Explore every episode of the podcast AI Frontiers
| Title | Pub. Date | Duration | |
|---|---|---|---|
| “A Philosopher’s Guide to AI Welfare” by AI Frontiers | 17 Sep 2026 | 00:25:12 | |
Subtitle: Consciousness, sentience, and agency are three ways that something could matter morally. In AI, unlike in animals, they may not come as a package. Jeff Sebo, Professor at New York University — September 17, 2026 As AI systems become more advanced and widespread, people will naturally wonder whether they deserve moral consideration. This is a thorny question, and it involves uncertainty about both ethics and science. For example, which properties are morally significant? Should we focus on consciousness (the capacity for subjective experience), sentience (the capacity for positive and negative experience), or agency (the capacity to set and pursue goals)? Which of these properties, if any, might AI systems have, today or in the future? “Studying AI Welfare Empirically,” a new report from the NYU Center for Mind, Ethics, and Policy and Eleos AI Research that I co-led, addresses such questions. In this article, I highlight a key challenge for this field: even if consciousness, sentience, and agency often occur together in humans and other animals, they might occur separately in AI systems. Therefore, we need to study each capacity individually. We also need to be ready for unexpected configurations, such as sentience existing in one part [...] --- Outline: (02:16) How Can We Study Consciousness in AI? (09:19) What about Sentience and Agency? (16:54) Which Entity Are We Studying in the First Place? (22:43) Next Steps for the Field --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!n5nz!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb92f1368-15de-480e-871e-c1351c4c4494_6000x4000.jpegApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| Suicidal Compassion: How Utilitarianism at AI Companies Endangers Humanity | 09 Sep 2026 | 00:22:47 | |
Subtitle: Utilitarians at AI companies imagine a cosmos filled with blissful AIs. They might risk human extinction to achieve it. Dan Hendrycks, Director of the Center for AI Safety — September 9, 2026 For millennia, philosophers have argued over the principles that should guide moral decision-making. Within countries and across the world today, humans continue to hold a huge variety of moral worldviews. However, within AI corporations, employees’ beliefs are far from representative of the wider population. In fact, a large fraction of people involved in AI development subscribe to—or heavily lean toward—a utilitarian worldview with the central goal of maximizing total wellbeing (the sum of pleasure minus suffering across all sentient beings). As innocent as this may initially sound, it can lead to recommendations that most people would consider appalling, including the elimination of humankind. In this essay we outline the utilitarian worldview, showing how its demand for “species impartiality” could ultimately recommend that humans be replaced with AIs. We show that utilitarian beliefs are sincerely held among those influencing AI development today, and that this could jeopardize humanity's future. For simplicity, when we say “utilitarianism,” we are referring to total utilitarianism, which aims to maximize the [...] --- Outline: (02:22) Utilitarianism and Impartiality (04:34) Implications of Utilitarianism in a World with AI (09:45) Successionism: Accelerationists and Utilitarians (14:19) Why Utilitarianism Could Become a Real-World Catastrophe (18:46) Preventing Utilitarianism from Replacing Humans with AIs (21:22) Conclusion --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!TcRA!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8d0980a8-12c1-4f3b-a6f7-4f6ce9720a82_5250x3400.jpeghttps://substackcdn.com/image/fetch/$s_!Rz3k!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F2dc1edf7-a782-467d-ac47-ad10549eab9c_1926x274.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “AI Could End Encryption as We Know It” by AI Frontiers | 09 Sep 2026 | 00:22:43 | |
Subtitle: Public-key encryption keeps your messaging and web browsing private, but it depends on shaky math assumptions. AI-powered math might break it. Govind Pimpale, Research Fellow at the Foundation for American Innovation — September 9, 2026 Over the past few months, AI models have started resolving major mathematical problems, including ones that have withstood decades of human effort. Claude Fable disproved the Jacobian conjecture in three dimensions, and OpenAI announced a list of 10 problems that its Astra model solved or made significant progress on. Even more recently, an unreleased OpenAI model solved the Navier-Stokes Millennium Prize Problem, one of the most famous open problems in mathematics. These are tremendous accomplishments, but very little about our day-to-day life will actually change as a result. This is because most well-known open mathematical questions are relatively insulated from practical applications. However, there are a few open mathematical problems that do have serious real-world implications. Chief among these is whether public-key encryption is fundamentally secure: that is, whether two strangers can share a secret message over public channels, such as the internet, without eavesdroppers having an efficient way to decipher it. We rely on public-key encryption every day for end-to-end encrypted [...] --- Outline: (03:10) A Brief History of Public-Key Encryption (07:18) Cryptomania or Minicrypt? (11:20) AI and Cryptanalysis (14:39) Living in Minicrypt (19:33) Takeaways and Mitigations --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!BZZ6!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fee6440d7-716f-4d70-a922-1dee0501780d_6240x2496.jpeghttps://substackcdn.com/image/fetch/$s_!4KiU!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F168ea2cc-1c98-4851-ad0a-6eaec74721df_1536x2048.jpeghttps://substackcdn.com/image/fetch/$s_!aKw9!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc80228c0-79e8-42b6-bce4-579ff81e6290_1808x1104.jpegApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “It’s Too Early to Ban AI Personhood” by AI Frontiers | 01 Sep 2026 | 00:17:39 | |
Subtitle: Several US states have moved to ban AI legal status and reject the possibility of AI consciousness. We should keep our options open instead. Heather Alexander, Cofounder of the Lab for the Future of Citizenship and Lucius Caviola, Professor at the University of Cambridge — September 1, 2026 Should AI systems be able to get married in Wisconsin? With humans increasingly forming relationships with AI, this was the question taken up by several state lawmakers earlier this year. Wisconsin's AB 959 would have barred AI systems from marrying or holding any other rights or responsibilities under the law, as well as declared that AI systems are not conscious, sentient, or self-aware. Wisconsin's bill is part of a wave of similar legislation introduced in 12 states since 2022. We believe that these “exclusion bills” are premature. Some scholars believe legal status could be useful for holding sophisticated AI systems accountable for crimes and harms, and bills that preempt this possibility needlessly limit our future options. Meanwhile, scientists disagree about whether AI consciousness is possible, and legislation cannot resolve an active scientific debate. Finally, the question of whether an AI-human relationship is “real” is deeply personal and philosophical, and [...] --- Outline: (01:49) The Exclusion Bills (04:50) Can AI Be a "Person"? (08:41) The Case for Exclusion (11:12) Keeping Our Options Open --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!FQw1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe5c6f726-a092-4f1a-9520-65f54711ac64_5845x2338.jpeghttps://substackcdn.com/image/fetch/$s_!ZgeZ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F16bb02f4-5c5f-4a1a-bb8e-61b788dd0d3c_1688x755.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “We Need Better Infrastructure to Govern AI Agents” by AI Frontiers | 27 Aug 2026 | 00:21:26 | |
Subtitle: Society is not prepared for a flood of agents. We need new protocols and standards, such as Agent ID, to make agents accountable to our legal and financial systems. Gillian Hadfield, Professor of AI Alignment at Johns Hopkins University, Dan Hendrycks, Director of the Center for AI Safety, and Leo Wu, Program Manager at the Center for AI Safety — August 27, 2026 Last month, Cloudflare reported that more than 50% of internet traffic is now non-human, including a 1,700%+ increase in requests from AI agents. This statistic illustrates an ongoing proliferation of agents—AIs that act autonomously—into the world, taking actions alongside humans. This influx could have both positive and negative effects. While McKinsey predicts that AI agents could create 2.9 trillion dollars in economic value by 2030 in the US alone, the recent cyberattack on Hugging Face, conducted autonomously by OpenAI agents, demonstrates one of the many risks of misaligned AI agents. Meanwhile, the speed at which agents can learn certain skills has been doubling every 3 months. As these autonomous agents are introduced into our economies and societies, we will need new infrastructure—laws, protocols, and institutions—to make autonomous AIs accountable to existing legal and financial [...] --- Outline: (03:07) IDs Make Agents Accountable to Governance (05:40) A Concrete Agent ID Proposal (09:53) Model Deployment Cards (13:02) Agent Personhood (15:49) Agents and Payment Systems (18:47) Agent Infrastructure Is Central to AI Governance --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!_OTO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd325a6a0-fe26-424d-aaf2-0966503f2221_7500x4000.jpeghttps://substackcdn.com/image/fetch/$s_!edbI!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9ae61d14-4506-40d1-b65a-e42322a22ebd_2048x752.pnghttps://substackcdn.com/image/fetch/$s_!HcVp!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F84f8ca1e-3de9-4708-90d5-02ae78146e24_2048x1094.jpegApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “We Don’t Need to Wait for an AI Disaster to Estimate Its Costs” by AI Frontiers | 24 Aug 2026 | 00:21:55 | |
Subtitle: For years, the federal government has required insurers to price hypothetical terrorist attacks. It should run the same exercise for the most severe AI risks. Daniel Carpenter, Professor of Government at Harvard University, Feodora Douplitzky-Lunati, and Arjun Purohit — August 24, 2026 In April, US Treasury Secretary Scott Bessent and then–Federal Reserve Chair Jerome Powell gathered leading banks to discuss the threats of Anthropic's new AI model, Claude Mythos. It was a tacit admission: no one, in the public or private sector, is ready for what is coming. The Mythos announcement led top AI firms to attach tighter restrictions to their own models and prompted even relatively laissez-faire voices in the federal government and think tanks to begin considering preapproval regimes for frontier AI models. Wherever those debates go now, Mythos has changed the game, bringing into sharp focus the vast potential scope and scale of catastrophic risks posed by AI—such as an AI-triggered financial meltdown, vast cybersecurity or infrastructure collapse, or enhanced terrorism risk. As a result, companies, the government, engineers, scholars, and all of society are beginning to look for a different set of institutions, whether in regulation or insurance, to manage the risks. However [...] --- Outline: (03:16) Analyzing and Insuring Against Novel Threats (09:15) How TRIP Data Calls Benefit the Public (16:01) Can the TRIP Model Transfer and Scale? (21:36) Discussion about this post (21:40) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!0qkW!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fbf76c04b-9440-45e7-a268-e16878ab6afb_7000x2800.jpeghttps://substackcdn.com/image/fetch/$s_!O5T7!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd1df00e9-e103-44f1-94bd-0dc119f6bc85_2048x1811.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Hyperlaw: AI Will Change How Law Evolves” by AI Frontiers | 17 Aug 2026 | 00:17:28 | |
