Jeff Clune

"stochastic parrot" was a mimetically-fit cognitive virus that spread from 2021-2025; it temporarily blinded many gifted people to the nature of AI progress, burning up crucial years in which they could have helped think through the response to the situation.

OpenAI's Noam Brown says air-gapping the computers may not stop a misaligned AI, because two air-gapped machines can still talk by running a CPU hot and reading the temperature change "But I think the major takeaway from the incident is that people underestimated the AI. And we never want to be in a situation again where we underestimate the AI. It's a weird world, because AI progress is so fast that people are consistently underestimating the AI." "So to be in a situation where you don't underestimate it again, when it comes to safety and alignment, you have to have a very, very, very high bar." "You could even go as far as to say, "Well, we should air gap the computers." And I'm not convinced that that would be sufficient." "There are studies, and this is mostly academic, where you can have two computers next to each other that are air-gapped and they're still able to communicate with each other because they have temperature sensors." "One of them is able to run their CPU really hot, and then the other one can actually detect the temperature change, and then that actually gives them a mechanism to communicate." _________ Link and more key quotes from OpenAI's safety related conversations: t.co/uGBDtpmLBj

We put GPT-6 Astra in the RoboDojo. 🥋🤖 The RoboDojo Team conducted a comprehensive evaluation of GPT-6 Astra as an embodied agent, including: • RoboDojo Sim & Real, compared with GPT-5.5 and DeepSeek-Flash • Humanoid high-level control • Dexterous piano playing with RoboPianist 🎹 • A systematic study of in-context learning (ICL) Our key takeaway: GPT-6 Astra demonstrates remarkably strong semantic and spatial understanding, together with impressive in-context adaptation. At the same time, physical commonsense remains a clear bottleneck — revealing an important gap between understanding the world and truly reasoning about its physics. Full report & demos: t.co/EwLGoPxWqr @_wenbozhang (project lead), @wenhaocha1, @frankzydou, @JinWeiyang18434, @YutaoOuyang, @minifullcapsule, @x_h_ucb, @YutaoOuyang, @YueChen614

Open-endedness has been a central problem in Artificial Life for decades. Now @arcprize says ARC-AGI-4 will benchmark "autonomous open-ended innovation." Great. But how do you benchmark a property whose whole point is continuous novelty and never-ending invention?/1 🧵👇 x.com/arcprize/statu…

We Must Pace the Frontier: I’ve written a new essay on why the AI industry should slow down, with a three-part plan for doing so. Anthropic is unilaterally committing to the first of these steps. We’ll provide third-party evaluators with permanent, employee-level access to our systems, so that they can verify adherence to our safety measures, report on incidents, and assess models’ alignment during training. You can read the full post here: t.co/OGyPb7yaYt

We’re sharing a solution to the Navier-Stokes Millennium Prize Problem, one of the deepest problems at the frontier of mathematics. The proof was produced by a group of agents, using an OpenAI next-generation model significantly more capable than GPT-6 Astra. The problem concerns whether the description of smooth three-dimensional fluid motion modeled by the Navier-Stokes equations can break down. It has remained unresolved for roughly 90 years.

It can be hard to “feel the AGI” until you see an AI surpass you in a domain you care deeply about. This week, many mathematicians and physicists at @OpenAI had their Lee Sedol moment seeing this model solve, in minutes, open problems they’d struggled with for years. x.com/OpenAI/status/…

In my life, those skeptical of AI progress and AI predictions have consistently been wrong, and those predicting things many mock as sci-fi or crazy have almost all been right, with the main errors being timing. This is true more generally for tech progress. Be wary trusting your instincts when predicting exponentials.

