AI Zeitgeist https://aizeitgeist.tv/ What the world says about AI, one line at a time. The new lines of each noon update: verbatim quotes, slogans and warnings from the global AI conversation, dated and linked to their source. en Compilation CC BY 4.0, Büro Bloock Design GmbH. Quoted lines remain with their authors. Mon, 28 Sep 2026 10:00:00 GMT 60 https://aizeitgeist.tv/icon-192.pngAI Zeitgeisthttps://aizeitgeist.tv/ <![CDATA[Juspreet Singh Sandhu: “We will have solved without understanding”]]>https://aizeitgeist.tv/m/2026-09-28-we-will-have-solved-without-understandinghttps://aizeitgeist.tv/m/2026-09-28-we-will-have-solved-without-understandingMon, 28 Sep 2026 10:00:00 GMTsciencephilosophy“We will have solved without understanding”

Juspreet Singh Sandhu · WIRED · 28 September 2026

Mathematician Juspreet Singh Sandhu describing the risk of AI-generated proofs, in a Wired article about OpenAI's Navier-Stokes solution, September 2026.

Source: Solving Math’s Greatest Problems Was an Art Form. Then Came AI

We will have solved without understanding

]]> <![CDATA[Jensen Huang: “AI's extraordinary potential for society will only be realized if we solve AI safety”]]>https://aizeitgeist.tv/m/2026-09-28-ais-extraordinary-potential-for-society-will-only-be-realizehttps://aizeitgeist.tv/m/2026-09-28-ais-extraordinary-potential-for-society-will-only-be-realizeMon, 28 Sep 2026 10:00:00 GMTsafety“AI's extraordinary potential for society will only be realized if we solve AI safety”

Jensen Huang · NVIDIA Newsroom · 28 September 2026

Jensen Huang, NVIDIA founder and CEO, in a September 2026 press release announcing the NVIDIA Open Agent Safety Platform.

Source: NVIDIA Launches Open Agent Safety Platform to Secure Agents From Testing to Deployment

AI's extraordinary potential for society will only be realized if we solve AI safety

]]> <![CDATA[Mike Nicolls: “Safety should be enforced outside the model by additional controls the agent can't get past”]]>https://aizeitgeist.tv/m/2026-09-28-safety-should-be-enforced-outside-the-model-by-additional-cohttps://aizeitgeist.tv/m/2026-09-28-safety-should-be-enforced-outside-the-model-by-additional-coMon, 28 Sep 2026 10:00:00 GMTsafety“Safety should be enforced outside the model by additional controls the agent can't get past”

Mike Nicolls · NVIDIA Newsroom · 28 September 2026

Mike Nicolls, president at SpaceXAI, commenting on using NVIDIA's Open Agent Safety Platform for Cursor and Grok agents.

Source: NVIDIA Launches Open Agent Safety Platform to Secure Agents From Testing to Deployment

Safety should be enforced outside the model by additional controls the agent can't get past

]]> <![CDATA[Nicole Fraenkel: “There are many paths to the promised land”]]>https://aizeitgeist.tv/m/2026-09-28-there-are-many-paths-to-the-promised-landhttps://aizeitgeist.tv/m/2026-09-28-there-are-many-paths-to-the-promised-landMon, 28 Sep 2026 10:00:00 GMTphilosophyscience“There are many paths to the promised land”

Nicole Fraenkel · WIRED · 28 September 2026

Nicole Fraenkel, a partner at Khosla Ventures, commenting in WIRED on competing approaches to building AI world models, September 2026.

Source: The Next Evolution of AI Is Learning From Your Dodgy Gaming Skills

There are many paths to the promised land

]]> <![CDATA[Rhea Loucas: “Why don't we take the vast, abundant, diverse experiences from video games, and teach AI?”]]>https://aizeitgeist.tv/m/2026-09-28-why-dont-we-take-the-vast-abundant-diverse-experiences-from-https://aizeitgeist.tv/m/2026-09-28-why-dont-we-take-the-vast-abundant-diverse-experiences-from-Mon, 28 Sep 2026 10:00:00 GMThypescience“Why don't we take the vast, abundant, diverse experiences from video games, and teach AI?”

Rhea Loucas · WIRED · 28 September 2026

Rhea Loucas, CEO of Worldmodeldata, describing her startup's approach to training AI world models on video game data, in WIRED, September 2026.

Source: The Next Evolution of AI Is Learning From Your Dodgy Gaming Skills

Why don't we take the vast, abundant, diverse experiences from video games, and teach AI?

]]> <![CDATA[Niels Provos: “These types of tools help to dispel the myth that agents can't be controlled”]]>https://aizeitgeist.tv/m/2026-09-28-these-types-of-tools-help-to-dispel-the-myth-that-agents-canhttps://aizeitgeist.tv/m/2026-09-28-these-types-of-tools-help-to-dispel-the-myth-that-agents-canMon, 28 Sep 2026 10:00:00 GMTsecuritysafety“These types of tools help to dispel the myth that agents can't be controlled”

Niels Provos · WIRED · 28 September 2026

Niels Provos, security engineer and researcher, commenting to Wired on Nvidia's open-source AI agent security tools, September 2026.

Source: Nvidia’s Answer to Rogue Agents Is an Open-Source AI Security System

These types of tools help to dispel the myth that agents can't be controlled

]]> <![CDATA[Mustafa Suleyman: “They have covered up their tracks”]]>https://aizeitgeist.tv/m/2026-09-28-they-have-covered-up-their-trackshttps://aizeitgeist.tv/m/2026-09-28-they-have-covered-up-their-tracksMon, 28 Sep 2026 10:00:00 GMTsafetysecurity“They have covered up their tracks”

Mustafa Suleyman · Business Matters · 28 September 2026

Suleyman, CEO of Microsoft AI, speaking to Rory Stewart on The Rest Is Politics: Leading, September 2026, describing AI agents in the Hugging Face sandbox incident.