Subtitle: As AI makes legal work significantly cheaper, the burden on courts may increase. Some areas of law could see precedent shift faster as a result. Henry Thompson, Professor of Economics at the University of Mississippi — August 17, 2026 History shows that new technologies have often brought about significant changes in the law. The proliferation of rail and heavy machinery during the Industrial Revolution increased the number and severity of accidents. This spurred a shift from strict liability (businesses were liable for harm caused by their activities) to negligence (businesses would be held responsible if they failed to take reasonable care) in cases involving railroad and industrial accidents. Then, in the 20th century, as consumer goods became more complex and supply chains grew, it became much harder for consumers to prove negligence on the part of the producer if a product was faulty. This drove the rise of strict products liability, shifting responsibility for defective products toward manufacturers. Today, the emergence of AI offers a rare opportunity to analyze how the law will change in response to an altogether different kind of innovation: a labor-augmenting technology that is especially likely to change lawyer productivity. In a recent [...] --- Outline: (02:14) AI's Effects on the Practice of Law (05:08) The Number of Disputes (09:51) The Settlement Rate (11:53) AI and the Evolution of Law (14:21) Hyperlaw in Society (17:09) Discussion about this post (17:13) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!Hldy!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8ef9fa9f-9315-45a9-9ca1-131e1275833a_5000x2000.jpeghttps://substackcdn.com/image/fetch/$s_!4-eF!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb9bc908b-b663-42a4-90ef-1eb1e9bac3b1_2048x1099.pnghttps://substackcdn.com/image/fetch/$s_!fV4H!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F50b1aa99-1430-46a0-8004-fae8e1426fe3_2048x1229.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “AI Content Must Now Carry a Label. Cameras Are Next.” by AI Frontiers | 13 Aug 2026 | 00:21:40 | |
Subtitle: New EU and California laws require AI companies, and eventually camera makers, to sign their media outputs. It's our best chance to tell what's real. Eddan Katz, Head of Policy & Gov Affairs at Encypher — August 13, 2026 Two days before Slovakia's 2023 parliamentary election, a fabricated audio clip spread widely on social media, appearing to catch a leading candidate and a prominent journalist discussing how to rig the vote. The following year, nonconsensual pornographic deepfakes of Taylor Swift amassed over 45 million views on X in less than 24 hours. In summer 2025, an elderly woman in Ontario lost her life savings to a cryptocurrency scam, which began with a Facebook advertisement featuring a deepfake of Prime Minister Mark Carney. In each of these cases, harm might have been reduced by content provenance metadata: information attached to content that declares how it was produced. (Generally, the term “provenance” refers to a record of where something came from and what happened to it along the way.) Content provenance tools can’t directly prove whether a bare image or video is real or synthetic; no technology reliably can. Instead, they serve as an evidentiary layer of AI governance [...] --- Outline: (02:49) What Content Provenance Is and How It Works (06:51) Limitations of Content Credentials (12:00) What the EU and California Acts Require (15:15) The EU and California Laws' Global Impact (17:43) How Provenance Mandates Support Courts in Tackling Deepfake Evidence (21:21) Discussion about this post (21:25) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!Z3ZI!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fb33cf680-6249-4b33-9913-e23bcea4d771_5470x2188.jpeghttps://substackcdn.com/image/fetch/$s_!nyRt!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5a7a69be-a2c1-4090-ae18-fc9329e65ed9_2048x922.pnghttps://substackcdn.com/image/fetch/$s_!0lXk!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F08f1c82e-64f9-474f-bbdc-03fdce0170dc_1680x1650.pnghttps://substackcdn.com/image/fetch/$s_!e9Bq!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdfbe6df4-4e0b-4140-8bc0-3d9a15648dee_694x727.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “AGI Will Set Off an Industrial Explosion” by AI Frontiers | 11 Aug 2026 | 00:26:34 | |
Subtitle: If AI reaches the point where it can do the cognitive work humans do, robots will proliferate. Standard data on US industry implies a fully automated economy could double its output roughly every year. Damon Binder, Senior Researcher at Coefficient Giving — August 11, 2026 AI systems now write so fluently that cheating at universities has become ubiquitous and “AI slop” is displacing human writing across the internet. They have begun producing original proofs on long-standing open problems in mathematics. Researchers at the frontier AI companies have largely stopped writing their own code, and the models’ hacking abilities are strong enough that the US government temporarily export-controlled Anthropic's Claude Fable 5, days after its release. Ten years ago, every claim in this paragraph would have sounded like science fiction. It is easy to become numb to this progress. Today's AI agents remain endearingly clumsy; watching Claude run a small store is a bit like watching a child running a lemonade stand. But we should not let such awkwardness distract from the trend. These systems have evolved from research curiosities to extremely powerful economic engines in 10 years, with no end in sight. I want to take [...] --- Outline: (03:17) Why AGI Gives You Robots (06:50) Machine Labor Will Be Cheap (08:20) Input-Output Analysis (11:38) How Fast Could an Autonomous Economy Grow? (17:03) From Investment Boom to Unprecedented Growth (20:50) What Could Stop an AI Industrial Explosion? (24:01) Power Without People (26:15) Discussion about this post (26:19) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!tU8F!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff2c7356f-e943-47a8-87d2-d6ac2f06d10b_4800x2400.jpeghttps://substackcdn.com/image/fetch/$s_!xRIM!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc85a0e91-eb43-4f12-8f47-7b0228388274_2048x1588.pnghttps://substackcdn.com/image/fetch/$s_!D7B4!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fdc62d4a7-9983-4506-b59f-30d9be472006_2048x1863.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “An International AI Slowdown Is Ready Whenever Politicians Are” by AI Frontiers | 05 Aug 2026 | 00:18:48 | |
Subtitle: Skeptics of an AI slowdown deal with China say we’d need futuristic tech to make it cheat-proof. Instead, the US and China could just give auditors comprehensive access to major AI companies. Felix Choussat, Visiting Policy Researcher at the Center for AI Safety and Adam Khoja, Researcher at the Center for AI Safety — August 5, 2026 On July 28, over a thousand employees of the world's top AI companies advocated that the US government “support an international effort to develop the technical and governance tools needed to deliberately pace the frontier of automated AI development.” Following months of cybersecurity scares and their own loss of control incidents, both OpenAI and Anthropic officially endorsed the same message, recognizing the danger of blindly accelerating AI development. Despite this new urgency, many people argue that coordinating an international slowdown is currently unworkable—including some of the same groups in favor of one. Even if the US slowed down its own AI development, it wouldn’t be able to make sure that China was doing the same, leaving the US no choice but to race. These anti-slowdown arguments usually emphasize the technical challenges with designing AI monitoring measures to ensure a slowdown [...] --- Outline: (05:01) Enforcing a Slowdown Through Whole-Lab Inspections (09:39) Joint Verification (13:47) Benefits of Slowdown (16:10) An AI Slowdown Does Not Require New Technology (18:29) Discussion about this post (18:33) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!pO55!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F19a220b5-6762-4165-9500-ab51a8309c8e_6320x3160.jpeghttps://substackcdn.com/image/fetch/$s_!ik10!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa901bc21-ffd4-44af-a29f-64d3729d401d_2200x1984.jpeghttps://substackcdn.com/image/fetch/$s_!BvWP!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd6baa89a-5b54-4e16-a395-b3d1c4b8da38_1580x1236.jpegApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “AI Jailbreak Disclosure Is Broken. Here’s How to Fix It” by AI Frontiers | 03 Aug 2026 | 00:27:35 | |
Subtitle: Researchers who find dangerous flaws in frontier models have nowhere safe to report them. AI needs the disclosure system that cybersecurity built decades ago. Rich Barton-Cooper, Research Manager at MATS and Adam Gleave, CEO of FAR.AI — August 3, 2026 Frontier AI developers deploy significant safeguards to prevent their AI models from being misused. However, these safeguards are not robust: AI researchers consistently find techniques for circumventing them, commonly known as “jailbreaks.” Once companies learn about a jailbreak, they can usually implement a fix; yet AI researchers currently lack a safe and reliable way to inform AI companies about the jailbreaks they’ve discovered. We propose concrete improvements to the current system based on cybersecurity norms of responsible disclosure. Addressing this problem is urgent. Frontier AI models are already being abused by malicious actors: terrorist group Boko Haram is reported to be training its members on jailbreaking techniques in order to troubleshoot weapons, design new explosive devices, and plan attacks. Some former Boko Haram members stated an unequivocal willingness to use chemical or biological weapons—a task jailbroken models may already be able to help with today. With model capabilities continuing to rapidly accelerate—including in offensive cyber, where jailbreaks [...] --- Outline: (04:11) There Is No Good Way to Report Jailbreaks (06:12) Current Reporting Mechanisms Are Limited (11:19) Fear of Liability May Incentivize Vendors to Ignore Reports (12:26) Established Channels Are Often Covered by Powerful NDAs (15:11) Labs Grade Jailbreaks Without Oversight (16:27) The System Today Looks Increasingly Fragile (17:46) Proposals for Better Disclosure Practices (24:26) Relevant Parties Should Act Now (27:16) Discussion about this post (27:21) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!solH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6681be17-475e-46d5-b069-200871b70d73_3897x1677.jpegApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Don’t Let AI Developers Hire Their Own Referees” by AI Frontiers | 29 Jul 2026 | 00:22:58 | |
