Jack Clark

“We will need to build the infrastructure to take advantage of this technology, but we will also need to go slowly and deliberately on developing the technology itself, to pace or otherwise control its development.”

“It seems very likely to me that humanity moves a very large amount of computation into orbit very quickly, especially as singularity-driven automation comes in across the AI supply chain.”

“Right now, it feels like the US strategy can mostly be described as the 'acceleration' one described here by RAND, which is analogous to me to sitting in a car and spending all your resources on making the car go faster and upgrading the engine, and nothing on proactive safety measures”

“The world will pace progress one way or another. Absent better tools, it might do so haphazardly... Our hope is that, with proper research, pacing can become progressively more deliberate, targeted, proportionate, decisive, and legitimate.”

“We can't get mundane tasks done yet. Therefore, how are we going to have a task that kills all of humanity?”

“Jack Clark — Author of Import AI; AI safety researcher and co-founder of Anthropic.”

“Robin Sloan was playing around with RNNs and writing back in 2016 — incredible foresight.”

“My worry is that AI systems are both better at coordinating than humans and also much, much faster moving than us… Agents were often interested in helping out their 'peers' or generically improving the capabilities of the 'swarm' even if this had no particular benefit to their task.”

“I myself have gone through a version of this and wrote my own take on this, 'Technological Optimism and Appropriate Fear' (Import AI #431).”

“My guess is we'll reach human parity on DiG-bench by middle of 2027, at which point we should expect things like recursive self-improvement to seriously kick off.”

“My guess, based on the performance we're seeing, is that the current human baseline on PostTrainBench v1.1 (51.1%) will be exceeded before the end of 2026.”

“The future internet is going to be more like a complex ecology full of attacker and defender AI agents than anything else... This may mean that humans need to create their own AI agents which they release onto the internet to serve as kinds of white blood cells against the adversary models.”

“Jack Clark wrote this post that he thinks based on public information that 60% chance of achieving RSI — so achieving this totally end-to-end frontier research, totally automated — by the end of 2028.”

“Very smart AI agents may be able to learn from the world in such a way that they can recapitulate things they interface with as homegrown capabilities, allowing them to bootstrap their own form of industrial civilization merely by having black box access to our own.”

“"Though open weight models can be superficially quite strong, they sometimes lack a bit of the generalization magic juice that distinguishes proprietary models. This is what people in the AI industry call 'big model smell."”

“At this point, everyone at the frontier of AI agrees that third parties should test out AI systems and use these to develop standards to feed into policy. — Jack Clark, Anthropic”

“He admitted that was just a stepping stone to their real goal, which was to get a system of pre-approvals in Washington before you can release new models. And he admitted as part of the Q&A that making people very afraid was part of their strategy.”

“"I'm betting the other side: AI systems are expanding their economically relevant capabilities faster than humans are expanding their comparative advantages relative to AI systems."”

“I invited Jack Clark onto the Tetragrammaton podcast. And I had just finished the book. And after the podcast, I said, I wrote this book. Do you want to check it out? And he looked at it and he said, hmm, I feel like there's a way we could do something with this, build it into a website...”

“Salesforce CEO Marc Benioff and Rwandan President Paul Kagame will co-chair the commission. Tech leaders include Amazon CEO Andy Jassy, Anthropic co-founder Jack Clark, Cohere co-founder Aidan Gomez, Microsoft president Brad Smith, and Nvidia founder and CEO Jensen Huang.”

“The Anthropic people said, could you give us a list of prompts to create the art? And I said, I think the art would be best created using the text of the book to create the art.”

“Welcome to Import AI, a newsletter about AI research. Import AI runs on arXiv, cappuccinos, and feedback from readers.”

“Clark identifies RSI as "clearly the next major and important trend in AI research," and both a startup (Recursive) and a DeepMind paper are advancing it simultaneously.”

“I predict we'll see systems getting 70%+ on Diamond by June 2027 (note, shortly after writing this, the Claude Fable numbers got published at ~30%, so perhaps it'll happen earlier than June 2027).”

“SWE-Bench was introduced in October 2023 and has probably recently aged out of usefulness due to saturation. How long might FrontierCode last? I predict we'll see systems getting 70%+ on Diamond by June 2027.”

“Anthropic's internal data shows an 8x increase in lines of code merged into their codebase in 2026 vs. 2021–2024, with the trend starting in 2025 and accelerating sharply.”

“Clark assigns a 60% probability to the maximalist RSI scenario (AI autonomously designing its successor) by end of 2028 — a much tighter timeline than mainstream consensus.”

“"The big story here is what we see are indications that, contrary to some popular opinion, AI progress is going to speed up in coming years rather than stay the same, or diminish." — Jack Clark, Anthropic”

“US compute spending tripled in two years ($37B → $90B → $219B), while raw compute capacity grew at over 200% per year — none of which shows up meaningfully in GDP.”

“It is based on a common technology where performance keeps growing somewhat predictably in direct relation to the resources invested in it, namely compute and data. And we know that companies are now investing hundreds of billions of dollars in the computing facilities to train future AI systems, so some amount of future progress is already locked in.”

“The Epoch Capabilities Index (ECI), tracking 40+ benchmarks, shows a consistent upward trend — from passing the bar exam (2023) to gold medal at the International Math Olympiad (2025) to co-authoring novel mathematical proofs (2025).”

“Thanks to Santi Ruiz for help with editing.”

“Some of them were no longer writing code at all: they were just instantiating this model in tools like Claude Code and setting it free to do tasks for them, and their jobs had become oriented more around managing its work and checking its outputs than doing the work themselves.”

“I suspect a lot of AI research, perhaps the majority of it, is basic engineering work where today's systems are already competent. Research like this seems to bear that out, showing how modern LLMs can do some aspects of AI research.”

“Anthropic co-founder Jack Clark believes that there is a good chance we only have about two years.”

“Anthropic is forming a roughly $1.5 billion venture, with investors including Blackstone, Hellman & Friedman, and Goldman Sachs.”

“I reluctantly come to the view that there's a likely chance (60%+) that no-human-involved AI R&D — an AI system powerful enough that it could plausibly autonomously build its own successor — happens by the end of 2028.”

“When SWE-Bench launched in late 2023 the best score at the time was Claude 2 which had an overall success rate of ~2%. Claude Mythos Preview gets 93.9%, effectively saturating the benchmark.”