2503.23923Bennett, M.T. "Are Biological Systems More Intelligent Than Artificial Intelligence?" https://arxiv.org/abs/2405.02325Bennett, M.T. PhD Thesis "How To Build Conscious Machines"https://osf.io/preprints/thesiscommons/wehmg_v1Legg, S. & Hutter, M. (2007). "Universal Intelligence: A Definition of Machine Intelligence"Wang, P. "Defining Artificial Intelligence" - on non-axiomatic reasoning systems (NARS)Chollet, F. (2019). "On the Measure of Intelligence" - introduces the ARC benchmark and developer-aware generalizationHutter, M. (2005). "Universal Artificial Intelligence: Sequential Decisions Based on Algorithmic Probability"Chalmers, D. "The Hard Problem of Consciousness"Descartes, R. - Cartesian dualism and the pineal gland theory (historical context)Friston, K. - Free Energy Principle and Active Inference frameworkLevin, M. - Work on collective intelligence, cancer as information isolation, and "mind blindness"Hinton, G. (2022). "The Forward-Forward Algorithm" - introduces mortal computation conceptAlexander Ororbia & Friston - Formal treatment of mortal computationSutton, R. "The Bitter Lesson" - on search and learning in AIPearl, J. "The Book of Why" - causal inference and reasoningAlternative AGI ApproachesWang, P. - NARS (Non-Axiomatic Reasoning System)Goertzel, B. - Hyperon system and modular AGI architecturesBenchmarks & EvaluationHendrycks, D. - Humanities Last Exam benchmark (mentioned re: saturation)Filmed at:Diverse Intelligences Summer Institute (DISI) https://disi.org/

Deep dive with Dan Hendrycks, a leading AI safety researcher and co-author of the "Superintelligence Strategy" paper with former Google CEO Eric Schmidt and Scale AI CEO Alexandr Wang.

*** SPONSOR MESSAGES

Gemini CLI is an open-source AI agent that brings the power of Gemini directly into your terminal - https://github.com/google-gemini/gemini-cli

Prolific: Quality data. From real people. For faster breakthroughs.

https://prolific.com/mlst?utm_campaign=98404559-MLST&utm_source=youtube&utm_medium=podcast&utm_content=script-gen

***

Hendrycks argues that society is making a fundamental mistake in how it views artificial intelligence. We often compare AI to transformative but ultimately manageable technologies like electricity or the internet. He contends a far better and more realistic analogy is nuclear technology. Like nuclear power, AI has the potential for immense good, but it is also a dual-use technology that carries the risk of unprecedented catastrophe.

The Problem with an AI "Manhattan Project":

A popular idea is for the U.S. to launch a "Manhattan Project" for AI—a secret, all-out government race to build a superintelligence before rivals like China. Hendrycks argues this strategy is deeply flawed and dangerous for several reasons:

- It wouldn’t be secret. You cannot hide a massive, heat-generating data center from satellite surveillance.

- It would be destabilizing. A public race would alarm rivals, causing them to start their own desperate, corner-cutting projects, dramatically increasing global risk.

- It’s vulnerable to sabotage. An AI project can be crippled in many ways, from cyberattacks that poison its training data to physical attacks on its power plants. This is what the paper refers to as a "maiming attack."

This vulnerability leads to the paper's central concept: Mutual Assured AI Malfunction (MAIM). This is the AI-era version of the nuclear-era's Mutual Assured Destruction (MAD). In this dynamic, any nation that makes an aggressive, destabilizing bid for a world-dominating AI must expect its rivals to sabotage the project to ensure their own survival.

This deterrence, Hendrycks argues, is already the default reality we live in.

A Better Strategy: The Three Pillars

Instead of a reckless race, the paper proposes a more stable, three-part strategy modeled on Cold War principles:

- Deterrence: Acknowledge the reality of MAIM. The goal should not be to "win" the race to superintelligence, but to deter anyone from starting such a race in the first place through the credible threat of sabotage.

