Unverified50% confidenceOpinionExact time
Sam Altman stated that OpenAI and Anthropic have significant disagreements on safety views, and said the world should accept some bad things happening in exchange for the benefits of technology.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
10/6/2026
First Seen
Valid until: 1/4/2027
Sources
Related Entities
Related Claims
UnverifiedDeepMind, Anthropic, and OpenAI all list consequence-aware reasoning as a priority in their respective safety research agendas, but the field is still in its early stages.70% similarUnverifiedOpenAI创始人Sam Altman曾提出，未来最稀缺的能力不是执行，而是'knowing what to ask for'（知道该要什么）68% similarUnverifiedAnthropic's core narrative is "staying ahead to control the risk"—that is, if powerful AI is inevitable, it is better for a safety-minded team to lead the way—which distinguishes it from OpenAI and Google.66% similarUnverifiedThe author stated that they do not trust Anthropic or OpenAI to fairly open up frontier intelligence to everyone.66% similarUnverifiedWithin OpenAI there is a divide between an 'accelerationist' faction and a 'safety-first' faction: the accelerationists advocate discovering problems through rapid iteration and market feedback, while the safety faction argues that releases should be restricted until risks are sufficiently understood.65% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/982132API
curl https://kongchang.com/api/v1/knowledge/claims/982132MCP
get_claim(id=982132)