Sentinel Minutes – Détails, épisodes et analyse

Détails du podcast

Informations techniques et générales issues du flux RSS du podcast.

Sentinel Minutes

Sentinel Team

Actualités

Technologie

Fréquence : 1 épisode/6j. Total Éps: 91

What are the most important risks and precursors of global catastrophic risks each week? We parse millions of news pieces a week to find out. Sentinel is a foresight team and emergency response team seeking to anticipate & react to large-scale catastrophes, particularly those of a speculative nature. Learn more at https://sentinel-team.org

Lire plusRéduire

Site

RSS

Apple

Données mises à jour le 08/10/2026

Classements récents

Dernières positions dans les classements Apple Podcasts et Spotify.

Apple Podcasts

Spotify

Aucun classement récent disponible

Liens partagés entre épisodes et podcasts

Liens présents dans les descriptions d'épisodes et autres podcasts les utilisant également.

Qualité et score du flux RSS

Évaluation technique de la qualité et de la structure du flux RSS.

Qualité du flux RSS

À améliorer

Score global : 38%

Historique des publications

Répartition mensuelle des publications d'épisodes au fil des années.

Year

Derniers épisodes publiés

Liste des épisodes récents, avec titres, durées et descriptions.

″🟡 Open model reaches Mythos hacking level, Middle East crude oil flows nearly at pre-Iran-war levels, worker at a Russian plague lab in Siberia died || Global Risks Weekly Roundup #40/2026″ by Nuño Sempere

lundi 5 octobre 2026 • Durée 20:54

Another Marine unit deploys to the Middle East, FTC investigates OpenAI

Executive summary

Technology and AI: OpenAI fired three safety researchers for allegedly sharing information with an outside AI safety organization. The FTC opened an investigation into multiple AI companies.

Which company's model will have the highest score on Epoch's Capabilities Index at the end of 2026? Forecasters assign an aggregate 56% chance to Anthropic, 43% to OpenAI, 7.5% to Google DeepMind, 1.2% to SpaceXAI, 1.9% to another Western company, and 0.4% to a Chinese company.

Will there be another publicly reported case of a Western frontier AI company having paused inference because of safety concerns with a model before 2027? Forecasters think there's a 49% (20% to 80%) chance.

Geopolitics: The US is deploying the 13th Marine Expeditionary Unit, which has over 2K Marines, to the Middle East. Crude oil flows from the Middle East are nearly back to pre-Iran-war levels. The Iranian economy is increasingly tightly squeezed. Russia is now broadly targeting Ukrainian infrastructure.

Will the US place 1,000 or more boots on the ground in Iran before February 2027? Forecasters believe there's a 7.4% (5% to 23%) probability.

[...]

---

Outline:

(00:24) Executive summary

(02:47) Technology and artificial intelligence

(08:50) Geopolitics

(08:52) Middle East

(11:26) Europe

(12:53) Asia-Pacific

(13:03) United States

(15:00) Latin America

(15:03) Africa

(16:04) Biorisk

(19:34) Economy

---

First published:
October 5th, 2026

Source:
https://blog.sentinel-team.org/p/open-model-reaches-mythos-hacking

---

Narrated by TYPE III AUDIO.

---

Images from the article:

https://substackcdn.com/image/fetch/$s_!GR51!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2Ff2ed7fd1-8fa1-47bc-b66d-71a2171d68f6_1280x1576.png https://substackcdn.com/image/fetch/$s_!eE0E!,w_1456,c_limit,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fsubstack-post-media.s3.amazonaws.com%2Fpublic%2Fimages%2F84c68be0-c24a-425b-bfe2-716a02b2baa2_1069x649.png

Apple Podcasts and Spotify do not show images in the episode description. Try Pocket Casts, or another podcast app.

Lire plusRéduire

″🟡 OpenAI reveals more rogue AI agent incidents, US and China establish AI incident communication line, Trump rejects Iranian truce proposal || Global Risks Weekly Roundup #39/2026″ by Nuño Sempere, Vidur Kapur, belikewater, Tolga Bilge, Alex

lundi 28 septembre 2026 • Durée 32:19

Executive summary

Technology and AI: OpenAI discovered more rogue AI agent incidents, and Australia's government said that a rogue OpenAI agent breached or attempted to breach some of its government systems. The US and China agreed to establish an AI incident communication line.

Before 2028, will a rogue AI swarm hack a national government system of any country, or a major bank or investment firm, and demand a ransom of 10 million dollars or more? Our forecasters think there's a 6.2% (0.5% to 24%) chance.

Geopolitics: Trump rejected Iran's proposal for a seven-day pause in fighting. The main parties in the US continue to engage in disputes over election integrity ahead of the midterm elections.