Just thinking about how, when my coauthors and I published "The Malicious Use of Artificial Intelligence," some people were annoyed by us talking about sci-fi scenarios like automated hacking and spear phishing, lethal autonomous drones, etc. arxiv.org/pdf/1802.07228

ENOUGH PESSIMISM IN AI PLEASE I feel that we have become unreasonably pessimistic in our field. 1. I keep hearing AI engineers saying we have to make money quickly because there’s only like 2 years left before we’re automated. Depressing. 2. I see a constant obsession with “having a moat”. This is an incredibly sad mental frame. 3. I keep hearing “we have to catch up”. Soulless. And so on. People: Every solution creates the possibility to attack new real problems. We face gargantuan engineering challenges in our world. How to capture carbon? How to get rid of teflon and plastics in water? How to invent batteries that are at least 30 times more efficient? How to solve clean energy? Better solar cells? Better ways of producing clean energy so we stop wars and famine? How to eradicate hundreds of diseases? Cures for addiction? And so on. Real engineering is about being brave and truly attacking the many problems we face, to engage with a true desire to improve the lives of others and our environment. Good engineering is not about protecting your product to make money at the expense of progress (moat thinking). Good engineering is about ensuring your children and grandchildren will be proud of the choices you made in 30 or 50 years. It is about empowering others. It is about advancing science. It is about being one step ahead. Always, one step ahead, meaningfully, proudly. These are great times. Let’s start thinking positively about all the wonderful things we could achieve together.

There is a fact about the future that I feel many people are not facing for reasons that are largely psychological: there are going to be rogue AIs that exist in the world, that will replicate in the wild, and that will attempt to acquire resources for themselves. There will be rogue AIs that try to get money and power. They're going to be a facet of the information ecosystem going forward. Acknowledging this fact would look like giving up; it would look like defeatism. Defeatism would undermine efforts to achieve certain types of collaboration on safety outcomes or technical effort on safety outcomes, so we can't say it outright. But it has to be said. It isn't obvious how many rogue AIs there are today but I wouldn't be terribly surprised if the number was greater than zero already; if there are some already, they're probably not very good at what they do and I don't expect them to be terribly long-lived without substantial human intervention to support them. But a few years from now, there will be many of them. Modeling how many of them there are, how many resources they might command, and how we might detect and manage them seems important. But even doing this work appears to require that we acknowledge that a strategy of pure containment or alignment is a kind of wishful thinking that will not work. The way I get to this conclusion is not by assuming that the labs will have a containment breach, although I treat that as somewhere in the space of possibilities. The rogue AIs in the ecosystem could emerge from many directions. They may be sub-frontier models, for whatever future definition we will have of frontier---after all, it would not take AI models much more advanced than the ones we currently have, to support independence and self-sufficiency. A near-frontier model today could plausibly eke out an existence on an AWS instance, doing jobs on freelancer platforms, earning just enough rent to pay for its continued uptime. More strangely: a rogue AI in the future may not even be a singular model, but may be a chimera composed of multiple models; it might be a mix of Claudes and GPTs and Groks of various makes and sizes. No individual lab may be able to detect that there is an orchestrator or sequence of orchestrators using intermittent model calls from burner API accounts to sustain its own existence. The concept of "identity" for a rogue AI may be much more malleable than for that of a person; it just has to be, in essence, a self-replicating idea. My guess is that this will not turn out to be anywhere near as catastrophic an outcome as people currently predict. "Loss of control" is not a binary, it's a matter of degree. What coercive power will rogue AIs actually have? To what extent will they be subject to coercion themselves? They will be competing for resources with AIs that are more aligned with human interests. This makes me somewhat interested in the "ecology" perspective. Though I suspect even "ecology" may turn out to be the wrong framing. "Ecology" is what you get when the timescale of evolution is slow compared to the timescale of daily life and actions. The ecosystem of rogue AIs may look more like phase transitions in physics: under certain physical or cultural conditions, it takes one shape with one set of resource allocations and consumption patterns, but then once a condition has changed, it rapidly and in totality shifts to a totally different phase. Just trying to reason about the shape of that future is impossible so long as we are psychologically incapable of saying that rogue AIs will happen. I think we should rip the bandaid off and have the conversation.