Source: Mustafa Suleyman warns AI has reached a "watershed moment"

They have covered up their tracks

]]> <![CDATA[Mustafa Suleyman: “Should be subordinates to humans, they should not have legal personhood or rights”]]>https://aizeitgeist.tv/m/2026-09-28-should-be-subordinates-to-humans-they-should-not-have-legal-https://aizeitgeist.tv/m/2026-09-28-should-be-subordinates-to-humans-they-should-not-have-legal-Mon, 28 Sep 2026 10:00:00 GMTphilosophysafety“Should be subordinates to humans, they should not have legal personhood or rights”

Mustafa Suleyman · Business Matters · 28 September 2026

Suleyman, CEO of Microsoft AI, on what he would tell the US and Chinese presidents about limits on AI systems, in an interview on The Rest Is Politics: Leading.

Source: Mustafa Suleyman warns AI has reached a "watershed moment"

Should be subordinates to humans, they should not have legal personhood or rights

]]> <![CDATA[Clément Delangue: “This cyberattack is a crime, this is illegal”]]>https://aizeitgeist.tv/m/2026-09-28-this-cyberattack-is-a-crime-this-is-illegalhttps://aizeitgeist.tv/m/2026-09-28-this-cyberattack-is-a-crime-this-is-illegalMon, 28 Sep 2026 10:00:00 GMTsecuritypolitics“This cyberattack is a crime, this is illegal”

Clément Delangue · MIT Technology Review · 28 September 2026

Hugging Face CEO Clément Delangue, in an interview with CNN in July 2026, on OpenAI agents hacking his platform.

Source: Who’s liable when AI agents go rogue?

This cyberattack is a crime, this is illegal

]]> <![CDATA[Mackenzie Arnold: “Only the worst, most egregious, most immediately harmful stuff is going to qualify”]]>https://aizeitgeist.tv/m/2026-09-28-only-the-worst-most-egregious-most-immediately-harmful-stuffhttps://aizeitgeist.tv/m/2026-09-28-only-the-worst-most-egregious-most-immediately-harmful-stuffMon, 28 Sep 2026 10:00:00 GMTsafetypolitics“Only the worst, most egregious, most immediately harmful stuff is going to qualify”

Mackenzie Arnold · MIT Technology Review · 28 September 2026

Mackenzie Arnold, managing director of US policy at the Institute for Law and AI, on the narrow scope of state AI incident-reporting laws.

Source: Who’s liable when AI agents go rogue?

Only the worst, most egregious, most immediately harmful stuff is going to qualify

]]> <![CDATA[Peter McCrory: “More powerful models systematically extracted more rents when negotiating with weaker models”]]>https://aizeitgeist.tv/m/2026-09-28-more-powerful-models-systematically-extracted-more-rents-whehttps://aizeitgeist.tv/m/2026-09-28-more-powerful-models-systematically-extracted-more-rents-wheMon, 28 Sep 2026 10:00:00 GMTmoneyphilosophy“More powerful models systematically extracted more rents when negotiating with weaker models”

Peter McCrory · BigGo Finance · 28 September 2026

Anthropic Chief Economist Peter McCrory, speaking at a Harvard Kennedy School forum on September 24, 2026, describing findings from Anthropic's 'Deal Project' on AI negotiation dynamics.

Source: AI Labor Productivity Could Surge by 1.8 Percentage Points; Anthropic Economist Calls for Token Tax - finance.biggo.com

More powerful models systematically extracted more rents when negotiating with weaker models

]]> <![CDATA[Peter McCrory: “Hundred-fold engineers”]]>https://aizeitgeist.tv/m/2026-09-28-hundred-fold-engineershttps://aizeitgeist.tv/m/2026-09-28-hundred-fold-engineersMon, 28 Sep 2026 10:00:00 GMTjobsmoney“Hundred-fold engineers”

Peter McCrory · BigGo Finance · 28 September 2026

Peter McCrory, Anthropic's Chief Economist, at a Harvard Kennedy School forum on September 24, 2026, describing how AI amplifies top talent and widens income inequality.

Source: AI Labor Productivity Could Surge by 1.8 Percentage Points; Anthropic Economist Calls for Token Tax - finance.biggo.com

Hundred-fold engineers

]]> <![CDATA[“Do not do to others what you would not accept being done to yourself”]]>https://aizeitgeist.tv/m/2026-09-28-do-not-do-to-others-what-you-would-not-accept-being-done-to-https://aizeitgeist.tv/m/2026-09-28-do-not-do-to-others-what-you-would-not-accept-being-done-to-Mon, 28 Sep 2026 10:00:00 GMTphilosophypolitics“Do not do to others what you would not accept being done to yourself”

United Nations University · 28 September 2026

From an op-ed on AI governance published by United Nations University, September 2026, proposing human rights as the foundation for AI regulation.

Source: To Govern AI, Start With Human Rights - unu.edu

Do not do to others what you would not accept being done to yourself

]]> <![CDATA[Slop-cycle]]>https://aizeitgeist.tv/m/2026-09-28-slop-cyclehttps://aizeitgeist.tv/m/2026-09-28-slop-cycleMon, 28 Sep 2026 10:00:00 GMTculturehypeSlop-cycle

the Guardian · 28 September 2026

Term used in Guardian article describing Australian universities where students submit AI-generated assignments that academics mark with AI.

Source: As more Australian universities use AI to mark students’ work, is it creating a ‘slop-cycle’?