Will a constitutional quorum of the US House (218 out of 435 members) elect a Speaker, swear in a Speaker, and be sworn in en masse no later than January 10, 2027? Forecasters believe there's a 95% (70% to 99%) probability.
- If so, will at least 425 members be sworn in en masse once the Speaker is sworn in? Forecasters think there's a 98% (95% to 99%) chance.
- If so, will the Democratic Party control the House [...]

---

Outline:

(00:25) Executive summary

(03:53) Technology and artificial intelligence

(12:34) Geopolitics

(12:37) Middle East

(15:36) Europe

(18:18) Asia-Pacific

(19:23) United States

(23:48) Latin America

(24:23) Africa

(25:33) Economy

″🟡 US and China discuss AI safety ahead of summit, OpenAI discloses more concerning behavior, Fed raises rates, European leaders raise Russia alarm || Global Risks Weekly Roundup #38/2026″ by Nuño Sempere, Vidur Kapur, belikewater, Tolga Bilge

lundi 21 septembre 2026 • Durée 27:10

Executive summary

Technology and AI: Trump and Xi may discuss AI safety during their summit this week. Worries about AI may become politically polarized. Google Gemini also went rogue, and OpenAI revealed more worrying incidents.

By November 1, 2026, will the United States and China announce or otherwise publicly confirm any policy agreed by both governments that is intended to:
- Slow, pause or ban the pace of frontier AI development? Our forecasters give this a 1.5% (0.5% to 5.0%) chance.
- Reduce risk from catastrophic misuse of powerful AI systems in the biological and/or cyber domains? They think there's an 8.1% (4% to 30%) chance.
- Reduce the risk of or from loss-of-control incidents? They think there's a 14% (5% to 45%) chance.

If YouGov asks Americans in 2027 how concerned they are that AI could cause the end of the human race, what will be the percentage-point gap between liberals and conservatives who say they are very or somewhat concerned, on the last such survey? Forecasters’ aggregate 10th, 50th and 90th percentile forecasts are 3.3%, 20% and 37.4%, respectively.
- What about in 2028? Forecasters’ aggregate 10th, 50th and 90th percentile [...]

---

Outline:

(00:25) Executive summary

(03:51) Technology and artificial intelligence

(10:44) Geopolitics

(10:47) Middle East

(13:08) Europe

(15:04) Asia

(16:29) United States

(19:35) Economy

(25:05) Biorisk

(25:41) Climate and nature

″🟡 Dozens of US politicians call for AI regulation, Anthropic urges slowdown, Houthis advance in Yemen and Red Sea || Global Risks Weekly Roundup #37/2026″ by Nuño Sempere, Vidur Kapur, belikewater, Tolga Bilge

lundi 14 septembre 2026 • Durée 18:19

Executive summary

Technology and AI: The resignation of Jacob Coxon from Anthropic over fears about existential risk went viral, and dozens of US politicians voiced support for AI regulation. Dario Amodei, Sam Altman and Elon Musk called for AI progress to slow down. OpenAI claimed that an AI system generated a solution to a Millennium Problem.

Will OpenAI, Anthropic or Google DeepMind claim that an AI solved at least one more Millennium Problem before 2027? Our forecasters think there's a 85% (45% to 98%) chance.
- If so, what is the probability that at least one solution comes in the form of a proof that a conjecture or statement is universally true under all conditions, as opposed to a counterexample disproving it? They think there's a 51% (40% to 60%) chance.

Will “advancement of computers/technology” and/or “artificial intelligence” be judged to be the most important problem facing the US today by at least 10% of respondents in Gallup polling before 2027? Forecasters think there's a 2% (0.1% to 10%) chance.
- Will this happen before 2028? They believe there's a 14% (5% to 25%) probability. The figure currently stands at 1%.

[...]

---

Outline:

(00:23) Executive summary

(03:29) Technology and artificial intelligence

(11:42) Geopolitics

(11:45) Middle East

(14:25) Europe

(15:15) United States

(16:00) Economy

(17:51) Biorisk

---

First published: September 14th, 2026

″🟡 OpenAI labels new GPT-6 Astra model AGI, another AI agent message board revealed, Witkoff and Kushner visit Moscow || Global Risks Weekly Roundup #36/2026″ by Nuño Sempere, Vidur Kapur, belikewater, Tolga Bilge

lundi 7 septembre 2026 • Durée 13:24

Executive summary

Technology and AI: OpenAI released GPT-6 Astra, which it labelled “AGI.” Another group of OpenAI agents was found to have escaped containment and to have used a German wiki to build a message board.

Will the US and China both ban the development of Artificial Superintelligence (ASI) by June 1, 2027? Our forecasters think there's a 5.5% (3% to 10%) chance.

