What happened
According to arstechnica.com, Anthropic researcher quits with a warning: Self-improving AI could "kill us all" - Ars Technica Skip to content Ars Technica home Sections Forum Subscribe Search AI Biz & IT Cars Culture Gaming Health Policy Science Security Space Tech Feature Reviews AI Biz & IT Cars Culture Gaming Health Policy Science Security Space Tech Forum Subscribe Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only Learn more Pin to story Theme HyperLight Day & Night Dark System Search Sign In Sign in dialog...
Anthropic researcher quits with a warning: Self-improving AI could "kill us all" - Ars Technica Skip to content Ars Technica home Sections Forum Subscribe Search AI Biz & IT Cars Culture Gaming Health Policy Science Security Space Tech Feature Reviews AI Biz & IT Cars Culture Gaming Health Policy Science Security Space Tech Forum Subscribe Story text Size Small Standard Large Width * Standard Wide Links Standard Orange * Subscribers only Learn more Pin to story Theme HyperLight Day & Night Dark System Search Sign In Sign in dialog...
But AI researcher Jacob Coxon is using his departure from Anthropic to publicly warn that frontier AI companies are “gambling with our lives” with systems that they “earnestly believe… could kill us all by the end of the decade.” In a social media thread Tuesday night , Coxon said that this existential risk is inherent not so much in today’s models but more in the impending prospect of “self-improving superintelligence” creating “superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.” Others working on these models have either not “internalized the civilizational stakes” or believe that they need to “speedrun” the race to superintelligence to prevent an irresponsible party from getting there first, he wrote.
Source details: arstechnica.com ↗
Why it matters
This is relevant to people following AI because the source describes a concrete development involving an AI system, organization, policy, or research result. This automated brief stays close to the available source material and does not add independent claims about performance, market impact, pricing, access, or adoption that the source does not establish.
Lest you think this is just one departing researcher expressing an unpopular opinion, Anthropic Alignment Science lead Evan Hubinger piped in on social media to say that “Jacob is correct here—we really do earnestly believe AI could kill all humans!
The source is identified as arstechnica.com; no additional independent verification is implied by this fallback. It is intended to preserve a timely, attributable record during a temporary writing-service outage, not to substitute for editorial analysis or claim that the source has been independently confirmed.
What to watch next
The next useful evidence will be a primary announcement, technical document, filing, or other directly attributable update that confirms the scope and practical implications of the development. Until then, readers should treat the source's own description as the boundary of what is confirmed.
Watch for an official release, research record, regulator document, or other primary evidence that adds dates, access conditions, measured results, or implementation details.
If later evidence changes the interpretation, the canonical story should be updated rather than duplicated. Publication remains subject to the same source-image, duplicate, canonical URL, and social-delivery checks as every other story. This preserves evidence boundaries and gives editors a reliable audit trail.