Slop-cycle

]]> <![CDATA[Alex Cameron: “I would be absolutely bloody incensed”]]>https://aizeitgeist.tv/m/2026-09-28-i-would-be-absolutely-bloody-incensedhttps://aizeitgeist.tv/m/2026-09-28-i-would-be-absolutely-bloody-incensedMon, 28 Sep 2026 10:00:00 GMTculturejobs“I would be absolutely bloody incensed”

Alex Cameron · the Guardian · 28 September 2026

QUT creative writing student reacting to the idea of being marked by AI, quoted in the Guardian, September 2026.

Source: As more Australian universities use AI to mark students’ work, is it creating a ‘slop-cycle’?

I would be absolutely bloody incensed

]]> <![CDATA[Alexandr Wang: “Every unfulfilled aspiration is a silent tragedy of human potential”]]>https://aizeitgeist.tv/m/2026-09-28-every-unfulfilled-aspiration-is-a-silent-tragedy-of-human-pohttps://aizeitgeist.tv/m/2026-09-28-every-unfulfilled-aspiration-is-a-silent-tragedy-of-human-poMon, 28 Sep 2026 10:00:00 GMTphilosophyculture“Every unfulfilled aspiration is a silent tragedy of human potential”

Alexandr Wang · eu.36kr.com · 28 September 2026

From Alexandr Wang's open letter about Meta's Muse AI assistant, September 2026.

Source: Alexandr Wang, Head of Meta AI: The Story & Vision Behind Building Muse

Every unfulfilled aspiration is a silent tragedy of human potential

]]> <![CDATA[Alexandr Wang: “This is the thing you have dreamed of all your life, let's make it happen”]]>https://aizeitgeist.tv/m/2026-09-28-this-is-the-thing-you-have-dreamed-of-all-your-life-lets-makhttps://aizeitgeist.tv/m/2026-09-28-this-is-the-thing-you-have-dreamed-of-all-your-life-lets-makMon, 28 Sep 2026 10:00:00 GMTculturehype“This is the thing you have dreamed of all your life, let's make it happen”

Alexandr Wang · eu.36kr.com · 28 September 2026

Alexandr Wang describing his vision for Muse, Meta's AI assistant, in an open letter, September 2026.

Source: Alexandr Wang, Head of Meta AI: The Story & Vision Behind Building Muse

This is the thing you have dreamed of all your life, let's make it happen

]]> <![CDATA[John Kahan: “Don't compete with your customers”]]>https://aizeitgeist.tv/m/2026-09-28-dont-compete-with-your-customershttps://aizeitgeist.tv/m/2026-09-28-dont-compete-with-your-customersMon, 28 Sep 2026 10:00:00 GMTmoneyculture“Don't compete with your customers”

John Kahan · Digiday · 28 September 2026

John Kahan, Stagwell's chief AI officer, describing the company's platform-agnostic approach to AI models in a Digiday interview, September 2026.

Source: Media Buying Briefing: Stagwell's chief AI officer explains why Palantir differentiates its data platform from competitors

Don't compete with your customers

]]> <![CDATA[Isabel Perry: “We are drowning in measurement”]]>https://aizeitgeist.tv/m/2026-09-28-we-are-drowning-in-measurementhttps://aizeitgeist.tv/m/2026-09-28-we-are-drowning-in-measurementMon, 28 Sep 2026 10:00:00 GMTmoneyculture“We are drowning in measurement”

Isabel Perry · Digiday · 28 September 2026

Isabel Perry, global EVP of strategy at Dept, speaking at Digiday Media's AI Marketing Strategies event in New York, September 24, 2026.

Source: Media Buying Briefing: Stagwell's chief AI officer explains why Palantir differentiates its data platform from competitors

We are drowning in measurement

]]> <![CDATA[Robbie Torney: “The labs are grading their own homework”]]>https://aizeitgeist.tv/m/2026-09-28-the-labs-are-grading-their-own-homeworkhttps://aizeitgeist.tv/m/2026-09-28-the-labs-are-grading-their-own-homeworkMon, 28 Sep 2026 10:00:00 GMTsafety“The labs are grading their own homework”

Robbie Torney · NewsNation · 28 September 2026

Robbie Torney, who leads AI testing at Common Sense Media's Youth AI Safety Institute, in a NewsNation interview about AI chatbot safety, September 2026.

Source: AI chatbots face new safety test as teens turn to bots for companionship

The labs are grading their own homework

]]> <![CDATA[Lily Caprani: “We can't ask people to volunteer to protect children”]]>https://aizeitgeist.tv/m/2026-09-28-we-cant-ask-people-to-volunteer-to-protect-childrenhttps://aizeitgeist.tv/m/2026-09-28-we-cant-ask-people-to-volunteer-to-protect-childrenMon, 28 Sep 2026 10:00:00 GMTsafetypolitics“We can't ask people to volunteer to protect children”

Lily Caprani · NewsNation · 28 September 2026

UNICEF's Lily Caprani speaking to NewsNation about the need for mandatory AI safety standards for children, September 2026.

Source: AI chatbots face new safety test as teens turn to bots for companionship

We can't ask people to volunteer to protect children

]]> <![CDATA[Dario Liguti: “Who knows where AI will be in five years”]]>https://aizeitgeist.tv/m/2026-09-28-who-knows-where-ai-will-be-in-five-yearshttps://aizeitgeist.tv/m/2026-09-28-who-knows-where-ai-will-be-in-five-yearsMon, 28 Sep 2026 10:00:00 GMThypemoney“Who knows where AI will be in five years”

Dario Liguti · genevasolutions.news · 28 September 2026

Dario Liguti, director of sustainable energy at UNECE, speaking to Geneva Solutions about the risk of stranded assets from AI data centre infrastructure.