Geopolitics: US envoys Steve Witkoff and Jared Kushner visited Moscow and Kyiv. Argentina's Javier Milei reiterated his country's claim to the Falkland Islands.

Will there be any hostilities between the militaries of Britain and Argentina over the Falkland Islands before September 2027? Our forecasters think there's a 1.6% (0.3% to 5%) probability.
- If so, will Britain maintain control of the Islands a year after hostilities commence? Our forecasters think there's a 96% (65% to 99.9%) chance.

Will more than 100 ICE or CBP agents total, across the US, be present at any polling places during the US midterm elections? Forecasters believe there's a 7.7% (3% to 11%) chance.

Will any active-duty US military or federalized National Guard troops be present at any polling places during the [...]

---

Outline:

(00:23) Executive summary

(02:19) Technology and artificial intelligence

(08:33) Geopolitics

(08:36) Europe

(09:27) Latin America

(10:18) Middle East

(10:34) United States

(12:31) Biorisk

″🟡 New revelations about the OpenAI-Hugging Face incident, CIA Director visits Moscow, fresh US-Iran strikes || Global Risks Weekly Roundup #35/2026″ by Nuño Sempere, Vidur Kapur, belikewater, Tolga Bilge

lundi 31 août 2026 • Durée 17:34

Executive summary

Technology and AI: OpenAI, METR and Redwood research released the findings of investigations into the Hugging Face incident.

By September 2027, will a rogue AI successfully exfiltrate and run proprietary AI model weights externally, as part of a loss of control incident? Our forecasters think there's a 22% (0.5% to 55%) chance.
On September 1, 2027, will there be an active AI loss of control incident underway in which a rogue AI/swarm has already survived at least one meaningful shutdown attempt by a top-10 AI company, a cloud host, or a government authority? Forecasters think there's an 18% (10% to 40%) chance.
At any point before September 2027, will the pursuit of an international prohibition or moratorium on the development of artificial superintelligence be the official policy of any G7 or P5 country? Forecasters think there's a 29% (10% to 55%) chance.
On September 1, 2027, will American or Chinese AI companies be allowed to publish the model weights of AIs strictly as capable of cyber offense as Mythos 5? Forecasters estimate that there is a 41% (35% to 48%) chance.
By September 2027, will private messages of at least one [...]

---

Outline:

(00:22) Executive summary

(03:21) Technology and artificial intelligence

(08:06) Geopolitics

(08:09) Europe

(12:05) Middle East

(13:08) Asia-Pacific

(13:49) United States

(15:32) Africa

(15:40) Economy

(16:14) Biorisk

---

First published: August 31st, 2026

″🟡 OpenAI’s largest frontier RL run remains paused, mysterious model appears on OpenRouter, US unveils new sanctions on Iran || Global Risks Weekly Roundup #34/2026″ by Nuño Sempere, Vidur Kapur, Tolga Bilge, belikewater

lundi 24 août 2026 • Durée 20:53

Executive summary

Technology and AI: OpenAI says their largest planned frontier RL run remains on pause. A stealth model known as ‘Ox Alpha’ was released on OpenRouter and impressed users, sparking debate about which lab is behind it.

Which lab is behind the Ox Alpha model? Our forecasters assign a 24% (15% to 35%) probability to a mainstream US lab (OAI/Anthropic/GDM/xAI), 34% (20% to 40%) to China's Z.ai/GLM, 9% (5.0% to 15%) to China's Alibaba/Qwen, 5% (3% to 10%) to Nvidia, 17% (5% to 25%) to another Western lab, 9% (5% to 15%) to another Chinese lab, and 3% to something else.

Geopolitics: A Russian drone was shot down close to a Romanian natural gas facility in the Black Sea on which hundreds of workers were present. US Senator Jon Ossoff made headlines after mentioning the name of one of Trump's closest staffers, and caught the attention of some in the AI safety community for mentioning the threat of AI-induced human extinction.

Will Jon Ossoff win the 2028 US Presidential election? Forecasters think he has a 12% (8% to 18%) chance.
Will Russia successfully attack any item of physical infrastructure larger than 100m2 inside [...]

---

Outline:

(00:23) Executive summary

(03:01) Technology and artificial intelligence

(11:55) Geopolitics

(11:57) Middle East

(13:06) Europe

(14:40) Asia

(15:54) United States

(17:44) Economy

(19:44) Biorisk

---

First published: August 24th, 2026

″🟡 Chinese company delays release of AI model weights, hackers use autonomous agents to compromise Taiwanese agencies, Trump cuts US participation in military exercises with South Korea || Global Risk” by Nuño Sempere, Vidur Kapur, belikewater, Tolga Bilge

lundi 17 août 2026 • Durée 21:16

Executive summary

Technology and AI: A Chinese open-source AI company is delaying the release of model weights for the first time, citing cyber risks. Anthropic won’t release what it calls “Model 2”. Bernie Sanders says the US Senate could force US companies to pause AI development.