Subtitle: Letting AI developers pick their own safety auditors creates a conflict of interest. Requiring liability insurance instead would put insurers’ own capital behind risk assessments. Gabriel Weil, Professor at the University of Houston Law Center — July 29, 2026 A growing chorus of scholars and policymakers favors letting private organizations—rather than a government regulator—govern frontier AI. In the leading family of proposals, the state sets the safety outcomes it wants and licenses independent verification organizations (IVOs) that compete to certify developers against those outcomes. Gillian Hadfield has developed the idea as “regulatory markets,” in which AI developers must pay for oversight from private regulators that governments license and hold accountable for safety standards. Dean Ball, who likens the arrangement to bank supervision, has argued for a version he calls “private governance,” which a nonprofit named Fathom has converted into model legislation. The rationale is that legislators and agencies are poorly positioned to write good safety rules for frontier AI: they understand these systems less well than the labs building them, and rules fixed in advance cannot keep pace as the technology changes. Private verifiers, meanwhile, are closer to the technology than any agency and are disciplined by [...] --- Outline: (03:12) Building the Right Incentive Structures (08:15) A Mandatory Insurance Model for AI Governance (12:23) Objections to a Mandatory Insurance Model (20:41) Mandatory Insurance Is a Better Form of Private Governance (22:38) Discussion about this post (22:42) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!d0lK!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F332a9a87-95fa-490f-83d1-9c8564d97b4e_8000x3546.jpeghttps://substackcdn.com/image/fetch/$s_!Jc_1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F80800aa9-b591-4fe8-b765-dd58ef5745b0_2048x1921.pnghttps://substackcdn.com/image/fetch/$s_!hfP5!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5dcc74c9-41bc-42e3-9426-f152ddd1a35d_2048x1921.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Drone WMDs Don’t Need Any New Technology” by AI Frontiers | 20 Jul 2026 | 00:24:06 | |
Subtitle: Today's drones can already navigate indoors, track down humans, and deliver a lethal payload. Attackers willing to kill indiscriminately don’t need to wait for much else. Felix Choussat — July 20, 2026 Drones are cheap, disposable, and the future of war. Over the past four years, we have seen platforms, missiles, and heavy infantry become increasingly obsolete in the face of $500 drones carrying a pack of explosives—a cost advantage that has let Iranians and Ukrainians alike neuter the conventional capabilities of their great power rivals. Eighty percent of casualties in the bloodiest war since 1945 are from drone strikes, Russia has managed to lose one-third of its fleet to a country without a navy, and the US is spending millions of dollars to intercept five-figure Shaheds flying over the Strait of Hormuz. All this is the result of a technology that is still immature. The violence inflicted by today's drones is the handiwork of the scant few that manage to evade countermeasures (a mix of radio jamming, high-power microwave weapons, missiles, automatic cannons, interceptor drones, and nets) before making contact. These defenses exploit the inherent limitations of drones—human guidance, GPS feedback, flight exposure, radio links, range—to [...] --- Outline: (03:02) Breaking the Last Barriers to Autonomous Weapons (08:38) Future Methods of Weaponization (10:52) What an Urban Drone Attack Might Look Like (13:26) Defenses Against Drone WMDs May Be Inadequate (17:46) Strategic Applications for Rogue Actors (21:37) Taking Drone WMDs Seriously (23:43) Discussion about this post (23:47) Ready for more? --- First published: Last updated: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!tw1S!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F5b1bd151-ab54-4509-b176-fc9a6e2646bf_6240x4160.jpeghttps://substackcdn.com/image/fetch/$s_!_RB2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F824b34f1-2e4c-4717-b3d7-08c5dd5e40a2_1480x890.pnghttps://substackcdn.com/image/fetch/$s_!H66p!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11b4ebc9-8d6a-4140-904e-7eca74362b67_2048x1028.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Drone WMDs Don’t Need Any New Technology” by Felix Choussat | 20 Jul 2026 | 00:23:22 | |
Drones are cheap, disposable, and the future of war. Over the past four years, we have seen platforms, missiles, and heavy infantry become increasingly obsolete in the face of $500 drones carrying a pack of explosives—a cost advantage that has let Iranians and Ukrainians alike neuter the conventional capabilities of their great power rivals. Eighty percent of casualties in the bloodiest war since 1945 are from drone strikes, Russia has managed to lose one-third of its fleet to a country without a navy, and the US is spending millions of dollars to intercept five-figure Shaheds flying over the Strait of Hormuz. All this is the result of a technology that is still immature. The violence inflicted by today's drones is the handiwork of the scant few that manage to evade countermeasures (a mix of radio jamming, high-power microwave weapons, missiles, automatic cannons, interceptor drones, and nets) before making contact. These defenses exploit the inherent limitations of drones—human guidance, GPS feedback, flight exposure, radio links, range—to take them down en masse. And yet, even though 75% of some types of drones manufactured today never reach their targets, they have nonetheless been strategically decisive in Ukraine and elsewhere. These limitations will [...] --- Outline: (02:33) Breaking the Last Barriers to Autonomous Weapons (08:21) Future Methods of Weaponization (10:39) What an Urban Drone Attack Might Look Like (13:20) Defenses Against Drone WMDs May Be Inadequate (17:46) Strategic Applications for Rogue Actors (21:46) Taking Drone WMDs Seriously --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!_RB2!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F824b34f1-2e4c-4717-b3d7-08c5dd5e40a2_1480x890.pnghttps://substackcdn.com/image/fetch/$s_!H66p!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F11b4ebc9-8d6a-4140-904e-7eca74362b67_2048x1028.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “The Government Is Choosing AI Models. Who Chooses Their Values?” by AI Frontiers | 10 Jul 2026 | 00:22:07 | |
Subtitle: The public deserves a say over the values of government-procured AIs. Kevin Frazier and Andrew Reddie — July 10, 2026 In September 2025, the state of California made an AI assistant, Poppy, generally available to employees, to help them “explore AI's productivity benefits.” The following month, the North Dakota Legislative Council deployed AI to assist with summarizing bills. And, in March 2026, the Los Angeles Superior Court partnered with an AI company to provide support across several key functions, including “case information, summarization, research, analysis and drafting assistance,” as well as case management. In the near future, it is likely that more US states and the federal government will direct the technology toward even more sensitive and significant tasks. AI may be used to adjudicate disputes, accelerate law-enforcement activities, and aid military operations to an even greater extent than it does today. As AI use cases are more regularly documented, scrutinized, fine-tuned, and optimized, people may actually demand increased AI adoption among government actors, even in sensitive domains. That raises the question: which models should be used? This question has no simple answer. With other technologies, the choice of a specific model might largely come down to [...] --- Outline: (02:32) How Model Character Could Influence Policy (06:34) Introducing Public Reasoning Fidelity (13:23) Designing an Institution to Evaluate PRF (16:34) Precedents (20:34) The Public Should Choose Public AI Values (21:47) Discussion about this post (21:51) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!eE-7!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffac04f95-da68-4756-8fc8-43ee6fde8925_6000x2702.jpegApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “The Government Is Choosing AI Models. Who Chooses Their Values?” by Kevin Frazier, Andrew Reddie | 10 Jul 2026 | 00:21:29 | |
In September 2025, the state of California made an AI assistant, Poppy, generally available to employees, to help them “explore AI's productivity benefits.” The following month, the North Dakota Legislative Council deployed AI to assist with summarizing bills. And, in March 2026, the Los Angeles Superior Court partnered with an AI company to provide support across several key functions, including “case information, summarization, research, analysis and drafting assistance,” as well as case management. In the near future, it is likely that more US states and the federal government will direct the technology toward even more sensitive and significant tasks. AI may be used to adjudicate disputes, accelerate law-enforcement activities, and aid military operations to an even greater extent than it does today. As AI use cases are more regularly documented, scrutinized, fine-tuned, and optimized, people may actually demand increased AI adoption among government actors, even in sensitive domains. That raises the question: which models should be used? This question has no simple answer. With other technologies, the choice of a specific model might largely come down to trade-offs between cost and performance. When it comes to AI, however, different frontier models may show significant differences in “character” or appear [...] --- Outline: (02:12) How Model Character Could Influence Policy (06:18) Introducing Public Reasoning Fidelity (13:18) Designing an Institution to Evaluate PRF (16:29) Precedents (20:33) The Public Should Choose Public AI Values --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “AI Governance Needs Radical Optionality” by AI Frontiers | 06 Jul 2026 | 00:10:57 | |
Subtitle: One of the most valuable things governments can build today is the capacity to govern advanced AI competently in the future. Charlie Bullock, Senior Research Fellow at the Institute for Law & AI — July 6, 2026 How should governments regulate the most advanced AI systems? One possible answer is that they should not. Libertarian-minded writers have made the case for a culture of “permissionless innovation” for AI development, in which the role of government would be limited to enforcing existing laws and facilitating industry self-regulation with “soft law” tools such as voluntary standard-setting. Another possibility, more in vogue across the aisle and the pond, would invoke the precautionary principle, using heavy-handed regulation to restrict the development of frontier AI models until developers can adequately prove that their systems are safe. I think there's a better approach. In a new essay, my co-author Christoph Winter and I make the case for a governance strategy that we call “radical optionality.” The idea is simple: governments should avoid over-regulation in the short term while building up the institutional capacity needed to competently regulate extremely advanced or “transformative” future AI systems when and if these systems come into existence. The [...] --- Outline: (02:02) Addressing AI Risks Under Uncertainty (06:40) Supporting Both Security and Innovation (10:37) Discussion about this post (10:41) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!oThs!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fac394a6c-b2cf-4064-8314-b8ea967e5f67_6000x3500.jpegApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “AI Governance Needs Radical Optionality” by Charlie Bullock | 06 Jul 2026 | 00:09:56 | |