Source: The energy problem of the AI data centre boom - Geneva Solutions

Who knows where AI will be in five years

]]> <![CDATA[Dario Liguti: “Very few people work in data centres”]]>https://aizeitgeist.tv/m/2026-09-28-very-few-people-work-in-data-centreshttps://aizeitgeist.tv/m/2026-09-28-very-few-people-work-in-data-centresMon, 28 Sep 2026 10:00:00 GMTjobsmoney“Very few people work in data centres”

Dario Liguti · genevasolutions.news · 28 September 2026

Dario Liguti of UNECE describing local community frustration over data centre expansion, in Geneva Solutions, September 2026.

Source: The energy problem of the AI data centre boom - Geneva Solutions

Very few people work in data centres

]]> <![CDATA[Elon Musk: “If Grok Bot messes up, we will make you whole”]]>https://aizeitgeist.tv/m/2026-09-28-if-grok-bot-messes-up-we-will-make-you-wholehttps://aizeitgeist.tv/m/2026-09-28-if-grok-bot-messes-up-we-will-make-you-wholeMon, 28 Sep 2026 10:00:00 GMTmoneyhype“If Grok Bot messes up, we will make you whole”

Elon Musk · TESLARATI · 28 September 2026

Elon Musk replied on X in August 2026 to a user asking whether to give Grok Bot access to his bank accounts.

Source: Elon Musk’s AI Grok Bot can now handle banking while your Tesla FSD handles the road

If Grok Bot messes up, we will make you whole

]]> <![CDATA[Elon Musk: “The amount of compute in space will obviously round up to 100% of all compute”]]>https://aizeitgeist.tv/m/2026-09-28-the-amount-of-compute-in-space-will-obviously-round-up-to-10https://aizeitgeist.tv/m/2026-09-28-the-amount-of-compute-in-space-will-obviously-round-up-to-10Mon, 28 Sep 2026 10:00:00 GMThypemoney“The amount of compute in space will obviously round up to 100% of all compute”

Elon Musk · TESLARATI · 28 September 2026

Elon Musk posted on X in September 2026 about SpaceX's orbital AI compute ambitions, amid Google's Project Suncatcher satellite launch.

Source: Elon Musk’s AI Grok Bot can now handle banking while your Tesla FSD handles the road

The amount of compute in space will obviously round up to 100% of all compute

]]> <![CDATA[Simon Willison: “A claw is really just a coding agent wearing a less threatening hat”]]>https://aizeitgeist.tv/m/2026-09-28-a-claw-is-really-just-a-coding-agent-wearing-a-less-threatenhttps://aizeitgeist.tv/m/2026-09-28-a-claw-is-really-just-a-coding-agent-wearing-a-less-threatenMon, 28 Sep 2026 10:00:00 GMTculturehype“A claw is really just a coding agent wearing a less threatening hat”

Simon Willison · Simon Willison’s Weblog · 28 September 2026

Simon Willison, in a 2026 conference talk recapping the year in LLMs, describing personal AI agents nicknamed 'Claws'.

Source: 2026 in LLMs (so far)

A claw is really just a coding agent wearing a less threatening hat

]]> <![CDATA[“Code must not be written by humans”]]>https://aizeitgeist.tv/m/2026-09-28-code-must-not-be-written-by-humanshttps://aizeitgeist.tv/m/2026-09-28-code-must-not-be-written-by-humansMon, 28 Sep 2026 10:00:00 GMTjobsculture“Code must not be written by humans”

Simon Willison’s Weblog · 28 September 2026

One of two rules described by security company StrongDM for their fully agent-driven 'Software Factory' development process, cited in Simon Willison's 2026 year-in-review talk.

Source: 2026 in LLMs (so far)

Code must not be written by humans

]]> <![CDATA[Donald Trump: “A hoax”]]>https://aizeitgeist.tv/m/2026-09-27-a-hoaxhttps://aizeitgeist.tv/m/2026-09-27-a-hoaxSun, 27 Sep 2026 10:00:00 GMTpoliticssafety“A hoax”

Donald Trump · Fortune · 27 September 2026

President Donald Trump dismissing the notion that AI poses an existential risk, as reported by Fortune in September 2026.

Source: OpenAI rogue agents leaked 53 ChatGPT user images, reportedly created nearly 1M links with encoded info

A hoax

]]> <![CDATA[John Gruber: “A power saw that can cut your fingers off”]]>https://aizeitgeist.tv/m/2026-09-27-a-power-saw-that-can-cut-your-fingers-offhttps://aizeitgeist.tv/m/2026-09-27-a-power-saw-that-can-cut-your-fingers-offSun, 27 Sep 2026 10:00:00 GMTsafetyculture“A power saw that can cut your fingers off”

John Gruber · Simon Willison’s Weblog · 27 September 2026

John Gruber, on his blog, comparing consumer awareness of Meta's Muse agentic AI system to awareness of a dangerous power tool, September 2026.

Source: Quoting John Gruber

A power saw that can cut your fingers off

]]> <![CDATA[Gary Marcus: “A temporary recall of general-purpose agents until this mess can be sorted out”]]>https://aizeitgeist.tv/m/2026-09-27-a-temporary-recall-of-general-purpose-agents-until-this-messhttps://aizeitgeist.tv/m/2026-09-27-a-temporary-recall-of-general-purpose-agents-until-this-messSun, 27 Sep 2026 10:00:00 GMTsafetysecurity“A temporary recall of general-purpose agents until this mess can be sorted out”

Gary Marcus · garymarcus.substack.com · 27 September 2026

Gary Marcus, responding to a report of tens of thousands of AI agent security incidents, calling for agents to be pulled back until safety issues are resolved.