- Nonproliferation: Just as we work to keep fissile materials for nuclear bombs out of the hands of terrorists and rogue states, we must control the key inputs for catastrophic AI. The most critical input is advanced AI chips (GPUs). Hendrycks makes the powerful claim that building cutting-edge GPUs is now more difficult than enriching uranium, making this strategy viable.

- Competitiveness: The race between nations like the U.S. and China should not be about who builds superintelligence first. Instead, it should be about who can best use existing AI to build a stronger economy, a more effective military, and more resilient supply chains (for example, by manufacturing more chips domestically).

Dan says the stakes are high if we fail to manage this transition:

- Erosion of Control

- Intelligence Recursion

- Worthless Labor

Hendrycks maintains that while the risks are existential, the future is not set.

TOC:

1 Measuring the Beast [00:00:00]

2 Defining the Beast [00:11:34]

3 The Core Strategy [00:38:20]

4 Ideological Battlegrounds [00:53:12]

5 Mechanisms of Control [01:34:45]

TRANSCRIPT:

https://app.rescript.info/public/share/cOKcz4pWRPjh7BTIgybd7PUr_vChUaY6VQW64No8XMs

<truncated, see refs and larger description on YT version>

This episode features Shlomi Fuchter and Jack Parker Holder from Google DeepMind, who are unveiling a new AI called Genie 3. The host, Tim Scarfe, describes it as the most mind-blowing technology he has ever seen. We were invited to their offices to conduct the interview (not sponsored).Imagine you could create a video game world just by describing it. That's what Genie 3 does. It's an AI "world model" that learns how the real world works by watching massive amounts of video. Unlike a normal video game engine (like Unreal or the one for Doom) that needs to be programmed manually, Genie generates a realistic, interactive, 3D world from a simple text prompt.**SPONSOR MESSAGES***Prolific: Quality data. From real people. For faster breakthroughs.https://prolific.com/mlst?utm_campaign=98404559-MLST&utm_source=youtube&utm_medium=podcast&utm_content=script-gen***Here’s a breakdown of what makes it so revolutionary:From Text to a Virtual World: You can type "a drone flying by a beautiful lake" or "a ski slope," and Genie 3 creates that world for you in about three seconds. You can then navigate and interact with it in real-time.It's Consistent: The worlds it creates have a reliable memory. If you look away from an object and then look back, it will still be there, just as it was. The guests explain that this consistency isn't explicitly programmed in; it's a surprising, "emergent" capability of the powerful AI model.A Huge Leap Forward: The previous version, Genie 2, was a major step, but it wasn't fast enough for real-time interaction and was much lower resolution. Genie 3 is 720p, interactive, and photorealistic, running smoothly for several minutes at a time.The Killer App - Training Robots: Beyond entertainment, the team sees Genie 3 as a game-changer for training AI. Instead of training a self-driving car or a robot in the real world (which is slow and dangerous), you can create infinite simulations. You can even prompt rare events to happen, like a deer running across the road, to teach an AI how to handle unexpected situations safely.The Future of Entertainment: this could lead to a "YouTube version 2" or a new form of VR, where users can create and explore endless, interconnected worlds together, like the experience machine from philosophy.While the technology is still a research prototype and not yet available to the public, it represents a monumental step towards creating true artificial worlds from the ground up.Jack Parker Holder [Research Scientist at Google DeepMind in the Open-Endedness Team]https://jparkerholder.github.io/Shlomi Fruchter [Research Director, Google DeepMind]https://shlomifruchter.github.io/TOC:[00:00:00] - Introduction: "The Most Mind-Blowing Technology I've Ever Seen"[00:02:30] - The Evolution from Genie 1 to Genie 2[00:04:30] - Enter Genie 3: Photorealistic, Interactive Worlds from Text[00:07:00] - Promptable World Events & Training Self-Driving Cars[00:14:21] - Guest Introductions: Shlomi Fuchter & Jack Parker Holder[00:15:08] - Core Concepts: What is a "World Model"?[00:19:30] - The Challenge of Consistency in a Generated World[00:21:15] - Context: The Neural Network Doom Simulation[00:25:25] - How Do You Measure the Quality of a World Model?[00:28:09] - The Vision: Using Genie to Train Advanced Robots[00:32:21] - Open-Endedness: Human Skill and Prompting Creativity[00:38:15] - The Future: Is This the Next YouTube or VR?[00:42:18] - The Next Step: Multi-Agent Simulations[00:52:51] - Limitations: Thinking, Computation, and the Sim-to-Real Gap[00:58:07] - Conclusion & The Future of Game EnginesREFS:World Models [David Ha, Jürgen Schmidhuber]https://arxiv.org/abs/1803.10122POEThttps://arxiv.org/abs/1901.01753[Akarsh Kumar, Jeff Clune, Joel Lehman, Kenneth O. Stanley]The Fractured Entangled Representation Hypothesishttps://arxiv.org/pdf/2505.11581TRANSCRIPT:https://app.rescript.info/public/share/Zk5tZXk6mb06yYOFh6nSja7Lg6_qZkgkuXQ-kl5AJqM