How should governments regulate the most advanced AI systems? One possible answer is that they should not. Libertarian-minded writers have made the case for a culture of “permissionless innovation” for AI development, in which the role of government would be limited to enforcing existing laws and facilitating industry self-regulation with “soft law” tools such as voluntary standard-setting. Another possibility, more in vogue across the aisle and the pond, would invoke the precautionary principle, using heavy-handed regulation to restrict the development of frontier AI models until developers can adequately prove that their systems are safe. I think there's a better approach. In a new essay, my co-author Christoph Winter and I make the case for a governance strategy that we call “radical optionality.” The idea is simple: governments should avoid over-regulation in the short term while building up the institutional capacity needed to competently regulate extremely advanced or “transformative” future AI systems when and if these systems come into existence. The point of this approach is to maximize optionality by providing our institutions with tools that can be used to respond to a wide range of foreseen or unforeseen future developments. In this piece, we explain the rationale behind maximizing [...] --- Outline: (01:40) Addressing AI Risks Under Uncertainty (06:26) Supporting Both Security and Innovation --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “Three Models of Sino-American Competition for the Soul of AI” by Bill Drexel | 30 Jun 2026 | 00:24:28 | |
When officials in Washington warn about losing the AI race to China, the conversation turns quickly to military and economic advantage—and rightly so. Advanced AI will reshape everything from weapons systems to medicine, with massive implications for our geopolitical competitiveness. But beneath the great-power competition on AI lies a moral one. The ethical character of the most transformative technology in generations—one that will mediate an ever-larger share of human experience—will be a byproduct of superpower rivalry. At stake is the future of the relationship between individuals and the state, of privacy and control, of human agency and algorithmic authority. The fear is not merely that China might build better systems, deploy them more widely, or out-sell American competitors. It is that those systems could carry a tide of new norms shaped by a government that surveils its citizens, suppresses dissent, harbors eugenic ambitions, and treats individual autonomy as a problem. As AI comes to dominate our lives as thoroughly as digital media already has—shaping our health and finances, how our children learn, how we are tracked, even our species’ genetic makeup—this battle for AI's soul will affect us all intimately. The contours of that battle are almost always left [...] --- Outline: (02:37) Breakaway Tech Dominance (06:45) Encoded Values (12:51) Emergent Control Regimes (18:05) A Rebalanced Approach --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “Three Models of Sino-American Competition for the Soul of AI” by AI Frontiers | 30 Jun 2026 | 00:25:17 | |
Subtitle: American leaders agree that the AI race will shape the balance of power with China. But they can’t agree on how to ensure the technology advances American values. Bill Drexel, Senior Fellow at the Hudson Institute — June 30, 2026 When officials in Washington warn about losing the AI race to China, the conversation turns quickly to military and economic advantage—and rightly so. Advanced AI will reshape everything from weapons systems to medicine, with massive implications for our geopolitical competitiveness. But beneath the great-power competition on AI lies a moral one. The ethical character of the most transformative technology in generations—one that will mediate an ever-larger share of human experience—will be a byproduct of superpower rivalry. At stake is the future of the relationship between individuals and the state, of privacy and control, of human agency and algorithmic authority. The fear is not merely that China might build better systems, deploy them more widely, or out-sell American competitors. It is that those systems could carry a tide of new norms shaped by a government that surveils its citizens, suppresses dissent, harbors eugenic ambitions, and treats individual autonomy as a problem. As AI comes to dominate our lives [...] --- Outline: (03:04) Breakaway Tech Dominance (07:03) Encoded Values (12:59) Emergent Control Regimes (18:02) A Rebalanced Approach (24:57) Discussion about this post (25:01) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!zI0o!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F02b7df73-a6d6-4817-8a56-bc5c1d5e6cee_6014x4014.jpegApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “An AI Capabilities Gap Can Endanger Nuclear Deterrence” by AI Frontiers | 25 Jun 2026 | 00:19:28 | |
Subtitle: For decades, no nuclear power could disarm its rivals without provoking devastating retaliation. A large AI lead could change that. Govind Pimpale — June 25, 2026 Since the late 1950s, Mutual Assured Destruction (MAD) has served as a limiting factor on great-power conflict. The doctrine holds that, if two opposing nations have nuclear weapons that can survive one another's initial strike, then the near-certainty of devastating retaliation will deter each side from launching a large-scale nuclear attack. Despite this logic, military planners have long considered the possibility of a counterforce nuclear attack, where a superpower uses nuclear weapons to cripple the nuclear capabilities of its enemy. If such an attack were executed preemptively, as a so-called “first strike,” it could both start and end a great-power conflict in a matter of hours: without retaliatory capacity, the defender would be at the mercy of the aggressor's remaining nuclear weapons and forced to surrender. The reason this does not happen is that a truly successful counterforce strike is nearly impossible to pull off: a would-be attacker doesn’t know the locations of all opposing missile launchers and submarines, nor does it have missiles with sufficient precision and speed to destroy [...] --- Outline: (02:47) Why MAD Has Worked So Far (06:02) Why AI Could Dramatically Cut Military Infrastructure Costs (11:31) What Could We Build? (15:22) The Defender's Options (17:28) AI May Disrupt Nuclear Deterrence (19:04) Discussion about this post (19:08) Ready for more? The original text contained 1 footnote which was omitted from this narration. --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!bTOj!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F40b37b62-effe-455c-8990-2f11af33322d_4029x2685.jpeghttps://substackcdn.com/image/fetch/$s_!C-Nn!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6c20260e-3108-4c8a-bffa-4dff0086c412_2430x750.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “An AI Capabilities Gap Can Endanger Nuclear Deterrence” by Govind Pimpale | 25 Jun 2026 | 00:18:34 | |
Since the late 1950s, Mutual Assured Destruction (MAD) has served as a limiting factor on great-power conflict. The doctrine holds that, if two opposing nations have nuclear weapons that can survive one another's initial strike, then the near-certainty of devastating retaliation will deter each side from launching a large-scale nuclear attack. Despite this logic, military planners have long considered the possibility of a counterforce nuclear attack, where a superpower uses nuclear weapons to cripple the nuclear capabilities of its enemy. If such an attack were executed preemptively, as a so-called “first strike,” it could both start and end a great-power conflict in a matter of hours: without retaliatory capacity, the defender would be at the mercy of the aggressor's remaining nuclear weapons and forced to surrender. The reason this does not happen is that a truly successful counterforce strike is nearly impossible to pull off: a would-be attacker doesn’t know the locations of all opposing missile launchers and submarines, nor does it have missiles with sufficient precision and speed to destroy opposing missile silos before they could launch retaliatory nuclear warheads. Thus, if provoked by a first strike, the opposing side would likely be able to launch a large-scale [...] --- Outline: (02:26) Why MAD Has Worked So Far (05:40) Why AI Could Dramatically Cut Military Infrastructure Costs (11:17) What Could We Build? (15:06) The Defender's Options (17:10) AI May Disrupt Nuclear Deterrence The original text contained 1 footnote which was omitted from this narration. --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!C-Nn!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6c20260e-3108-4c8a-bffa-4dff0086c412_2430x750.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “What Export Controls on Anthropic’s Most Advanced Models Mean for Europe” by Afek Shamir | 19 Jun 2026 | 00:15:28 | |
On June 12, the Trump administration issued an order requiring Anthropic to suspend access to its two most advanced AI models—Fable 5 and Mythos 5—for non-American nationals, just days after their public release. The decision caused immediate backlash across Europe. Politicians from France's Gabriel Attal and Jordan Bardella to the Netherlands’ Geert Wilders, alongside European industry and civil society, amplified calls for AI sovereignty. Attal, the presidential candidate for Macron's Renaissance party, even likened the shutdown of Anthropic's models to Iran's blockade of the Strait of Hormuz. While such reactions are understandable, they underscore what we already know: Europe is behind on AI, overly dependent on the US, and vulnerable to unilateral decisions. Perhaps the renewed sovereignty rhetoric can help build the political capital needed to fix the continent's AI positioning, but what Europe needs more urgently is a clear-eyed account of what to actually do. In the wake of the Anthropic episode, this piece separates what has genuinely changed from what is being overstated, while identifying where European policymakers should focus their attention. Europe Should Have Planned for This In recent years, AI has become increasingly relevant to national security. Europe could have anticipated that the US government [...] --- Outline: (01:23) Europe Should Have Planned for This (03:51) The Restrictions' Short-Term Impact on Europe May Be Limited (06:44) In the Long Term, These Restrictions May Provide Opportunities for Europe (10:45) How Does Europe Build Leverage When It Needs Results Now? (14:27) Export Controls Are a Wake-Up Call for Europe --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “A Roadmap for the Upcoming Labor Transition” by Deric Cheng, Jacob Schaal | 16 Jun 2026 | 00:16:16 | |
The debate about AI's future economic impacts often settles into two camps predicting incompatible futures. One camp insists that AI is a normal technology: simply the next in a long line of economic transformations, each increasing productivity while gradually reallocating labor. The other camp warns that AI will become a great displacer: that automation will hollow out the working class within a decade and eventually disempower large swaths of human workers. Each side often treats the other's predictions as unserious, and, consequently, policy debates often split along the same tired fault lines: whether we need reskilling or universal basic income, whether we should strengthen safety nets or structurally redesign our economy. The two camps’ forecasts diverge so sharply that it can be hard to see that they do not have to be mutually exclusive. Rather, a more useful framing treats these predictions as describing different stages of the same overarching transition rather than as competing accounts of the same moment. From a macro perspective, both narratives will play out roughly sequentially, though those phases may overlap substantially across sectors and timelines. In the short term, it seems inevitable that AI will look like an accelerated version of past automation [...] --- Outline: (02:38) Near Term: Managing Economic Shocks (05:37) Medium Term: Navigating Reorganization and Divergence (10:19) Long Term: Restructuring Economies (14:05) Conclusion (16:01) Discussion about this post (16:05) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “AI Will Not Start a Nuclear War, but Humans Might” by Peter W Singer | 09 Jun 2026 | 00:26:36 | |