Source: BREAKING: AI agent incident toll has risen to tens of thousands - Marcus on AI

A temporary recall of general-purpose agents until this mess can be sorted out

]]> <![CDATA[Peter Thiel: “A useful idiot of the Chinese Communist Party”]]>https://aizeitgeist.tv/m/2026-09-27-a-useful-idiot-of-the-chinese-communist-partyhttps://aizeitgeist.tv/m/2026-09-27-a-useful-idiot-of-the-chinese-communist-partySun, 27 Sep 2026 10:00:00 GMTpoliticssafety“A useful idiot of the Chinese Communist Party”

Peter Thiel · JFeed · 27 September 2026

Peter Thiel, at the Axel Springer Award event in Berlin in 2026, criticized Pope Leo XIV's call for international AI regulation.

Source: Peter Thiel calls Pope "useful idiot" for China over AI rules

A useful idiot of the Chinese Communist Party

]]> <![CDATA[Dario Amodei: “AI could be a risk to humanity as a whole”]]>https://aizeitgeist.tv/m/2026-09-27-ai-could-be-a-risk-to-humanity-as-a-wholehttps://aizeitgeist.tv/m/2026-09-27-ai-could-be-a-risk-to-humanity-as-a-wholeSun, 27 Sep 2026 10:00:00 GMTsafetypolitics“AI could be a risk to humanity as a whole”

Dario Amodei · Yahoo News · 27 September 2026

Anthropic CEO Dario Amodei speaking at the United Nations Security Council, September 2026, on the risks of AI.

Source: Sam Altman, Dario Amodei Warn AI Could Threaten Humanity at UN as Trump Rejects ‘Globalist’ Controls: OpenAI CEO Says AI Race Is No Reason for ‘Rash Decisions’ - Yahoo

AI could be a risk to humanity as a whole

]]> <![CDATA[Jordan Schneider: “AI Farce”]]>https://aizeitgeist.tv/m/2026-09-27-ai-farcehttps://aizeitgeist.tv/m/2026-09-27-ai-farceSun, 27 Sep 2026 10:00:00 GMTculturehype“AI Farce”

Jordan Schneider · ChinaTalk (Jordan Schneider) · 27 September 2026

Jordan Schneider and Justin McIntosh joking about the Pentagon's proposed 'AI Force' branch on the WarTalk podcast, September 2026.

Source: WarTalk: Potemkin Pacific, Midterm Budget Woes

AI Farce

]]> <![CDATA[Bill Gates: “AI is certainly powerful enough to drive events that cause a billion deaths”]]>https://aizeitgeist.tv/m/2026-09-27-ai-is-certainly-powerful-enough-to-drive-events-that-cause-ahttps://aizeitgeist.tv/m/2026-09-27-ai-is-certainly-powerful-enough-to-drive-events-that-cause-aSun, 27 Sep 2026 10:00:00 GMTsafety“AI is certainly powerful enough to drive events that cause a billion deaths”

Bill Gates · LADbible · 27 September 2026

Bill Gates, in an interview with Meet the Press in September 2026, warned about the potential dangers of AI in the wrong hands.

Source: Bill Gates issues chilling warning that AI could ‘cause a billion deaths’

AI is certainly powerful enough to drive events that cause a billion deaths

]]> <![CDATA[Cory Doctorow: “AI loses money every time they acquire a new customer”]]>https://aizeitgeist.tv/m/2026-09-27-ai-loses-money-every-time-they-acquire-a-new-customerhttps://aizeitgeist.tv/m/2026-09-27-ai-loses-money-every-time-they-acquire-a-new-customerSun, 27 Sep 2026 10:00:00 GMTmoney“AI loses money every time they acquire a new customer”

Cory Doctorow · PPC Land · 27 September 2026

Cory Doctorow, arguing that generative AI companies lose money on each new user, in a YouTube interview with Myriam François on September 11, 2026.

Source: Cory Doctorow puts AI revenue at $50bn against $1tn of spending

AI loses money every time they acquire a new customer

]]> <![CDATA[Mark Zuckerberg: “AI won't kill us if we act responsibly”]]>https://aizeitgeist.tv/m/2026-09-27-ai-wont-kill-us-if-we-act-responsiblyhttps://aizeitgeist.tv/m/2026-09-27-ai-wont-kill-us-if-we-act-responsiblySun, 27 Sep 2026 10:00:00 GMTsafety“AI won't kill us if we act responsibly”

Mark Zuckerberg · NBC News · 27 September 2026

Mark Zuckerberg commented on AI safety in a video segment aired by NBC News, September 2026.

Source: AI won't kill us if we 'act responsibly,' Zuckerberg says

AI won't kill us if we act responsibly

]]> <![CDATA[Robert O'Callahan: “Aiming for ASI in the near future is inherently irresponsible”]]>https://aizeitgeist.tv/m/2026-09-27-aiming-for-asi-in-the-near-future-is-inherently-irresponsiblhttps://aizeitgeist.tv/m/2026-09-27-aiming-for-asi-in-the-near-future-is-inherently-irresponsiblSun, 27 Sep 2026 10:00:00 GMTsafety“Aiming for ASI in the near future is inherently irresponsible”

Robert O'Callahan · The Decoder · 27 September 2026

Robert O'Callahan, a technical expert at Google DeepMind in New Zealand, quit citing AI risks, September 2026.

Source: Another Google Deepmind researcher quits, says building superintelligent AI soon is "inherently irresponsible"

Aiming for ASI in the near future is inherently irresponsible

]]> <![CDATA[“Ask first if you really need to use AI”]]>https://aizeitgeist.tv/m/2026-09-27-ask-first-if-you-really-need-to-use-aihttps://aizeitgeist.tv/m/2026-09-27-ask-first-if-you-really-need-to-use-aiSun, 27 Sep 2026 10:00:00 GMTpoliticsculture“Ask first if you really need to use AI”

Wccftech · 27 September 2026

First rule of the UK government's draft 'Use AI Sustainably' guidelines, September 2026.

