AI robotics spark stock rally
3m
Analysts predicted a stock-market rally as AI moved out of data centers into physical robotics, while some Austrian real-estate firms experimented with AI only cautiously.
Chad Jones has a useful way to think about AI safety:
If AI poses a 1% chance of catastrophe, and we value a statistical life at ~$10M, reducing that risk by 1 percentage point is worth about $100,000 per person. https://newsletter.platypuseconomics.com/p/if-we-pay-for-safer-cars-why-wont
If AI poses a 1% chance of catastrophe, and we value a statistical life at ~$10M, reducing that risk by 1 percentage point is worth about $100,000 per person. https://newsletter.platypuseconomics.com/p/if-we-pay-for-safer-cars-why-wont
How Much Would You Pay to Prevent the AI Apocalypse?
And what economists say we should spend. Hint: we're nowhere close.
newsletter.platypuseconomics.com
September 23, 2026 at 7:15 PM
Chad Jones has a useful way to think about AI safety:
If AI poses a 1% chance of catastrophe, and we value a statistical life at ~$10M, reducing that risk by 1 percentage point is worth about $100,000 per person. https://newsletter.platypuseconomics.com/p/if-we-pay-for-safer-cars-why-wont
If AI poses a 1% chance of catastrophe, and we value a statistical life at ~$10M, reducing that risk by 1 percentage point is worth about $100,000 per person. https://newsletter.platypuseconomics.com/p/if-we-pay-for-safer-cars-why-wont
BREAKING: Heads of major AI firms warn the UN that without intervention, "AI could be a risk to humanity as a whole".
Heads of AI firms tell UN Security Council that it could be a risk to all humanity
The heads of major artificial intelligence firms pleaded with the United Nations to set controls over the burgeoning technology they are designing in order to save the world — or at least its people.
bit.ly
September 23, 2026 at 8:02 PM
BREAKING: Heads of major AI firms warn the UN that without intervention, "AI could be a risk to humanity as a whole".
Reposted by Simon Lester, Chris Hanretty, Christian Odendahl
My latest for @economist.com: AI writing is taking over politics.
10% of what was is said in Parliament is now AI-drafted, per Pangram. Yet more in the House of Representatives, or in Australia and Canada.
Link: www.economist.com/britain/202...
And a 🧵 below, w/ some main offenders
10% of what was is said in Parliament is now AI-drafted, per Pangram. Yet more in the House of Representatives, or in Australia and Canada.
Link: www.economist.com/britain/202...
And a 🧵 below, w/ some main offenders
September 23, 2026 at 7:50 PM
My latest for @economist.com: AI writing is taking over politics.
10% of what was is said in Parliament is now AI-drafted, per Pangram. Yet more in the House of Representatives, or in Australia and Canada.
Link: www.economist.com/britain/202...
And a 🧵 below, w/ some main offenders
10% of what was is said in Parliament is now AI-drafted, per Pangram. Yet more in the House of Representatives, or in Australia and Canada.
Link: www.economist.com/britain/202...
And a 🧵 below, w/ some main offenders
Reposted by Fabián Muniesa, Michael Larkin
There seems to be a coordinated talking point among the EA folks that the "standard" approach to "AI" on "the Left" is "AI is just stochastic parrots so no need to worry about it".
(Sorry for all the scare quotes, but y'know)
>>
(Sorry for all the scare quotes, but y'know)
>>
September 23, 2026 at 1:03 PM
There seems to be a coordinated talking point among the EA folks that the "standard" approach to "AI" on "the Left" is "AI is just stochastic parrots so no need to worry about it".
(Sorry for all the scare quotes, but y'know)
>>
(Sorry for all the scare quotes, but y'know)
>>
The AI Productivity Delusion: As with previous tech revolutions, it is natural to focus on AI's net effects: Will it create more than it destroys? But unlike previous eras of creative destruction, the real question this time is, “Create more of what?” www.project-syndicate.org/commentary/a...
The AI Productivity Delusion
Yanis Varoufakis sees the technology converting markets into fiefs and profits into rents, ending capitalism as we know it.
www.project-syndicate.org
September 23, 2026 at 11:22 AM
The AI Productivity Delusion: As with previous tech revolutions, it is natural to focus on AI's net effects: Will it create more than it destroys? But unlike previous eras of creative destruction, the real question this time is, “Create more of what?” www.project-syndicate.org/commentary/a...
Just got my first AI slop paper to review & am totally ripshit per @novosad.bsky.social - what a waste of time! AI is amazing but can destroy science while revolutionizing it.
Editors: We are looking for journal partners to study how to use AI to make scientific publishing better (not worse). LMK!
Editors: We are looking for journal partners to study how to use AI to make scientific publishing better (not worse). LMK!
September 23, 2026 at 10:24 PM
Just got my first AI slop paper to review & am totally ripshit per @novosad.bsky.social - what a waste of time! AI is amazing but can destroy science while revolutionizing it.