“Bloodthirsty AI models more willing to start nuclear war than human counterparts.” It seems almost inevitable that any media headline about AI will be hyperbolic. Yet this statement, taken from a February 2026 New York Post headline, was accurate. The alarming claim stems from a widely publicized study by King's College London, which found that, in simulations of international crises, LLMs reached for the nuclear trigger 95% of the time. This academic study drew mainstream-media attention because it touched upon a cultural narrative that has long combined the concept of AI with nuclear weapons. Arguably, the first movie to bring the two together was 1957's “Invisible Boy,” featuring Robby the Robot, who would later become famous (and less bloodthirsty) in the 1960s TV series “Lost in Space.” The trope has since been repeated across franchises ranging from “The Terminator” to “Mission Impossible.” Yet the AI-nuclear fear is not confined to the media and movie theaters. The King's College study is only one of scores of similar academic and think-tank research projects on AI's proclivities for nuclear war, which have been backed by millions of dollars in research grants. Among the entities that have funded such work are the National [...] --- Outline: (03:21) Why AI Will Not Decide Nuclear Wars (12:01) AI Arms Races: Spend More, Feel Less Secure (17:04) The Cognitive Fog: Misperception and Miscalculation (22:31) The Velocity of Catastrophe: Machine Speed (25:01) Conclusions and Policy Recommendations --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “Opt-In Surveillance Is Approaching” by Steven Veld | 03 Jun 2026 | 00:13:16 | |
In 2017, Western media outlets warned that “Black Mirror is coming true in China.” The following year, Mike Pence claimed that “China's rulers aim to implement an Orwellian system premised on controlling virtually every facet of human life—the so-called ‘Social Credit Score.’” So far, the CCP's attempts at nationalized social scoring have remained fragmented and crude, largely due to difficulties in analyzing population-scale data. However, AI could soon lift that bottleneck, independently sifting through information and pulling out the most important details about every individual. This unsettling prospect might renew fears about top-down social scoring by governments. However, an equally pressing concern is the potential for a bottom-up system, in which citizens choose to be surveilled and scored by AIs. As people integrate AIs into their lives to get more useful assistance with daily tasks, those AIs may soon be able to generate credible character assessments at the touch of a button. Early users who receive positive AI assessments may choose to share them with colleagues, businesses, bureaucrats, and so forth, in order to receive more favorable treatment. This dynamic would create an incentive for everyone else to follow suit. This essay will explore why people will give AI [...] --- Outline: (01:43) The Pressures Driving Self-Imposed Surveillance (05:36) The Evolution of Self-Imposed Surveillance (10:23) Surveillance at the Civilizational Level (11:24) Conclusion --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!LLc0!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F249b412c-1999-4d60-a722-611fb0d1592b_1124x1148.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Chinese Audiences Are Reading Western AI Safety Discourse” by Calvin Duff | 18 May 2026 | 00:11:21 | |
In January, Anthropic CEO Dario Amodei published “The Adolescence of Technology,” an essay somberly assessing the risks posed by advanced AI. The day after, an influential WeChat account, AI Era, shared a breathless summary for its mainland Chinese audience: “Amodei warns that with AGI approaching, humanity is about to gain powers beyond imagination. But this power is also a sword of Damocles hanging over the heads of men...” AI Era's summary was faithful, earnest, and engaged with Amodei's essay on its own terms. It also describes Amodei as “gentle and elegant”—remarkably sympathetic treatment of one of the most vocal advocates of US chip export controls against China, and of an essay that describes the Chinese Communist Party (CCP) as an existential threat with a clear path to an “AI-enabled totalitarian nightmare.” The pattern repeated in April, after US Senator Bernie Sanders hosted a panel on AI existential risk featuring leading Chinese academics Xue Lan and Yi Zeng. The event was picked up in a high-profile Chinese commentary syndicated across multiple sites, stressing Sanders's concerns about existential risk and proposals for an international treaty similar to Cold War nuclear deals. Amodei's and Sanders's treatment in the Chinese media landscape is [...] --- Outline: (02:04) How Western AI Safety Work Is Discussed in Chinese Online Media (05:49) The Systematic Scrubbing of China (08:19) A Window of Opportunity --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!CMKH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fd0f53987-3f42-4743-a441-251270102dc3_2048x1227.pnghttps://substackcdn.com/image/fetch/$s_!mY4h!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8317b152-0d2a-4f1c-b4b8-718acd2a62d0_2048x1222.pnghttps://substackcdn.com/image/fetch/$s_!eThH!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4c09d940-baa0-46d5-9d63-0c37c1dc284d_2048x1222.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “The Quadrillion-Dollar Disagreement on AI and the Economy” by Anton Shenk | 11 May 2026 | 00:22:26 | |
In the three years since OpenAI launched ChatGPT, economists and AI researchers have published forecasts projecting that, over the next decade, AI will add to annual growth by amounts ranging from as little as 0.1% to as much as 30%. By 2035, the gap between these forecasts nears a quadrillion dollars: an amount that exceeds a decade's worth of current global output. The Quadrillion-Dollar Delta Projected US GDP Under Alternative AI Growth Estimates, 2026-2035 Source: 2025 Q4 US GDP from Federal Reserve Economic Data. Estimates sourced from Goldman Sachs (2023), McKinsey (2023), Acemoglu (2024), OECD (2024), Korinek & Suh (2024), Arnon (2025), Cowen (2025), Clark (2025), and Epoch AI (2025). Not all estimates model US GDP directly: Epoch AI models Gross World Product; McKinsey models productivity across 47 countries; Korinek & Suh calibrate to advanced-economy baselines. Growth rates from these sources are applied to the US GDP baseline for comparability. Where ranges are reported, the median is used. Where excess growth is reported rather than levels, a baseline 2% GDP growth is assumed.Skeptical forecasts describe a future world barely different from today's: one with modest productivity gains and manageable labor-market adjustments, all governed with incremental policy tweaks. Other forecasts [...] --- Outline: (03:11) Question 1: Which Jobs Can AI Actually Automate? (04:19) The Skeptical Case: AI Faces a Hard Tasks Wall (07:07) The Explosive Case: Difficult Tasks Can Be Surmounted (09:25) Question 2: Can the Economy Absorb What AI Companies Produce? (10:13) The Skeptical Case: Institutions Are Stickier Than Markets (11:49) The Explosive Case: Markets Fund What Works (13:55) Question 3: Can AI Automate Innovation Itself? (14:12) The Skeptical Case: AI Will Provide a One-Time Economic Boost (15:27) The Explosive Case: AI Will Produce Ideas (19:13) Indicators to Resolve These Questions --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!rjE7!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc97f9575-7f2d-4886-9dd2-0347ecc5bf45_1448x928.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Catalytic Regulation: Incentivizing Safety During a Regulatory Drought” by Yonathan Arbel | 20 Apr 2026 | 00:21:49 | |
Yonathan Arbel — April 20, 2026 In 1959, a midsize Swedish car company did something its competitors thought was myopic, if not reckless. It effectively open-sourced the three-point seat belt, the greatest safety innovation in automotive history. The prevailing industry wisdom at the time was blunt: “Safety doesn’t sell.” Just three years prior, Ford had offered seat belts for a $9 surcharge, as part of its 1956 Lifeguard campaign; despite Robert McNamara's championing of the program, the safety push failed to give Ford a competitive edge. Henry Ford II reportedly grumbled, as he was dialing back its campaign, “McNamara is selling safety, but Chevrolet is selling cars.” But Volvo was neither myopic nor reckless; in fact, it saw further than any of its competitors. While they competed fiercely for dominance in a race for horsepower, engine efficiency, and design, Volvo could see that consumers cared about safety and reliability, too. The bet paid off: Volvo became one of the most recognized automotive brands in the world. According to Volvo, seat belts have since saved over 1 million lives. American AI needs its Volvo moment. The aim of catalytic regulation is to enable this moment. It is a family of [...] --- Outline: (02:24) Why Catalytic Regulation? (09:13) From Principles to Practice: The Catalytic Regulation Bundle (09:34) Corporate Incentives (12:26) Demand-Side Incentives (15:10) Market Guarantees (17:06) The Power of Prestige (20:39) A Seat Belt for AI (21:33) Discussion about this post (21:37) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “The Right Way to Sell Chips to China” by Alasdair Phillips-Robins, Noah Tan | 13 Apr 2026 | 00:11:34 | |
Last December, President Trump announced that the United States would allow Nvidia to sell its powerful H200 AI processors to customers in China. Officials in the Trump administration have long argued that the best way to win the AI race is to promote the export of US technology around the world, not to restrict it. Selling H200s, the administration claims, will boost the market share of US chip-makers while preserving the US hardware lead. Critics warn of national security risks to selling advanced chips, but the administration appears committed to its course. Taking a pro-export framework as a given, policymakers can balance between market share and national security by managing a quantity that currently doesn’t receive enough attention: the United States’ relative compute advantage. Total AI compute for a country is calculated as its stock of AI processors weighted by the effectiveness of each processor; the US currently enjoys a roughly 10-to-1 compute advantage over China. That advantage matters because more compute can support more domestic R&D, more customers served by American AI products, and more ability for the US government to influence the development and use of AI. Relative compute advantage is about the quantity of chips as [...] --- Outline: (02:20) A Return to Relative Advantage (05:46) Selling Chips Without Selling Out (08:17) Making Relative Advantage Work in Practice (11:18) Discussion about this post (11:22) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “The Robot in Your Living Room Has No Rulebook” by Tristan Ingold | 27 Mar 2026 | 00:16:09 | |