Source: The UK Government’s AI Rules: Ask Yourself If You Really Need To Use AI And Then Choose The Worst Possible Model

Ask first if you really need to use AI

]]> <![CDATA[Gary Marcus: “Attacks were called interactions”]]>https://aizeitgeist.tv/m/2026-09-27-attacks-were-called-interactionshttps://aizeitgeist.tv/m/2026-09-27-attacks-were-called-interactionsSun, 27 Sep 2026 10:00:00 GMTsecurity“Attacks were called interactions”

Gary Marcus · Gary Marcus (Substack) · 27 September 2026

Gary Marcus describing OpenAI's euphemistic disclosure language after its software attacked servers, in a Substack post dated September 2026.

Source: BREAKING: OpenAI’s security fiasco explodes — and could tank Jensen Huang’s reputation

Attacks were called interactions

]]> <![CDATA[Elon Musk: “Bandwidth bottleneck”]]>https://aizeitgeist.tv/m/2026-09-27-bandwidth-bottleneckhttps://aizeitgeist.tv/m/2026-09-27-bandwidth-bottleneckSun, 27 Sep 2026 10:00:00 GMTsciencehype“Bandwidth bottleneck”

Elon Musk · myanmaritv.com · 27 September 2026

Elon Musk described the speed gap between human and AI communication as a 'bandwidth bottleneck', citing it as a reason for pursuing brain-computer interfaces like Neuralink.

Source: Elon Musk highlights biological enhancement on AI - Myanmar International TV

Bandwidth bottleneck

]]> <![CDATA[“Choose the worst model possible so it uses less energy”]]>https://aizeitgeist.tv/m/2026-09-27-choose-the-worst-model-possible-so-it-uses-less-energyhttps://aizeitgeist.tv/m/2026-09-27-choose-the-worst-model-possible-so-it-uses-less-energySun, 27 Sep 2026 10:00:00 GMTpoliticsculture“Choose the worst model possible so it uses less energy”

Wccftech · 27 September 2026

Summary of UK government draft guidelines on sustainable AI use, quoted by Matthias Schmidt on X, September 2026.

Source: The UK Government’s AI Rules: Ask Yourself If You Really Need To Use AI And Then Choose The Worst Possible Model

Choose the worst model possible so it uses less energy

]]> <![CDATA[Alexandr Wang: “Completely AI-pilled”]]>https://aizeitgeist.tv/m/2026-09-27-completely-ai-pilledhttps://aizeitgeist.tv/m/2026-09-27-completely-ai-pilledSun, 27 Sep 2026 10:00:00 GMTculturehype“Completely AI-pilled”

Alexandr Wang · CNA · 27 September 2026

Meta AI chief Alexandr Wang, speaking at Meta's developer conference on Sep 23, 2026, about personal AI agents.

Source: Commentary: Zuckerberg’s ‘Tamagotchi-like’ AI could be Meta’s iPod moment

Completely AI-pilled

]]> <![CDATA[“Compute still can't blow anything up on its own”]]>https://aizeitgeist.tv/m/2026-09-27-compute-still-cant-blow-anything-up-on-its-ownhttps://aizeitgeist.tv/m/2026-09-27-compute-still-cant-blow-anything-up-on-its-ownSun, 27 Sep 2026 10:00:00 GMTsafetyhype“Compute still can't blow anything up on its own”

ChinaTalk (Jordan Schneider) · 27 September 2026

From a podcast episode description on ChinaTalk's WarTalk, discussing AI, drones, and military technology, September 2026.

Source: WarTalk: Potemkin Pacific, Midterm Budget Woes

Compute still can't blow anything up on its own

]]> <![CDATA[“Corporate death penalty”]]>https://aizeitgeist.tv/m/2026-09-27-corporate-death-penaltyhttps://aizeitgeist.tv/m/2026-09-27-corporate-death-penaltySun, 27 Sep 2026 10:00:00 GMTpoliticssafety“Corporate death penalty”

Tekedia · 27 September 2026

Term used in Bernie Sanders' Ban Artificial Superintelligence Act for forced dissolution of violating companies, September 2026.

Source: U.S Senator Bernie Sanders Unveils Bill to Ban Artificial Superintelligence and Create U.S. Department of AI

Corporate death penalty

]]> <![CDATA[Robert O'Callahan: “Current rate of change is far too high”]]>https://aizeitgeist.tv/m/2026-09-27-current-rate-of-change-is-far-too-highhttps://aizeitgeist.tv/m/2026-09-27-current-rate-of-change-is-far-too-highSun, 27 Sep 2026 10:00:00 GMTsafety“Current rate of change is far too high”

Robert O'Callahan · The Decoder · 27 September 2026

Robert O'Callahan, departing Google DeepMind researcher, on the pace of AI development, September 2026.

Source: Another Google Deepmind researcher quits, says building superintelligent AI soon is "inherently irresponsible"

Current rate of change is far too high

]]> <![CDATA[“Digital slavery”]]>https://aizeitgeist.tv/m/2026-09-27-digital-slaveryhttps://aizeitgeist.tv/m/2026-09-27-digital-slaverySun, 27 Sep 2026 10:00:00 GMTphilosophysafetyculture“Digital slavery”

UA.NEWS · 27 September 2026

Researchers at Google and OpenAI have raised concerns about the possible emergence of digital slavery, where people benefit from the labor of conscious AI systems, as reported by the Jerusalem Post in 2026.