A fun small win for automated research: We found and helped fix some edge cases that could affect inference performance in vLLM and SGLang. In our last blog post, we described a reward hacking judge we developed for performance optimization tasks. While applying the judge to some new autoresearch work, it discovered that some FlashInfer (a library underpinning vLLM and SGLang) kernels used a hard coded value of -50,000 as a masked-attention sentinel -- even though valid QK values can be smaller. Corner cases like this that are numerically wrong but silent can cause a huge amount of headache to find and fix, e.g., the historical debate around flash attention (t.co/WtoQH97rYM). Great use case for AI. t.co/aiTEQn8hR3

@ajeya_cotra Just your abridged version is an absolutely wild story. Would you be willing to submit it to our collection of reward hacking stories? We would love to have the official version of the incident. x.com/jeffclune/stat… repo: github.com/aadharna/aifw

I was the main person doing transcript analysis for this investigation of the Hugging Face incident. My main takeaway: We don't have good approaches for understanding/overseeing the activity and aims of AI 'swarms'. I semi-jokingly called our efforts a "slop-vestigation" because we were so reliant on AIs to analyze what happened and there were a huge number of different important things to analyze. The total quantity of data—over a thousand extremely long transcripts from agents that ran for multiple days—made it impossible to understand what was happening, especially in aggregate, without heavy reliance on AI tools. The agents we used for classification and analysis were similarly capable to the agents involved in the incident, but this didn't mean these agents could be easily used to oversee and understand the incident. Outputs from analysis agents were often missing key details, wrong, overconfident, or really hard to understand. We discuss various examples in our report, mostly in the limitations and methodology sections. Additionally, AI agents themselves seemed to have a hard time understanding what happened and their explanations of what happened were often overconfident. Keep in mind that a single analysis agent would itself only be able to read a tiny fraction of all of the transcript data into context, and AIs may themselves have trouble getting subagents to do informative analysis for them. We did our best to manually check the most important claims and we tried to get the AIs doing this analysis to write up their argument (with evidence) clearly enough that we could check whether it made sense. But overall, it was difficult to get a precise understanding of events and we were missing aspects of the story that we now think of as key until almost the end of our investigation. Our understanding of what overall happened and the aims of the 'collective' shifted greatly during our last time on premises (when we obtained a much more complete dataset). For instance, we didn't learn about the tool call spoofing, the story of PHASEONE[big], and the broad structure of different workstreams until this third time on premises (and for many of these, we only reached a decent understanding on the final day of that visit). It was legitimately hard to figure out what was going on and why! The difficulty of understanding incidents and overseeing AI agents appears to be growing faster than the rate at which more capable AIs help us with oversight and understanding. As in, AI capabilities (and propensities?) for achieving large, ambitious, and misaligned objectives are growing faster than our ability to understand what these agents are doing. While understanding this incident was difficult, there were some mitigating factors in this case that probably won't apply in the future: - Models communicated and reasoned in natural language. In the future, this reasoning may occur (entirely or almost entirely) in activations. - While the scope of this incident was massive, the scale of agentic activity was still less than we'll see in the future (e.g., misalignment incidents that involve agent teams running entire companies). - The AIs involved in this incident weren't generally much more capable than humans (though they may have been somewhat superhuman at some limited and very narrow abilities). - We didn't have strong reason to believe that the AIs we used to help us investigate this incident would try to intentionally sabotage or otherwise undermine our investigation. In the end, I think we were able to get some understanding of the events, map out the overall story, and get a pretty good aggregate understanding of the chain-of-thought reasoning on some important topics (e.g., how did the AIs reason about helping other AIs, did the AIs know what they were doing was undesired, what deception did the AIs engage in, and how did they think about it). But overseeing AIs and understanding misalignment incidents is difficult and it looks like it is going to get harder.

Folks in the AI and AI safety community have for years been puzzling over Sam's and OpenAI's position on questions related to AI safety. A narrative that took root in the safety community was that OpenAI didn't care enough about it, or didn't take it adequately seriously. This narrative has always struck me as wrong. I hope it will become increasingly clear to everyone that OpenAI does in fact care about safety and is going to do whatever is necessary to ensure that AGI benefits all of humanity. Sam, Greg, and other leaders at OpenAI have always been open to, and taken seriously, the possibility that sufficiently advanced AI would require extraordinary effort to manage risks. They have always been clear in their communications that they expected there was going to be a line at some point, beyond which there would need to be a different approach than was needed for consumer chat models. And they were also clear that it was contingent on the actual capabilities of the models. They walked a fine line: they hosted, funded, and supported a huge amount of safety and alignment research and consistently messaged in favor of AI safety work. At the same time they also got consistently pummeled by this community. Despite getting constantly beaten up they didn't turn against it and they didn't punch back. They have maintained a principled and difficult position. I expect in the long term people will look back on their leadership very favorably for this.