Editors: We are looking for journal partners to study how to use AI to make scientific publishing better (not worse). LMK!
Editors: We are looking for journal partners to study how to use AI to make scientific publishing better (not worse). LMK!
Nice reality check on the AI hype and why the techbros promote this BS: www.technologyreview.com/2026/09/22/1...
Don’t be fooled by this summer of AI hype
Breathless claims about AGI and new capabilities fall apart pretty quickly under scrutiny.
www.technologyreview.com
September 23, 2026 at 3:39 AM
Nice reality check on the AI hype and why the techbros promote this BS: www.technologyreview.com/2026/09/22/1...
@jamellebouie.net draws appropriate parallels between AI in the 21st century and railroads in the 19th: “If A.I. promises to transform the entire society, then the entire society is entitled to have a say, even a decisive one, over the use, development and direction of the technology.”
Opinion | New Matrix, Meet the Old Matrix (Gift Article)
The railroads were once the octopus that A.I. is now.
www.nytimes.com
September 23, 2026 at 10:21 AM
@jamellebouie.net draws appropriate parallels between AI in the 21st century and railroads in the 19th: “If A.I. promises to transform the entire society, then the entire society is entitled to have a say, even a decisive one, over the use, development and direction of the technology.”
by Emily Grubert Reposted by Kathleen Kennedy, Olúfẹ́mi Táíwò
I think I’ve underestimated how radical it is to be an “actually, we don’t need more data” guy in the age of AI
September 23, 2026 at 1:17 AM
I think I’ve underestimated how radical it is to be an “actually, we don’t need more data” guy in the age of AI
In @fortune.com - The backwards AI pacing debate & how far business is from the frontier - DC & Silicon Valley have found a new fight over AI in pacing...Corporate USA is yrs behind the AI frontier & labs’ fortunes will be decided by trust & adoption, not raw capability.
fortune.com/2026/09/22/a...
fortune.com/2026/09/22/a...
The backwards AI pacing debate and how far business is from the frontier | Fortune
In every tech revolution, the greatest fortunes went to those who understood that a frontier is worthless until the settlers have arrived.
fortune.com
September 23, 2026 at 12:35 AM
In @fortune.com - The backwards AI pacing debate & how far business is from the frontier - DC & Silicon Valley have found a new fight over AI in pacing...Corporate USA is yrs behind the AI frontier & labs’ fortunes will be decided by trust & adoption, not raw capability.
fortune.com/2026/09/22/a...
fortune.com/2026/09/22/a...
Amongst all the deluge of stuff I have seen about AI today, this was the most enlightening and enjoyable to read.
September 23, 2026 at 4:32 PM
Amongst all the deluge of stuff I have seen about AI today, this was the most enlightening and enjoyable to read.
The question for AI is to what extend returns are increasing. A new paper by Tom Cunningham et al. puts a number on it.
arxiv.org/abs/2609.15802
This has big implications for law. One does not regulate self-improved systems as others, see papers.ssrn.com/sol3/papers....
arxiv.org/abs/2609.15802
This has big implications for law. One does not regulate self-improved systems as others, see papers.ssrn.com/sol3/papers....
September 23, 2026 at 12:03 PM
The question for AI is to what extend returns are increasing. A new paper by Tom Cunningham et al. puts a number on it.
arxiv.org/abs/2609.15802
This has big implications for law. One does not regulate self-improved systems as others, see papers.ssrn.com/sol3/papers....
arxiv.org/abs/2609.15802
This has big implications for law. One does not regulate self-improved systems as others, see papers.ssrn.com/sol3/papers....
Reposted by William J. Brady, Laurent Dubois, Olúfẹ́mi Táíwò
AI is the Gyges' ring of student work: it offers perfect concealment of wrongdoing.
If Glaucon is right, then everyone will use it to do injustice (cheat), & will be wise to do so.
If Socrates/Plato is right, then doing so also harms the wrong-doer, whose soul will be malformed into misery by it.
If Glaucon is right, then everyone will use it to do injustice (cheat), & will be wise to do so.
If Socrates/Plato is right, then doing so also harms the wrong-doer, whose soul will be malformed into misery by it.
September 23, 2026 at 12:43 PM
AI is the Gyges' ring of student work: it offers perfect concealment of wrongdoing.
If Glaucon is right, then everyone will use it to do injustice (cheat), & will be wise to do so.
If Socrates/Plato is right, then doing so also harms the wrong-doer, whose soul will be malformed into misery by it.
If Glaucon is right, then everyone will use it to do injustice (cheat), & will be wise to do so.
If Socrates/Plato is right, then doing so also harms the wrong-doer, whose soul will be malformed into misery by it.
Reposted by Lisa Janicke Hinchliffe
Something that worries me more every day: I suspect most fields have 5-25 breakthrough ideas latent in existing data and literature. Say you’re a midsize AI company looking for cred. Why not hire a *small* team of domain experts + 1000 agents … and promote your models as the ones that broke Econ?