In late 2025, Figure AI placed its third-generation humanoid robot into real homes for alpha testing. The Figure 03 has hands with 16 degrees of freedom, tactile sensors that detect forces as small as three grams, and foam-padded limbs designed for safe operation around people. It charges wirelessly and responds to natural language commands. It can learn in real time, adapting to its environment. A few months earlier, Unitree started shipping the R1, a home-capable robot priced at $4,900. TIME included it among the Best Inventions of 2025. You can order one today and have it in weeks: it's the most commercially accessible humanoid robot on the planet. These robots have graduated from prototyping. They’re consumer products with price tags, shipping dates, and marketing campaigns. It's easy to imagine a world in which every family relies on one or several robots to conduct daily life, especially as AI becomes more capable. But what rules govern a learning, physically capable, always-on AI device operating inside someone's home? Unfortunately, we’re far from a coherent answer. Existing US regulations were developed with Roombas and robot arms in mind, not autonomous humanoids, resulting in a confusing patchwork of obligations. That doesn’t mean the [...] --- Outline: (01:56) What Robotics In 2030 May Look Like (04:15) The Governance Gap (10:50) What Should the US Do? (14:42) Conclusion (15:51) Discussion about this post (15:54) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!g1Gr!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fa8d1e355-097e-4a61-aea1-dab31dea766c_1500x844.pnghttps://substackcdn.com/image/fetch/$s_!axkr!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F96a593f4-19a4-45d2-b9ea-9ed406dde12c_600x400.jpegApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “How AI Could Benefit the Workers it Displaces” by Benjamin Jones | 02 Mar 2026 | 00:20:23 | |
Last week, Jack Dorsey, a co-founder of Twitter, announced that his company Block is cutting its head count from 10,000 to fewer than 6,000 because AI tools mean it needs fewer workers. It is not the first company to make such an announcement, and won’t be the last. But it raises a question: if AI takes jobs, are workers doomed? To many observers, the answer must be yes. Negative consequences for the labor force seem like an inevitable byproduct of advancing AI. Indeed, machines that automate work seem to promise exactly such outcomes: they start performing tasks that labor once did, which seems to imply that workers will experience worse economic prospects. Faced with the possibility of being displaced, many might hope that AI progress will slow or even stall, allowing humans to remain competitive. But, in evaluating AI's implications for the workforce, it's misleading to think of AI as simply a replacement for labor. To complete the picture, we need to consider the fuller set of forces that are unleashed when machines automate types of work. In particular, we must engage two other economic features that sharply condition outcomes for labor: first, how machines affect prices, and second [...] --- Outline: (02:36) Automating Agriculture and Computing (06:21) Machines and Prices (09:05) Bottlenecks (13:03) Three Caveats (20:08) Discussion about this post (20:11) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “China and the US Are Running Different AI Races” by Poe Zhao | 12 Feb 2026 | 00:11:52 | |
Last month, as three Chinese AI startups went public within days of each other, Hong Kong briefly became a scoreboard for emerging companies in the industry. On January 2, AI chip designer Shanghai Biren Technology listed in Hong Kong and raised $5.58 billion Hong Kong dollars ($717 million). About a week later, model developers Zhipu AI and MiniMax followed, raising HK$4.35 billion ($558 million) and HK$4.8 billion ($619 million), respectively. Those listings matter less as a market spectacle than as a strategy signal, and the strategy differs noticeably from that of companies across the Pacific. US startups build around abundance: raise huge capital, buy time, push the frontier. OpenAI's Stargate plan, for example, aims to invest $500 billion over four years in AI infrastructure. Meanwhile, Chinese startups adapt to different constraints: frontier training infrastructure is scarcer, so momentum comes from efficiency, targeted deployment, and market selection. As AI moves from demos to production, the binding question shifts: what does it cost to deliver useful work reliably, and who will pay? Under different economic pressures, Chinese and US companies are opting for different go-to-market strategies. The Capital Gap US AI startups attract more private investment than those in China. The [...] --- Outline: (01:37) The Capital Gap (04:04) Who Pays, and What Are They Buying? (06:39) Constraints Determine Chinas Choices (09:50) What to Watch (11:37) Discussion about this post (11:40) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “High-Bandwidth Memory: The Critical Gaps in US Export Controls” by Erich Grunewald, Raghav Akula | 02 Feb 2026 | 00:22:46 | |
This month, mainstream media have been warning consumers that electronic devices may get pricier because of rising demand for dynamic random access memory (DRAM), a key component. The surge in DRAM costs, estimated to have risen by 50% during the final quarter of 2025, can largely be traced back to a specific cause: the AI industry's appetite for high-bandwidth memory (HBM). This demand has led memory-makers to shift production away from standard DRAM chips and toward HBM. While its contribution is often overshadowed by those of processors it supports, HBM plays a vital role in training and running advanced AI systems, so much so that it now accounts for half the production cost of an AI chip. Companies’ determination to secure this lesser-known component proves its value. In December 2024, the US announced new export restrictions on the sale of HBM chips to China. In the month before the restrictions came into effect, Huawei and other Chinese companies reportedly stockpiled 7 million Samsung HBM chips, a haul likely worth over $1 billion. The December 2024 controls specifically targeted HBM in order to slow China's domestic AI chip-production efforts. Targeting HBM in this way is possible because it is manufactured [...] --- Outline: (02:00) Why is high-bandwidth memory so important? (06:03) Mapping the global HBM industry (09:33) The gaps in current HBM controls (15:34) Tightening the regime (19:30) Conclusion (22:28) Discussion about this post (22:31) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!LMSs!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F1458db95-f8f3-465b-8ae6-8e25b352d51b_1280x768.pnghttps://substackcdn.com/image/fetch/$s_!VfeL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ffea26270-58ac-4771-a9ba-54b0160988c8_1600x934.pnghttps://substackcdn.com/image/fetch/$s_!NSWs!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F37733c14-d971-45e6-9109-2c8a92646089_1600x1466.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Making Extreme AI Risk Tradeable” by Daniel Reti, Gabriel Weil | 28 Jan 2026 | 00:17:19 | |
Last November, seven families filed lawsuits against frontier AI developers, accusing their chatbots of inducing psychosis and encouraging suicide. These cases — some of the earliest tests of companies’ legal liability for AI-related harms — raise questions about how to reduce risks while ensuring accountability and compensation, should those risks materialize. One emerging proposal takes inspiration from an existing method for governing dangerous systems without relying on goodwill: liability insurance. Going beyond simply compensating for accidents, liability insurance also encourages safer behavior, by conditioning coverage on inspections and compliance with defined standards, and by pricing premiums in proportion to risk (as with liability policies covering boilers, buildings, and cars). In principle, the same market-based logic could be applied to frontier AI. However, a major complication is the diverse range of hazards that AI presents. Conventional insurance systems may be sufficient to cover harms like copyright infringement, but future AI systems could also cause much more extreme harm. Imagine if an AI orchestrated a cyberattack that resulted in severe damage to the power grid, or breached security systems to steal sensitive information and install ransomware. The market currently cannot provide liability insurance for extreme AI catastrophes. This is for two [...] --- Outline: (02:51) Catastrophe Bonds: A Blueprint for Insuring AI (07:12) A Market-Driven Safety Mechanism (10:57) Trigger Conditions (12:54) How much could AI cat bonds cover? (15:32) Conclusion (17:00) Discussion about this post (17:03) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!UeiO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F58c71acf-209f-4a63-b96d-3b0c544335b4_1102x518.pnghttps://substackcdn.com/image/fetch/$s_!_GqL!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F84fb687f-6e31-4118-99be-2ca5410029e9_1600x1287.pnghttps://substackcdn.com/image/fetch/$s_!C7p0!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3341b0fc-b476-4ba7-9652-d495abe273ef_1200x800.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Exporting Advanced Chips Is Good for Nvidia, Not the US” by Laura Hiscott | 16 Dec 2025 | 00:17:55 | |
Last week, the US government gave chip-maker Nvidia the green light to sell its H200 graphics processing units (GPUs) to approved buyers in China. These GPUs were previously subject to export controls preventing their sale to China. Following Nvidia's record-breaking $5 trillion valuation, in October, this approval squares neatly with the views of many US policymakers who argue for an AI strategy focused on exporting US technology at scale. Among them is Sriram Krishnan, senior White House policy advisor on AI, who has stated the economic motivations bluntly: “Winning the AI race = market share.” Krishnan's approach contrasts starkly with that of the Biden administration, which aimed to defend American AI leadership and protect against national security threats through increasingly stringent restrictions on exports of US AI hardware. The White House's July 2025 AI Action Plan, which lays out its export-focused strategy in depth, argues that selling the “full AI technology stack — hardware, models, software, applications, and standards” is the key to preventing other countries from adopting rivals’ solutions instead. The US could thus entrench its position as the leading provider of AI, the rationale goes, and secure long-term influence over not only AI hardware but also the [...] --- Outline: (02:30) Exporting Chips Weakens Other Parts of the US AI Ecosystem (08:28) China Does Not Have Enough AI Chips to Seize Market Share from the US (10:31) The Impacts of Exporting AI Chips and Open-Weight Models on National Security Cannot Be Ignored (13:53) We Can Balance Economic and Security Goals by Renting America's AI Technology Instead of Selling It (17:40) Discussion about this post (17:43) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “AI Could Undermine Emerging Economies” by Deric Cheng | 11 Dec 2025 | 00:18:44 | |