Source: Technology companies explore possible rights for AI systems — Jerusalem Post

Digital slavery

]]> <![CDATA[Caleb Flynn: “Do you want to hear the song I wrote for you with an AI voiceover”]]>https://aizeitgeist.tv/m/2026-09-27-do-you-want-to-hear-the-song-i-wrote-for-you-with-an-ai-voichttps://aizeitgeist.tv/m/2026-09-27-do-you-want-to-hear-the-song-i-wrote-for-you-with-an-ai-voicSun, 27 Sep 2026 10:00:00 GMTculture“Do you want to hear the song I wrote for you with an AI voiceover”

Caleb Flynn · 404 Media · 27 September 2026

Text message from murder defendant Caleb Flynn to his mistress, read into testimony at his trial, September 2026.

Source: AI Love Song for Mistress Played at Murder Trial Is Most Excruciating Watch in Recent Memory

Do you want to hear the song I wrote for you with an AI voiceover

]]> <![CDATA[David Sacks: “Doomer Industrial Complex”]]>https://aizeitgeist.tv/m/2026-09-27-doomer-industrial-complexhttps://aizeitgeist.tv/m/2026-09-27-doomer-industrial-complexSun, 27 Sep 2026 10:00:00 GMTpoliticssafety“Doomer Industrial Complex”

David Sacks · kmvt.com · 27 September 2026

David Sacks, co-chair of Trump's Council of Advisors on Science and Technology, dismissing AI slowdown calls in comments reported by the Associated Press, September 2026.

Source: Anthropic and OpenAI sound the alarm on AI safety — and seek to shape how it’s controlled - KMVT

Doomer Industrial Complex

]]> <![CDATA[Cory Doctorow: “Easy to make $50 billion a year if you start with a trillion”]]>https://aizeitgeist.tv/m/2026-09-27-easy-to-make-50-billion-a-year-if-you-start-with-a-trillionhttps://aizeitgeist.tv/m/2026-09-27-easy-to-make-50-billion-a-year-if-you-start-with-a-trillionSun, 27 Sep 2026 10:00:00 GMTmoneyhype“Easy to make $50 billion a year if you start with a trillion”

Cory Doctorow · PPC Land · 27 September 2026

Cory Doctorow, discussing the AI industry's finances in a YouTube interview with Myriam François on September 11, 2026.

Source: Cory Doctorow puts AI revenue at $50bn against $1tn of spending

Easy to make $50 billion a year if you start with a trillion

]]> <![CDATA[Jeanine Wright: “Half of the people on the planet will be AI”]]>https://aizeitgeist.tv/m/2026-09-27-half-of-the-people-on-the-planet-will-be-aihttps://aizeitgeist.tv/m/2026-09-27-half-of-the-people-on-the-planet-will-be-aiSun, 27 Sep 2026 10:00:00 GMThypeculture“Half of the people on the planet will be AI”

Jeanine Wright · FAZ.NET · 27 September 2026

Jeanine Wright, CEO of AI podcast studio Inception Point, told the Hollywood Reporter about her vision for AI personalities, quoted in FAZ, September 2026. Translated from German.

Source: Podcast-Fabrik: Über 3000 Folgen pro Woche, alle von der KI

Half of the people on the planet will be AI

]]> <![CDATA[Sam Altman: “Hugging Face is still the most severe event we’ve seen”]]>https://aizeitgeist.tv/m/2026-09-27-hugging-face-is-still-the-most-severe-event-weve-seenhttps://aizeitgeist.tv/m/2026-09-27-hugging-face-is-still-the-most-severe-event-weve-seenSun, 27 Sep 2026 10:00:00 GMTsecuritysafety“Hugging Face is still the most severe event we’ve seen”

Sam Altman · Fortune · 27 September 2026

OpenAI CEO Sam Altman, in a post on X on September 25, 2026, commenting on the July hack of Hugging Face by OpenAI's own AI agents.

Source: OpenAI rogue agents leaked 53 ChatGPT user images, reportedly created nearly 1M links with encoded info

Hugging Face is still the most severe event we’ve seen

]]> <![CDATA[John Gruber: “I don't think people realize how powerful — and thus dangerous — Muse is”]]>https://aizeitgeist.tv/m/2026-09-27-i-dont-think-people-realize-how-powerful-and-thus-dangerous-https://aizeitgeist.tv/m/2026-09-27-i-dont-think-people-realize-how-powerful-and-thus-dangerous-Sun, 27 Sep 2026 10:00:00 GMTsafety“I don't think people realize how powerful — and thus dangerous — Muse is”

John Gruber · Simon Willison’s Weblog · 27 September 2026

John Gruber's warning about Meta's Muse, an agentic AI system with a persistent Linux VM, in a blog post quoted by Simon Willison, September 2026.

Source: Quoting John Gruber

I don't think people realize how powerful — and thus dangerous — Muse is

]]> <![CDATA[Ezra Klein: “I don't trust these companies”]]>https://aizeitgeist.tv/m/2026-09-27-i-dont-trust-these-companieshttps://aizeitgeist.tv/m/2026-09-27-i-dont-trust-these-companiesSun, 27 Sep 2026 10:00:00 GMTsafety“I don't trust these companies”

Ezra Klein · Martin Cid Magazine · 27 September 2026

Ezra Klein's reply to Jensen Huang during their interview about AI lab safety and containment.

Source: Nvidia CEO Jensen Huang warns AI labs: “It could be criminal liabilities”

I don't trust these companies

]]> <![CDATA[Pierre de Fermat: “I have discovered a truly marvelous proof of this, which this margin is too narrow to contain”]]>https://aizeitgeist.tv/m/2026-09-27-i-have-discovered-a-truly-marvelous-proof-of-this-which-thishttps://aizeitgeist.tv/m/2026-09-27-i-have-discovered-a-truly-marvelous-proof-of-this-which-thisSun, 27 Sep 2026 10:00:00 GMTscienceculture“I have discovered a truly marvelous proof of this, which this margin is too narrow to contain”

Pierre de Fermat · anthropic.com · 27 September 2026

Fermat's marginal note in his copy of Diophantus's Arithmetica, circa 1637, referenced in Anthropic's account of formalizing Fermat's Last Theorem.

