If you are inclined to take @ohabryka on his word on issues like this, you may want to know some context... Oliver accuses me and ControlAI …

If you are inclined to take @ohabryka on his word on issues like this, you may want to know some context... Oliver accuses me and ControlAI of telling people to be more coy about extinction risks. I have personally testified in front of the Canadian parliament about extinction risks. ( https://www.ourcommons.ca/DocumentViewer/en/45-1/ETHI/meeting-20/evidence#Int-13276095 ) My pinned tweet is about extinction risks. The article I have most linked to my blog is named "Preventing Extinction from Superintelligence". ( https://cognition.cafe/p/preventing-extinction-from-superintelligence and https://www.google.com/search?q=https%3A%2F%2Fcognition.cafe%2Fp%2Fpreventing-extinction-from-superintelligence ) I have recently published an article that is specifically about people being too coy about extinction risks. ( https://cognition.cafe/p/the-spectre-haunting-the-ai-safety ) I have conceived the Direct Institutional Plan from ControlAI ( https://controlai.com/dip ), that is specifically about talking to policy makers about extinction risks, and helping people do so. ControlAI has started a campaign focused on extinction risks and superintelligence, supported by more than a hundred politicians. ( https://controlai.com/statement ) The statement literally contains the words “risks of extinction” and “superintelligent AI systems”, words that do not go unnoticed over the course of hundreds of briefings. I have cofounded an online community, whose core focus is to train people to talk about extinction risks to Policy Makers, following ControlAI's Direct Institutional Plan. ( https://dip.torchbearer.community/ ) I have consistently pushed people to be more candid about the risks in many organisations! You can ask around at FLI, CAIS, Palisade or MIRI. Me pushing other organizations to be more frank has been an extremely source of friction with other organisations, as many will vividly remember from interactions with me. It is hard for me to be charitable in this situation. I am not perfect, but I can't remember a more baseless public attack against me in my life! In 2023, I asked him and Ben to moderate two unreasonably mean parts (not whole comments) on a single post. This has now somehow been twisted into "trying to get people to be much more coy" and "they've tried quite hard". He has since then tried to discredit people associated with me, in a similar baseless way. See this response from Connor to Habryka's accusation that he has "lost the plot" ( https://www.lesswrong.com/posts/H26ndkABmGvoc9PTC/conjecture-a-roadmap-for-cognitive-software-and-a-humanist?commentId=C8QiYQi4wKaN5Zij9 ). Just consider the level of disconnect from reality here: his evidence against ControlAI is a screenshot of a comment he made about me (and I am not the director of ControlAI nor a spokesperson of it), on an unrelated topic, from 3 years ago, that he is misconstruing. Alas, they've historically also been at the forefront of trying to get people to be much more coy (they've in the past tried quite hard to get me to delete lab-critical comments from LessWrong). https://www.lesswrong.com/posts/vFqa8DZCuhyrbSnyx/integrity-in-ai-governance-and-advo…

Related

• I agree with Connor that most folks concerned about AI have been far too coy about extinction risk, and that ControlAI is one of few excepti…
• We think ControlAI can turn $50M / year into a 10% chance of banning ASI. Most of the AI safety community has been far too coy about extinct…
• ' If a superintelligence is built, humanity will lose control over its future.' Before Canadian Senators, ControlAI's US Director Connor Lea…
• @ESYudkowsky My horrendous nightmare of a political lifecycle, ladies and gentlemen and others.

Source: Connor Leahy (X) | 2026-04-21