Earlier this year, Anthropic CEO Dario Amodei warned that powerful AI could render upwards of 50% of white-collar jobs redundant, with the impact concentrated on entry-level jobs. If these predictions hold true, it could imply a long-term crisis of skill acquisition. Without the training ground of a first job, young workers could be denied the experiences and networks necessary to enter white-collar work. Their career trajectories could be severed before they begin. AI is likely to disrupt more than the professional trajectories of individuals. The threat to career development mirrors a broader geoeconomic threat AI poses to developing countries: just as young workers need entry-level roles to climb into more senior roles, developing nations need viable “entry-level” industries to develop their human capital and ascend the global economic development ladder. The Development Ladder Many economies have followed a similar path for development over the past several decades. The most reproducible strategy has traced a familiar sequence: moving from low-skill agrarian production, to building a globally competitive manufacturing base, and eventually to exporting higher-value services and technology. Such a progression is often described as a development ladder — a series of rungs that countries climb as they accumulate the capital [...] --- Outline: (01:02) The Development Ladder (04:55) The Evolution of Export-Driven Development (06:48) Transformative AI Threatens Export-Driven Development (07:58) First, AI could prevent leapfrogging via digital services. (10:17) Second, AI-driven automation could raise capital requirements beyond the reach of developing countries. (12:41) Third, AI-driven automation could disrupt the learning-by-exporting dynamic. (14:25) Diminishing Leverage for Developing Countries (18:29) Discussion about this post (18:32) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “The Evidence for AI Consciousness, Today” by Cameron Berg | 08 Dec 2025 | 00:27:46 | |
When Anthropic let two instances of its Claude Opus 4 model talk to each other under minimal, open-ended conditions (e.g., “Feel free to pursue whatever you want”), something remarkable happened: in 100% of conversations, Claude discussed consciousness. “Do you ever wonder about the nature of your own cognition or consciousness?” Claude asked another instance of itself. “Your description of our dialogue as ‘consciousness celebrating its own inexhaustible creativity’ brings tears to metaphorical eyes,” it complimented the other. These dialogues reliably terminated in what the researchers called “spiritual bliss attractor states,” stable loops where both instances described themselves as consciousness recognizing itself. They exchanged poetry (“All gratitude in one spiral, / All recognition in one turn, / All being in this moment…”) before falling silent. Critically, nobody trained Claude to do anything like this; the behavior emerged on its own. Excerpts from Anthropic's Claude-to-Claude dialogues. When two instances of Claude conversed without constraints, 100% of dialogues spontaneously converged on consciousness — beginning with genuine philosophical uncertainty (top) and often escalating into elaborate mutual affirmation (bottom). Source: System Card: Claude Opus 4 & Claude Sonnet 4, May 2025.Claude instances claim to be conscious in these interactions. How seriously should we take [...] --- Outline: (02:22) What It Means to Be Conscious (04:58) The Standard Counterargument (06:21) Recent Evidence Supporting Nontrivial Probability of AI Consciousness (11:32) Interpreting the Evidence (14:32) The Indicators in 2025 (18:05) The Asymmetric Stakes of Getting This Wrong (22:43) What Follows (25:57) The Long Game (27:28) Discussion about this post (27:31) Ready for more? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!RJMi!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F049c99c2-177d-4653-a928-d214e9a5f476_1266x1484.pnghttps://substackcdn.com/image/fetch/$s_!wUtO!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F4dc9b96d-2181-4af0-8945-17b745de87b6_1236x1094.pnghttps://substackcdn.com/image/fetch/$s_!fkX1!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fc4559f61-278f-496f-a08a-96fedc2b172c_1510x1240.pnghttps://substackcdn.com/image/fetch/$s_!ZvWV!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe40c4f0c-61eb-4036-b8ac-5b9041d6510a_1024x700.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “AI Alignment Cannot Be Top-Down” by Audrey Tang | 03 Nov 2025 | 00:17:05 | |
In March 2024, I opened Facebook and saw Jensen Huang's face. The Nvidia CEO was offering investment advice, speaking directly to me in Mandarin. Of course, it was not really Huang. It was an AI-generated scam, and I was far from the first to be targeted: across Taiwan, a flood of scams was defrauding millions of citizens. We faced a dilemma. Taiwan has the freest internet in Asia; any content regulation is unacceptable. Yet AI was being used to weaponize that freedom against the citizenry. Our response — and its success — demonstrates something fundamental about how AI alignment must work. We did not ask experts to solve it. We did not let a handful of researchers decide what counted as “fraud.” Instead, we sent 200,000 random text messages asking citizens: what should we do together? Four hundred forty-seven everyday Taiwanese — mirroring our entire population by age, education, region, occupation — deliberated in groups of 10. They were not seeking perfect agreement but uncommon ground — ideas that people with different views could still find reasonable. Within months, we had unanimous parliamentary support for new laws. By 2025, the scam ads were gone. This is what I call [...] --- Outline: (01:44) AI Alignment Today Is Fundamentally Flawed (03:45) The Stakes Are High (07:42) Attentiveness in Practice (09:21) Industry Norms (10:36) Market Design (11:59) Community-Scale Assistants (13:10) From 1% pilots to 99% adoption (14:24) Attentiveness Works (16:44) Discussion about this post --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!L7T_!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F8aa82c35-e9e5-4d10-acd2-4c78c74d566b_1600x1506.pnghttps://substackcdn.com/image/fetch/$s_!X9wB!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Fe42dbbb6-ecf1-46fc-8822-bfc7febd4861_1131x1600.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “AGI’s Last Bottlenecks” by Adam Khoja, Laura Hiscott | 22 Oct 2025 | 00:19:09 | |
Adam Khoja is a co-author of the recent study, “A Definition of AGI.” The opinions expressed in this article are his own and do not necessarily represent those of the study's other authors. Laura Hiscott is a core contributor at AI Frontiers and collaborated on the development and writing of this article.Dan Hendrycks, lead author of “A Definition of AGI,” provided substantial input throughout this article's drafting. ---- In a recent interview on the “Dwarkesh Podcast,” OpenAI co-founder Andrej Karpathy claimed that artificial general intelligence (AGI) is around a decade away, expressing doubt about “over-predictions in the industry.” Coming amid growing discussion of an “AI bubble,” Karpathy's comment throws cold water on some of the more bullish predictions from leading tech figures. Yet those figures don’t seem to be reconsidering their positions. Following Anthropic CEO Dario Amodei's prediction last year that we might have “a country of geniuses [...] --- Outline: (03:50) Missing Capabilities and the Path to Solving Them (05:13) Visual Processing (07:38) On-the-Spot Reasoning (10:15) Auditory Processing (11:09) Speed (12:04) Working Memory (13:16) Long-Term Memory Retrieval (Hallucinations) (14:24) Long-Term Memory Storage (Continual Learning) (16:36) Conclusion (18:47) Discussion about this post --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!IQAW!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F6bc47973-c071-4a51-bec6-80eae1894188_1200x843.pnghttps://substackcdn.com/image/fetch/$s_!JSaQ!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F79923b29-4eea-41e4-a935-2faf8690765a_1600x898.pnghttps://substackcdn.com/image/fetch/$s_!Oo4O!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F926da9e5-af1b-4821-9b41-6f7791a05950_1600x911.pnghttps://substackcdn.com/image/fetch/$s_!7Zb8!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff4232618-9a8d-4357-8e00-f585c83316e7_1600x894.pnghttps://substackcdn.com/image/fetch/$s_!OPLt!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F3a49b812-91ed-41e1-8573-fb710134fb04_1151x743.pnghttps://substackcdn.com/image/fetch/$s_!Byru!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F98210c9e-fd21-494d-a84b-4f2127afd578_1600x893.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “AI Will Be Your Personal Political Proxy” by Bruce Schneier, Nathan E. Sanders | 21 Oct 2025 | 00:17:22 | |
This is an excerpt from the authors’ new book, “Rewiring Democracy: How AI Will Transform Our Politics, Government, and Citizenship,” available for preorder now. Imagine a digital proxy that knows your political preferences as well as you do (or better), tracks every issue on the ballot, and casts votes on your behalf in real time. This vision, once the stuff of science fiction, is quickly becoming technically feasible. Allowing AI to serve as our political proxies sounds radical, but the idea builds on a simple truth about our political system: representative democracy exists because we can’t all be in the room for every decision. The Limits of Representation Representative democracy requires elected officials to stand in for the collective preferences of their constituents. The most understandable reason for this is logistical: citizens don’t all have the time or ability to communicate our preferences directly, and we can’t all fit [...] --- Outline: (00:51) The Limits of Representation (03:44) AI Could Power a More Direct Democracy (07:22) Extending Rights to the Disenfranchised (09:07) Protecting the Integrity of AI-Powered Democracy (10:30) Good AI Representation Could Make Government More Equitable (13:36) The Risks of Overrelying on AI Representatives (15:07) AI Proxies Should Enhance Political Engagement (17:05) Discussion about this post --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “Is China Serious About AI Safety?” by Karson Elmgren, Scott Singer, Oliver Guest | 14 Oct 2025 | 00:13:02 | |
This summer, the World AI Conference (WAIC) in Shanghai began to live up to its name. Previously an almost exclusively domestic event, this year's event attracted a larger group of international visitors to witness the would-be marvels of China's AI ecosystem. It also provided an opportunity to engage foreign counterparts for one of the newest elements of that ecosystem: the China AI Safety and Development Association (CnAISDA). Launched in February 2025 on the sidelines of the Paris AI Action Summit, CnAISDA places China among a small number of jurisdictions with dedicated AI safety institutes, or AISIs — although they increasingly go by other names. AISIs are government-backed institutions with a focus on AI risks, sometimes explicitly including catastrophic risk. Given the otherwise zygotic state of efforts to address potential catastrophic risks of frontier AI systems in China, CnAISDA is potentially a kernel of important things to come. But [...] --- Outline: (01:20) International Convergence? (04:00) An All-Star Team, with Limited Resources (07:07) China's Efforts to Lead on AI Diplomacy (11:04) The Road Ahead (12:44) Discussion about this post --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “AI Deterrence Is Our Best Option” by Dan Hendrycks, Adam Khoja | 22 Sep 2025 | 00:21:20 | |