Anthropic threw Claude at a DNA database. 21 hours, 950 agents, and 210mm tokens later, it emerged with a novel reverse transcriptase (similar to CRISPR). They're running a non-pathogenic (BSL-1/2) research lab in the Bay Area.
Claude discovers a novel enzyme system with CRISPR-like repeats
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
www.anthropic.com
September 23, 2026 at 9:57 PM
Something that worries me more every day: I suspect most fields have 5-25 breakthrough ideas latent in existing data and literature. Say you’re a midsize AI company looking for cred. Why not hire a *small* team of domain experts + 1000 agents … and promote your models as the ones that broke Econ?
Reposted by Robert T. Tally
As a result of @benjaminjriley.bsky.social's most recent Alpha School reporting, all who have been saying "AI has a Mechanical Turk problem" - that is, "artificial intelligence" is actually the cognitive labor of humans in the Global South - get to take another victory lap.
They must be winded.
They must be winded.
![They hired this data team of people that didn't live in the U.S. and they would watch us through our computers and it felt really sketchy and scary. It felt like it was more of a science experiment, and I was more of a lab rat rather than [having] someone that actually cared about my learning.
I was getting really frustrated about that throughout the year. Every time they would change something, I would be like, why are we changing this?
We couldn't tape over the computer or anything like that. They would get really frustrated if we taped over it. We had to keep the camera on.
[The software] would also track your facial expressions, and be like, "This is the emotion that you're evoking while you're talking about this." And every single time, it would tell me that I look really sad, which is fair because / was really sad, because / was crying every single day.
They would make me and a few other kids take standardized math tests pretty much every day, and then any question that we got wrong, we would have to get on a Zoom call with some random guy from normally India and explain to him every single thing that we got wrong and why. And I hated that.
I was a guinea pig, and I knew that I was a guinea pig. They were not subtle about it.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:776hcssquwcv2ihpptzvttxj/bafkreihcxzi7znw2yq5ux6mhcq5zbfpeydb5ys2syjpi7ojhhxiyg7l7zm)
September 23, 2026 at 5:19 PM
As a result of @benjaminjriley.bsky.social's most recent Alpha School reporting, all who have been saying "AI has a Mechanical Turk problem" - that is, "artificial intelligence" is actually the cognitive labor of humans in the Global South - get to take another victory lap.
They must be winded.
They must be winded.
Reposted by Henry Jones
The people who built AI with seemingly little care for the havoc it could cause and no guardrails to protect public should be held accountable for every problem AI is now causing. They want to make billions but don’t want to accept responsibility for their technology.
www.abc.net.au/news/2026-09...
www.abc.net.au/news/2026-09...
Breaking: OpenAI agent hacked Medicare portal, PM says
Prime Minister Anthony Albanese says an Open AI agent gained unauthorised accessed to the Medicare statistics reporting service portal.
www.abc.net.au
September 23, 2026 at 8:55 PM
The people who built AI with seemingly little care for the havoc it could cause and no guardrails to protect public should be held accountable for every problem AI is now causing. They want to make billions but don’t want to accept responsibility for their technology.
www.abc.net.au/news/2026-09...
www.abc.net.au/news/2026-09...
Reposted by Gilles Louppe
incredible progress in the field of AI
September 23, 2026 at 6:07 AM
incredible progress in the field of AI
by Zen Faulkes
Thinking about how some scientists might support AI 🤖 for research because they see it as a way to circumvent journal paywalls.
Haven’t seen anyone say this specifically, but I wonder if any are thinking 💭 along those lines.
Haven’t seen anyone say this specifically, but I wonder if any are thinking 💭 along those lines.
September 23, 2026 at 8:26 PM
Thinking about how some scientists might support AI 🤖 for research because they see it as a way to circumvent journal paywalls.
Haven’t seen anyone say this specifically, but I wonder if any are thinking 💭 along those lines.
Haven’t seen anyone say this specifically, but I wonder if any are thinking 💭 along those lines.
Interesting that loss of public support was most linked to AI ‘idea generation’ — suggests it is not just about people’s distrust of AI output, but their sense that science loses public value when torn from its root in human imagination and curiosity
Public support for research falls sharply when it uses AI, study finds.
Scepticism over artificial intelligence in taxpayer-funded research “demonstrates fragility” of support within UK.
www.researchprofessionalnews.com/rr-news-uk-i...
Scepticism over artificial intelligence in taxpayer-funded research “demonstrates fragility” of support within UK.
www.researchprofessionalnews.com/rr-news-uk-i...
Public support for research falls sharply when it uses AI, study finds - Research Professional News
Scepticism over artificial intelligence in taxpayer-funded research “demonstrates fragility” of support within UK
www.researchprofessionalnews.com
September 23, 2026 at 12:34 PM
Interesting that loss of public support was most linked to AI ‘idea generation’ — suggests it is not just about people’s distrust of AI output, but their sense that science loses public value when torn from its root in human imagination and curiosity