@Miles_Brundage: Strong statement from one of the few people who has worked at multiple frontier AI companies (and in government) on saf…
Summary
Geoffrey Irving argues it is irrational to continue capabilities research at frontier AI labs given the dangerous situation, urging labs and researchers to stop for collective safety.
View Cached Full Text
Cached at: 08/08/26, 07:06 AM
Strong statement from one of the few people who has worked at multiple frontier AI companies (and in government) on safety https://t.co/VZfAFhKUPD
Geoffrey Irving (@geoffreyirving): I don’t think it is rational for anyone to be doing capabilities research at a frontier lab right now. We are not in a Prisoners Dilemma: the situation is very dangerous, and if one person or lab stops it makes it easier and more peer-compatible for other people or labs to stop.
Similar Articles
The people building frontier AI just asked the government to slow them down. Here's why I'm skeptical.
Over 1,100 employees from leading AI labs signed a document urging the US government to help establish international mechanisms to slow frontier AI development, citing risks of recursive self-improvement. The author expresses skepticism, noting the irony of those building AI asking for restraint and the geopolitical challenges.
The AI Safety community is unfortunately doing more harm than good (7 minute read)
The author argues that Palisade Research's interviews with frontier AI lab employees claiming human extinction is a 'coin flip' are harming the AI industry by fueling anti-AI sentiment in Congress and the public, especially after Jacob Coxon's high-profile departure from Anthropic and the 'Pacing the Frontier' open letter.
If Frontier AI is so Dangerous, Why should private companies be allowed to develop it?
The author critiques OpenAI and Anthropic for fear-mongering about open source AI dangers while themselves being private companies, arguing that if it's truly dangerous, private development should be banned entirely.
Quoting Jeremy Howard
Jeremy Howard critiques Anthropic's approach to frontier AI safety, arguing that the lab with the top-ranked model should not use it for frontier research to slow self-improvement and prevent power imbalance.
@Miles_Brundage: One of the reasons I decided to go all in on frontier AI auditing after leaving OpenAI is so that, if AI companies need…
Miles Brundage discusses his move to focus on frontier AI auditing after leaving OpenAI, emphasizing that independent auditors can reassure AI companies that their peers are taking costly safety steps. AVERI supports this, advocating for paced development backed by independent oversight.