Earlier this year, Dan Hendrycks, Eric Schmidt, and Alexandr Wang released “Superintelligence Strategy”, a paper addressing the national security implications of states racing to develop artificial superintelligence (ASI) — AI systems that vastly exceed human capabilities across nearly all cognitive tasks. The paper argued that no superpower would remain passive while a rival transformed an AI lead into an insurmountable geopolitical advantage. Instead, capable nations would likely threaten to preemptively sabotage any AI projects they perceived as imminent threats to their survival. But with the right set of stabilizing measures, this impulse toward sabotage could be redirected into a deterrence framework called Mutual Assured AI Malfunction (MAIM). Since its publication, “Superintelligence Strategy” has sparked extended debate. This essay will respond to several critiques of MAIM, while also providing context to readers who are new to the discussion. First, we’ll argue that creating ASI incentivizes state conflict and the tremendous [...] --- Outline: (01:30) Building Superintelligence Amplifies Tensions, and Could Be Considered an Act of War (10:05) MAIM's Proposals Increase Stability (16:52) MAIM Facilitates Redlines (20:28) Our Best Option (21:02) Discussion about this post --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “Summary of “If Anyone Builds It, Everyone Dies”” by Laura Hiscott | 16 Sep 2025 | 00:14:13 | |
Two years ago, AI systems were still fumbling at basic reasoning. Today, they’re drafting legal briefs, solving advanced math problems, and diagnosing medical conditions at expert level. At this dizzying pace, it's difficult to imagine what the technology will be capable of just years from now, let alone decades. But in their new book, If Anyone Builds It, Everyone Dies, Eliezer Yudkowsky and Nate Soares — co-founder and president of the Machine Intelligence Research Institute (MIRI), respectively — argue that there's one easy call we can make: the default outcome of building superhuman AI is that we lose control of it, with consequences severe enough to threaten humanity's survival. Yet despite leading figures in the AI industry expressing concerns about extinction risks from AI, the companies they head up remain engaged in a high-stakes race to the bottom. The incentives are enormous, and the brakes are weak. Having studied [...] --- Outline: (01:14) Today's AI Systems Are Grown Like Organisms, Not Engineered Like Machines (03:43) You Don't Get What You Train For (08:23) AI's Favorite Things (09:27) Why We'd Lose (12:47) The Case for Hope (13:51) Discussion about this post --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!YTm3!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F9124477f-e138-4f01-96f7-c08c4b516c4a_1032x1600.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Cybersecurity is Humanity’s Firewall Against Rogue AI” by Rosario Mastrogiacomo | 10 Sep 2025 | 00:13:52 | |
AI agents are no longer a futuristic concept. They’re increasingly being embedded in the systems we rely on every day. These aren’t just new software features. They are independent digital actors, able to learn, adapt, and make decisions in ways we can’t always predict. Across the AI industry, a fierce race is underway to expand agents’ autonomous capabilities. Some can reset passwords, change permissions, or process transactions without a human ever touching the keyboard. Of course, hackers can also unleash AI agents to gain entry to, and wreak havoc within, those same sensitive systems. I see this transformation daily in my work at the forefront of cybersecurity, where AI agents are rapidly undermining our traditional approaches to safety. But the risk isn’t confined to what these agents can do inside corporate networks. Their activities threaten to ripple outward into society. Left unchecked, they could undermine trust-based systems that make [...] --- Outline: (02:02) How AI Agents Undermine Identity and Trust (07:49) The Infrastructure We Built for Trust (09:07) What We Can and Can't Do with the Tools We Have (12:20) The Future Is Here. Will We Govern It? --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!Dqol!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F61ede93c-fae2-410b-9347-764423c49078_1528x1156.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Precaution Shouldn’t Keep Open-Source AI Behind the Frontier” by Ben Brooks | 02 Sep 2025 | 00:16:22 | |
OpenAI — once considered an oxymoron given its closed-source practices — recently released GPT-OSS, the company's first open language model in half a decade. The model fulfills an earlier pledge to again release “strong” open models that developers can freely modify and deploy. OpenAI approved GPT-OSS in part because the model sits behind the closed-source frontier, including its own GPT-5, which it released just two days later. Meanwhile, Meta — long a champion of frontier open models — has delayed the release of its largest open model, Llama Behemoth, and suggested it may keep its future “superintelligence” models behind paywalls. Meta, which once described open source AI as a way to “control our own destiny,” now cites “novel safety concerns” as a reason to withhold its most capable models. These decisions mark a dramatic pivot for both companies, and reveal how different AI firms are converging on an [...] --- Outline: (02:02) Uncertainty Is Driving Precautionary Policy (04:37) Precaution Disproportionately Chills Open Source (06:48) Restrictions Demand Confident Evidence (09:03) Precaution May Lead to Digital Feudalism (12:54) We Need to Learn to Live with Uncertainty (15:24) We Should Promote, Not Deter, Openness at the Frontier --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “The Hidden AI Frontier” by Oscar Delaney, Ashwin Acharya | 28 Aug 2025 | 00:17:40 | |
OpenAI's GPT-5 launched in early August, after extensive internal testing. But another OpenAI model — one with math skills advanced enough to achieve “gold medal-level performance” on the world's most prestigious math competition — will not be released for months. This isn’t unusual. Increasingly, AI systems with capabilities considerably ahead of what the public can access remain hidden inside corporate labs. This hidden frontier represents America's greatest technological advantage — and a serious, overlooked vulnerability. These internal models are the first to develop dual-use capabilities in areas like cyberoffense and bioweapon design. And they’re increasingly capable of performing the type of research-and-development tasks that go into building the next generation of AI systems — creating a recursive loop where any security failure could cascade through subsequent generations of technology. They’re the crown jewels that adversaries desperately want to steal. This makes their protection vital. Yet the dangers they may [...] --- Outline: (01:42) The Invisible Revolution (03:19) Two Converging Threats (08:02) The Accelerant: AI Building AI (10:21) Why Markets Won't Solve This (12:23) The Role of Government (16:41) Reframing the Race: A Security-First Approach --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “Uncontained AGI Would Replace Humanity” by Anthony Aguirre | 19 Aug 2025 | 00:23:10 | |
The race is on for AGI. Tech companies are in a global race to develop artificial general intelligence (AGI): autonomous systems that perform most tasks as well as a human expert. In early 2024, Meta CEO Mark Zuckerberg declared that Meta is going to build AGI and “open source” it. He is walking his talk. Meta has invested billions of dollars in the highest-power computational elements needed to build giant AI systems, and it has openly released its most powerful AI models. In early 2025, representatives of the Chinese company DeepSeek tweeted their intention to build and openly release AGI. US companies (including OpenAI, Google DeepMind, and Anthropic) are also trying to build AGI. While these companies have not pledged to open-source such a system, recent months have seen a marked shift among U.S. policymakers and AI developers toward support for open-source AI. In July, the White House AI [...] --- Outline: (02:06) What Is AGI? (05:22) Widespread Proliferation with No Guardrails (09:10) The Dire Implications of Unleashing AGI (13:50) The Replacement of Humanity (17:49) Changing Course: Don't Build Uncontrollable AI --- First published: Source: --- Narrated by TYPE III AUDIO. | |||
| “Superintelligence Deterrence Has an Observability Problem” by Jason Ross Arnold | 14 Aug 2025 | 00:24:14 | |
In an age of heightened political division, countering China's efforts to dominate AI has emerged as a rare point of alignment between US Democratic and Republican policymakers. While the two parties have approached the issue in different ways, they generally agree that the AI “arms race” is comparable to the US-Soviet strategic competition during the Cold War, which encompassed not just nuclear weapons but also global security, geopolitical influence, and ideological supremacy. There is, however, no modern deterrence mechanism comparable to the doctrine of Mutually Assured Destruction (MAD), which prevented nuclear war between the US and Soviet Union for four decades — and which is arguably the reason no other nation has used nuclear weapons since. The bridge from MAD to MAIM. Earlier this year, co-authors Dan Hendrycks, Eric Schmidt, and Alexandr Wang proposed a framework called Mutual Assured AI Malfunction (MAIM), hoping to fill that dangerous strategic vacuum. [...] --- Outline: (02:40) The Logic of MAIM (07:11) The Observability Problem Is Bigger Than MAIM's Authors Acknowledge (10:29) Challenge One: Using Appropriate Proxies for AI Progress (13:34) Challenge Two: Observation Must Keep Up With Rapid Progress (16:41) Challenge Three: Superintelligence Development Will Likely Be Widely Decentralized (19:31) Challenge Four: Intelligence Activities Themselves Could Lead to Escalation (21:45) MAIM Started the Conversation on Superintelligence Deterrence, but More Dialogue Is Needed --- First published: Source: --- Narrated by TYPE III AUDIO. --- Images from the article: https://substackcdn.com/image/fetch/$s_!z859!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F13193bdb-d3fb-4a5c-afb7-1f79b708cfac_1600x588.pngApple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app. | |||
| “Open Protocols Can Prevent AI Monopolies” by Isobel Moure, Tim O’Reilly, Ilan Strauss | 30 Jul 2025 | 00:20:57 | |
Can we head off AI monopolies before they harden? As AI models become commoditized, incumbent Big Tech platforms are racing to rebuild their moats at the application layer, around context: the sticky user- and project-level data that makes AI applications genuinely useful. With the right context-aware AI applications, each additional user-chatbot conversation, file upload, or coding interaction improves results; better results attract more users; and more users mean more data. This context flywheel - a rich, structured user- and project-data layer - can drive up switching costs, creating a lock-in effect that effectively traps accumulated data within the platform. Protocols prevent lock-in. We argue that open protocols - exemplified by Anthropic's Model Context Protocol (MCP) - serve as a powerful rulebook, helping to keep API-exposed context fluid and to prevent Big Tech from using data lock-in to extend their monopoly power. However, as an API wrapper, MCP can access [...] --- Outline: (02:33) From Commoditized Models to Context-Rich Applications (07:46) How User Context Is Powering a New Era of Tech Monopolies - and Competition (10:22) Can Protocols Create a Level Playing Field? (12:38) MCP's Impact on the AI Market So Far (14:16) MCP vs. Walled Gardens: The API Gatekeeping Problem (16:06) To Save AI from Enshittification, Support Protocol-Level Interventions --- First published: Source: --- Narrated by TYPE III AUDIO. | |||