Daniel Kokotajlo, one of the authors of the AI 2027 scenario, now puts his median date for a Superhuman AI Researcher at May 2027. In mid-August he put it at December 2028, so his forecast moved 19 months earlier, by our arithmetic. The forecasts are the group’s own and assume no deliberate slowdown.
The AI Futures Project post, published at 16:15 UTC on 10 November, says “all of us forecast that Superhuman AI Researcher (SAR) will likely arrive at some point in 2027”. It names one input above the rest: “the single biggest source was the METR report” on Anthropic.
Business Pill · AI Agents Explained
An evergreen Business Engineer explainer on AI agents. It teaches the general idea only and does not discuss the AI Futures Project forecasts.
The key insight: Daniel Kokotajlo’s median for a Superhuman AI Researcher moved from December 2028 to May 2027, 19 months earlier, by our arithmetic. The six forecasters in the AI Futures Project update now give medians between May and November 2027.
The post defines SAR as “roughly the point at which a company would rather fire all their human researchers and rely solely on automated AI researchers than have a team of the best human researchers in the world with no AIs”.
Six forecasters give SAR medians: Daniel May 2027, Eli and Romeo August 2027, Brendan September 2027, Lukas and Thomas November 2027. That is 6 to 12 months after November 2026, by our arithmetic. The group also gives “a 35-65% chance” that Top-Expert-Dominating AI arrives before 2028.
For that later milestone, the post’s appendix says Kokotajlo’s median moved to September 2027 “(from late 2028)” and Eli Lifland’s to September 2028 “(from mid-2032)”. The post carries Lifland’s byline; both are among the authors listed on the AI 2027 site.
Each AI Futures Project forecaster’s median date for a Superhuman AI Researcher, from the group’s 10 November update (its own forecasts, assuming no deliberate slowdown). Months after November 2026 are our arithmetic.
What Was Posted on X
The AI Futures Project have updated their timelines. They say we are five months ahead of AI 2027, and they now believe a Superhuman AI Researcher will arrive sometime in 2027.
Andrew Curran posted at 17:05 UTC on 10 November that the AI Futures Project had updated its timelines, saying the group puts current progress five months ahead of AI 2027 and now expects a Superhuman AI Researcher sometime in 2027. This piece draws its forecasts and quotes from the group’s own post.
They now predict that 95% of Anthropic’s R&D will be fully automated by November 2027. This is what Anthropic describes as AL5, fully autonomous RSI. This would have many implications for policy, as this will eventually happen at every lab, not just Anthropic
Andrew Curran posted at 18:19 UTC on 10 November that the group now predicts 95% of Anthropic’s R&D will be fully automated by November 2027, which he says is what Anthropic describes as AL5, fully autonomous RSI, and that this would have many implications for policy. The update box further down sets out where the figure sits in the group’s own post.
The researchers behind AI 2027 have sharply shortened their timelines.
Daniel Kokotajlo has moved his median forecast for a superhuman AI researcher from December 2028 to May 2027, a 19-month shift.
All six researchers now expect such AI to likely arrive in 2027, assuming no substantial slowdown. Their single biggest reason is METR’s recent assessment of AI-driven research acceleration at Anthropic.
Choblin posted at 20:16 UTC on 10 November that the researchers behind AI 2027 have shortened their timelines: Daniel Kokotajlo moved his median for a superhuman AI researcher from December 2028 to May 2027, a 19-month shift, and all six researchers now expect it likely in 2027, assuming no substantial slowdown, with METR’s assessment of research acceleration at Anthropic as the biggest reason. This piece takes the medians from the group’s update.
Why They Moved: METR and Anthropic
The group says METR estimated that “Anthropic is now making AI progress about twice as fast as its historical rate”, with about 80% of that put down to AI. The group turns this into “a 1.7x AI-attributable acceleration” and says it is “5 months ahead of the AI 2027 scenario”, which had that pace in roughly April 2027.
The METR estimate fits trends to AECI, which the post calls “Anthropicโs internal version of Epochโs Capabilities Index (ECI)”; we covered the public Epoch ECI record for Claude Fable 5.2. The trend after October 2025 has “1.9x the slope” of the one before, the post says.
The group also notes METR’s line that “Anthropic redacted a majority of the evidence we found most informative for these conclusions”. It quotes the Fable 5.2 system card, where Anthropic says Claude Mythos 5.2 “does not double the internal rate of overall aggregate progress in AI research and development”. Separately, Anthropic’s IPO filing now says it met its AI R&D threshold.
Update (10 Nov): where the “95% by November 2027” figure comes from.
Andrew Curran posted that the group now predicts “95% of Anthropic’s R&D will be fully automated by November 2027”. In the AI Futures post, that date comes from one of several methods: a “logistic extrapolation” of the chart Anthropic shared of R&D tasks by automation level, with an assumption that each automation level begins registering about seven months after the previous, “unfortunately, based on little data”.
The post then says: “Assuming SAR comes when 95% of tasks are AL5, this predicts SAR in Nov 2027”. It adds that there is “high uncertainty”, especially around how far AL5 will lag behind AL4 and how the share of AL5 tasks relates to when SAR is achieved.
As we read the post, the 95% is the threshold the method assumes for a Superhuman AI Researcher inside one extrapolation, not a separate forecast from the group.
What the Group Asks For
The post argues that “If SAR is coming in 2027, it is incredibly urgent to pause or substantially โpaceโ frontier AI development”, ideally through an industry-wide pause and an international agreement. That is the group’s position, not a finding.
The Structural Read
As we read the update, the size of the move rests mostly on one estimate: METR’s 1.7x AI-attributable acceleration at Anthropic, much of whose evidence was redacted. The group’s own forecasts say so, and the dates are best read as their reading of that estimate.
AI Futures Project, Q3.5 Timelines Update, 10 November 2026
“Now, all of us forecast that Superhuman AI Researcher (SAR) will likely arrive at some point in 2027 (assuming no substantial deliberate slowdown).”
Three Implications
FOR READERS OF AI 2027 The group says the AI-driven acceleration it infers is about 5 months ahead of the AI 2027 scenario.
FOR ANTHROPIC WATCHERS The group asks Anthropic to release enough information for external observers to assess METR’s estimate; the post notes most of the key evidence was redacted.
FOR READING THE DATES Every date here is a median forecast that assumes no substantial deliberate slowdown; the group gives SAR medians from May to November 2027.
The framework reads: “The frontier moat has shifted.”
As we read it, the METR estimate behind this update concerns the same loop: models helping build the next frontier model. The framework predates this update and does not discuss these forecasts.
What Is Not Established
These are forecasts, not measurements. The post says “Itโs possible that there are some errors in the modeling”, and that much of METR’s evidence is redacted. Kokotajlo writes of the METR report: “I still donโt trust it completely, of course”. The post calls the system card quote “ambiguous as to whether they agree with the METR reportโs estimate”.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
The AI Futures Project now forecasts a Superhuman AI Researcher in 2027, with medians from May to November 2027. Kokotajlo’s own median moved 19 months earlier, from December 2028 to May 2027, by our arithmetic, mainly on METR’s report on Anthropic.
A note on sourcing. Forecasts, quotations and dates are from the AI Futures Project’s post of 10 November 2026; author names are from the AI 2027 site. The forecasts are the group’s own, not measurements. Month counts are our arithmetic.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