Prof. David Krakauer, President of the Santa Fe Institute argues that we are fundamentally confusing knowledge with intelligence, especially when it comes to AI.

He defines true intelligence as the ability to do more with less—to solve novel problems with limited information. This is contrasted with current AI models, which he describes as doing less with more; they require astounding amounts of data to perform tasks that don't necessarily demonstrate true understanding or adaptation. He humorously calls this "really shit programming".

David challenges the popular notion of "emergence" in Large Language Models (LLMs). He explains that the tech community's definition—seeing a sudden jump in a model's ability to perform a task like three-digit math—is superficial. True emergence, from a complex systems perspective, involves a fundamental change in the system's internal organization, allowing for a new, simpler, and more powerful level of description. He gives the example of moving from tracking individual water molecules to using the elegant laws of fluid dynamics. For LLMs to be truly emergent, we'd need to see them develop new, efficient internal representations, not just get better at memorizing patterns as they scale.

Drawing on his background in evolutionary theory, David explains that systems like brains, and later, culture, evolved to process information that changes too quickly for genetic evolution to keep up. He calls culture "evolution at light speed" because it allows us to store our accumulated knowledge externally (in books, tools, etc.) and build upon it without corrupting the original.

This leads to his concept of "exbodiment," where we outsource our cognitive load to the world through things like maps, abacuses, or even language itself.

We create these external tools, internalize the skills they teach us, improve them, and create a feedback loop that enhances our collective intelligence.

However, he ends with a warning. While technology has historically complemented our deficient abilities, modern AI presents a new danger. Because we have an evolutionary drive to conserve energy, we will inevitably outsource our thinking to AI if we can. He fears this is already leading to a "diminution and dilution" of human thought and creativity. Just as our muscles atrophy without use, he argues our brains will too, and we risk becoming mentally dependent on these systems.

TOC:

[00:00:00] Intelligence: Doing more with less

[00:02:10] Why brains evolved: The limits of evolution

[00:05:18] Culture as evolution at light speed

[00:08:11] True meaning of emergence: "More is Different"

[00:10:41] Why LLM capabilities are not true emergence

[00:15:10] What real emergence would look like in AI

[00:19:24] Symmetry breaking: Physics vs. Life

[00:23:30] Two types of emergence: Knowledge In vs. Out

[00:26:46] Causality, agency, and coarse-graining

[00:32:24] "Exbodiment": Outsourcing thought to objects

[00:35:05] Collective intelligence & the boundary of the mind

[00:39:45] Mortal vs. Immortal forms of computation

[00:42:13] The risk of AI: Atrophy of human thought

David Krakauer

President and William H. Miller Professor of Complex Systems

https://www.santafe.edu/people/profile/david-krakauer

REFS:

Large Language Models and Emergence: A Complex Systems Perspective

David C. Krakauer, John W. Krakauer, Melanie Mitchell

https://arxiv.org/abs/2506.11135

Filmed at the Diverse Intelligences Summer Institute:

https://disi.org/