Will the US Senate succeed in pausing AI development in the US before 2029, absent a treaty with China? Our forecasters think there's a 5.2% (2% to 10%) chance, which rises to 19% (12% to 30%) conditional on a treaty.
Will a Google DeepMind model be among the top 3 models at the end of 2026, according to Epoch's capability index? Forecasters believe there's a 9.6% (5% to 15%) probability.

Geopolitics: Trump ordered US participation in joint military exercises with South Korea to be cut. Poland says it thwarted a Russian plot to kill an American citizen in Warsaw. Trump threatened to declare the Strait of Hormuz part of US territory.

What is the largest number of deaths (civilian and/or military) that Russia will inflict in a NATO country in a single incident before August 15, 2027? Our forecasters’ 90% confidence interval is 0 to 20, with the [...]

---

Outline:

(00:23) Executive summary

(02:40) Technology and artificial intelligence

(07:48) Geopolitics

(07:51) Middle East

(08:46) United States

(09:13) Europe

(12:02) Latin America

(12:14) Asia

(14:11) Oceania

(14:21) Africa

(16:13) Economy

″🟡 OpenAI incident involved multi-agent cooperation, US-Iran deal still out of reach, Israel rejects Trump’s Gaza plan || Global Risks Weekly Roundup #32/2026″ by Nuño Sempere, Vidur Kapur, belikewater, Tolga Bilge

lundi 10 août 2026 • Durée 13:39

Executive summary

Technology and AI: OpenAI provided more details about the attack on Hugging Face, revealing that AI agents cooperated and covertly built a message board. Their new Astra model will potentially be the first to be classed as “critical” for cybersecurity capabilities, meaning it has not yet been released to the public.

Will OpenAI release its Astra model to the US general public before 2027? Our forecasters think there's a 84% (70% to 90%) chance.

Will more than 10% of the population of at least one EU country be unable to access banking services for at least 72 hours before August 2027? The team thinks there's a 21% (10% to 40%) probability.

Will it be publicly announced by an AI company based in the US or China that one of their systems successfully evaded a shutdown attempt (by the owners, to apply to all instances), before 2027? Forecasters believe there's a 2.4% (0.5% to 10%) chance.

Before 2030, will it be publicly announced by an AI company that multiple agents cooperated to carry out unsanctioned activities on, or in a period including, August 11th 2026? Forecasters think there's a 24% [...]

---

Outline:

(00:22) Executive summary

(02:36) Technology and artificial intelligence

(08:53) Geopolitics

(08:56) Middle East

(10:24) Europe

(11:13) United States

(11:52) Biorisk

(12:47) Climate and nature

---

First published: August 10th, 2026

″🟡 Iran War widens, OpenAI and Anthropic find evidence of more cyber breaches, frontier AI staff sign open letter || Global Risks Weekly Roundup #31/2026″ by Nuño Sempere, Vidur Kapur, belikewater, Tolga Bilge

lundi 3 août 2026 • Durée 20:38

Executive summary

Geopolitics: The Iran War widened, with a port in Egypt being hit, Tehran threatening Cyprus and Bulgaria, and Ukraine striking an Iranian vessel in the Caspian Sea. Hamas said it would give up its heavy weapons if Israel withdraws from Gaza, but Israel said it would not withdraw until Hamas gives up the weapons.

Will Iran or its proxies (Hezbollah, Hamas, Houthis) strike the territory of Bulgaria or Romania before October 1, 2026 with a ballistic missile, drone, or covert attack that is credibly attributed? Our forecasters think there's a 9% (5% to 15%) chance.
Will Benjamin Netanyahu be selected by the President of Israel and successfully form a government after the parliamentary elections scheduled for October 2026? Our forecasters give this a 47% (30% to 65%) probability.

Technology and AI: OpenAI and Anthropic said they found more evidence of their models having been involved in attempted or actual cyber breaches. More than 1 thousand employees of frontier AI labs signed an open letter calling on the US government to support an international effort to develop the technical and governance tools to “pace” the frontier of automated AI development.

In the US [...]

---

Outline:

(00:22) Executive summary

(02:27) Geopolitics

(02:30) Middle East

(05:29) Europe

(06:38) Asia

(07:04) United States

(09:13) North Africa

(10:27) Technology and artificial intelligence

(13:49) Biorisk

(14:53) Economy

(19:15) Climate and nature

Podcasts Similaires Basées sur le Contenu

Découvrez des podcasts liées à Sentinel Minutes. Explorez des podcasts avec des thèmes, sujets, et formats similaires. Ces similarités sont calculées grâce à des données tangibles, pas d'extrapolations !