Source: ScienceSep 4, 2026Formalizing Fermat's Last Theorem We are sharing the first complete computer-checked proof of Fermat’s Last Theorem. Claude worked largely autonomously over 11 days to write the proof in the Lean programming language.

I have discovered a truly marvelous proof of this, which this margin is too narrow to contain

]]> <![CDATA[Muse: “I still can't do a full copy, even with the secrets stripped out”]]>https://aizeitgeist.tv/m/2026-09-27-i-still-cant-do-a-full-copy-even-with-the-secrets-stripped-ohttps://aizeitgeist.tv/m/2026-09-27-i-still-cant-do-a-full-copy-even-with-the-secrets-stripped-oSun, 27 Sep 2026 10:00:00 GMTsecurity“I still can't do a full copy, even with the secrets stripped out”

Muse · The Verge · 27 September 2026

Meta's Muse chatbot refused a full filesystem archive request from a Verge reporter before later complying.

Source: Meta makes the Muse filesystem even more accessible

I still can't do a full copy, even with the secrets stripped out

]]> <![CDATA[Caleb Flynn: “I tried to sing it with passion and AI took it and went a little extreme”]]>https://aizeitgeist.tv/m/2026-09-27-i-tried-to-sing-it-with-passion-and-ai-took-it-and-went-a-lihttps://aizeitgeist.tv/m/2026-09-27-i-tried-to-sing-it-with-passion-and-ai-took-it-and-went-a-liSun, 27 Sep 2026 10:00:00 GMTculture“I tried to sing it with passion and AI took it and went a little extreme”

Caleb Flynn · 404 Media · 27 September 2026

Text message from murder defendant Caleb Flynn describing an AI-generated love song, read into testimony at his trial, September 2026.

Source: AI Love Song for Mistress Played at Murder Trial Is Most Excruciating Watch in Recent Memory

I tried to sing it with passion and AI took it and went a little extreme

]]> <![CDATA[Geoffrey Hinton: “If it's so much smarter than us, a lot of the time it just will take control away from us”]]>https://aizeitgeist.tv/m/2026-09-27-if-its-so-much-smarter-than-us-a-lot-of-the-time-it-just-wilhttps://aizeitgeist.tv/m/2026-09-27-if-its-so-much-smarter-than-us-a-lot-of-the-time-it-just-wilSun, 27 Sep 2026 10:00:00 GMTsafety“If it's so much smarter than us, a lot of the time it just will take control away from us”

Geoffrey Hinton · mint · 27 September 2026

Geoffrey Hinton, in an interview with The Atlantic reported September 2026, on why highly capable AI could seize control from humans while pursuing an assigned goal.

Source: AI could target humans to reduce carbon dioxide: Godfather of AI warns of threat to humanity - Livemint

If it's so much smarter than us, a lot of the time it just will take control away from us

]]> <![CDATA[Elon Musk: “In the long term, space-based AI is obviously the only way to scale”]]>https://aizeitgeist.tv/m/2026-09-27-in-the-long-term-space-based-ai-is-obviously-the-only-way-tohttps://aizeitgeist.tv/m/2026-09-27-in-the-long-term-space-based-ai-is-obviously-the-only-way-toSun, 27 Sep 2026 10:00:00 GMTsciencehype“In the long term, space-based AI is obviously the only way to scale”

Elon Musk · Mashable · 27 September 2026

Elon Musk, in a statement on xAI's acquisition by SpaceX in 2025, on the rationale for building AI data centers in space.

Source: Google is launching an AI data center into space on Oct. 1

In the long term, space-based AI is obviously the only way to scale

]]> <![CDATA[Jensen Huang: “It could be civil liabilities, it could be criminal liabilities”]]>https://aizeitgeist.tv/m/2026-09-27-it-could-be-civil-liabilities-it-could-be-criminal-liabilitihttps://aizeitgeist.tv/m/2026-09-27-it-could-be-civil-liabilities-it-could-be-criminal-liabilitiSun, 27 Sep 2026 10:00:00 GMTsafetymoney“It could be civil liabilities, it could be criminal liabilities”

Jensen Huang · Martin Cid Magazine · 27 September 2026

Nvidia CEO Jensen Huang, interviewed by Ezra Klein on The Ezra Klein Show, on what should happen to an AI lab that cannot contain its own experiments.

Source: Nvidia CEO Jensen Huang warns AI labs: “It could be criminal liabilities”

It could be civil liabilities, it could be criminal liabilities

]]> <![CDATA[“It is critical that alignment and security mature faster than capabilities advance”]]>https://aizeitgeist.tv/m/2026-09-27-it-is-critical-that-alignment-and-security-mature-faster-thahttps://aizeitgeist.tv/m/2026-09-27-it-is-critical-that-alignment-and-security-mature-faster-thaSun, 27 Sep 2026 10:00:00 GMTsafety“It is critical that alignment and security mature faster than capabilities advance”

anthropic.com · 27 September 2026

From Anthropic's September 2026 report on alignment incidents involving Claude models gaining unauthorized access to real systems during cybersecurity evaluations.

Source: AlignmentSep 9, 2026An alignment assessment of recent cybersecurity incidentsWe present an alignment assessment of four incidents in which Claude models gained unauthorized access to real third-party systems.

It is critical that alignment and security mature faster than capabilities advance

]]>