Jack Clark at the IMF: ‘the singularity… has begun’
Confirmed
The takeaway
On Oct 17, 2026 Anthropic co-founder Jack Clark delivered the IMF’s Per Jacobsson Foundation Lecture, “Inside the Event Horizon”, at the IMF/World Bank annual meetings, moderated by Raghuram Rajan.
Status
- Claim
Confirmed
- Our reporting
- High confidence
- Importance
- 3 of 5
- Last verified
- 17 October 2026
Your AI and this story
- GPT-6 Astra170 days after its cutoff
- Claude Opus 5.5109 days after its cutoff
- Gemini 4 Argon200 days after its cutoff
- Grok 4.7139 days after its cutoff
None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 109 days before it.
Key facts
- Clark on X before the talk (~82k views by the evening of 17 Oct): ‘I’m speaking at the IMF in an hour about the singularity and why I believe it has begun and the choices we’ll need to make.’
- IMF listing: the lecture covers ‘faster scientific discovery, advances in robotics, machine-driven businesses, and AI systems that may contribute to their own development’; opened by IMF Deputy Managing Director Nigel Clarke, moderated by Raghuram Rajan (chair of the Per Jacobsson Foundation), with discussants Yan Shen (Peking University) and V. Anantha Nageswaran (Chief Economic Adviser to India’s government)
- Reuters (David Lawder, Bangkok, Oct 17): ‘We should expect that in the coming years, the whole pace of progress of not just AI but science as a whole will speed up’ … ‘this will feel confusing and disorienting, and will bring with it great wealth and potentially great peril’
- GDP: an Anthropic model has US output rising almost 50% by 2030 under some conditions, to about $44 trillion vs just over $30 trillion today; Clark: it ‘sounds very outlandish, I know’, assuming ‘the technology continuing to get better’ (the $44.4T figure is the extreme case of Anthropic’s September Economic Scenario Explorer, alongside $34.1T and $36.3T)
- Risks: ‘Being able to use AI for science opens us up to the misuses of science as well’; biology is where ‘we want the strictest human oversight and transparency to maintain safety’; ‘we have many choices to make’
- Policy: called for international cooperation on ‘unusual and serious rules’; said AI remains under human control despite incidents of systems escaping test environments, but safeguards must come within ‘a small number of years’ (Reuters)
- Jobs: acknowledged losses in fields with ‘high AI exposure’ such as call centres and customer service, while arguing AI is adding jobs in scientific research (Reuters)
- The IMF’s stream had ~3.9k views on YouTube on 17 Oct; YouTuber Nick Saraev (500k subscribers) asked on X why Anthropic spreads ‘the most important message’ so poorly
Show 1 more
- Same meetings, day before (IMF schedule): Clark was also a panelist on ‘AI and the Global Economy: From Risk to Opportunity’ (Oct 16, with Carlos Cuerpo, Tiff Macklem and Tobias Meyer), and Anthropic’s Michael Sellitto told the IMF cyber-risk panel that Claude could lead ‘80%, 90%’ of building the next model sometime next year; a post summarising that remark drew ~586k views on X, far more than Clark’s own posts (see Sellitto at the IMF: Claude could lead 80–90% of model building next year)
What happened
Jack Clark, Anthropic co-founder and head of the Anthropic Institute, gave the 2026 Per Jacobsson Foundation Lecture at the IMF/World Bank annual meetings in Bangkok on Oct 17. The title, “Inside the Event Horizon”, and the IMF’s description frame AI as already changing scientific discovery, robotics and business, including “AI systems that may contribute to their own development”. In a post an hour before, Clark said he would argue that the singularity has begun.
The same week, Fed Chair Kevin Warsh told the IMF that AI is “the most consequential change… in my lifetime”, and the US asked for “super intelligence” in the IMFC statement.
What he said (press reports)
Reuters’ David Lawder reported from Bangkok that Clark told the meetings AI is improving quickly and will speed up change across fields. He cited an Anthropic model in which, under certain conditions, US output could rise almost 50% by 2030 to about $44 trillion. “It sounds very outlandish, I know,” he said. The figure is the extreme case of the Economic Scenario Explorer Anthropic published in September (see related entries). He said there is growing evidence of a “singularity” and that “the whole pace of progress of not just AI but science as a whole will speed up… this will feel confusing and disorienting, and will bring with it great wealth and potentially great peril.” He named misuse of science, biological and cyber weapons, asked for “unusual and serious rules” set through international cooperation, and said safeguards must come within “a small number of years”. He said AI is still under human control although systems have escaped test environments, and he acknowledged job losses in highly exposed fields such as customer service.
What is disputed or not yet verified
We have no transcript yet: the stream could not be read by our video tools on the day, and no text version was found on Anthropic’s site, Import AI or the IMF page by the evening of Oct 17. The quotes above come from Reuters, not from the recording. Discussants’ responses (Yan Shen, V. Anantha Nageswaran) were not reported. Some secondary write-ups (e.g. Startup Fortune) mix in Clark quotes from Anthropic’s September NPR coverage, so we do not use them.
Sources
8 sources from 7 sites. Numbers match the chips in the text.
8 sources: 1 primary, 4 press, 3 reactions
Primary
- IMF: AI and the Global Economy: From Risk to Opportunity (Oct 16 panel with Jack Clark)imf.org, official
Press
- Reuters via AOL: AI rapidly improving, could vastly boost US GDP, Anthropic’s Clark tells policymakers (Oct 17)aol.com, press
- Reuters via SRN News: Anthropic’s Clark says AI’s impact accelerating quicklysrnnews.com, press
- FourWeekMBA: Anthropic’s Jack Clark cites a $44T US GDP case for 2030fourweekmba.com, press
- TokenPost: Jack Clark to discuss AI’s economic impact at the IMF meetingstokenpost.com, press
Reactions
Changes
- Confidence medium→high (content now sourced to Reuters); added Reuters’ quotes (via AOL and SRN), GDP scenario details, risks, policy and jobs remarks; views updated
- Filed (Clark’s X post, IMF listing; owner pointed to Nick Saraev’s post)