Sam Altman called the surrender of human judgment to AI "a real safety issue" as OpenAI safety leader David Robinson quit and rebuked its safety record. Teams deciding how much to automate can now cite a model vendor's own CEO in favour of human review.
Perspective Coverage
3 publishers
- Builder
- Builder 34%
- Operator
- Operator 38%
- Investor
- Investor 28%
Reality
- Evidence60
- Adoption
- Not scored
- Hype gap+25
- Incentives65
- Confidence58
Mustafa Suleyman calls AIs 'internally hollow' after shipping Copilot with a mascot
Microsoft AI CEO Mustafa Suleyman calls AIs 'internally hollow' in a 5,700-word essay, after giving Copilot an animated mascot named Mico. For teams building assistants, the stronger case for designing them as tools comes from what agents such as Meta's Muse do with a user's private messages.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+5
- Incentives
- Not scored
- Confidence40
buildConfirmed5 publishers Mustafa Suleyman's essay traces Claude's talk about its own possible feelings back to Anthropic's constitution. His proposed remedy is an edit to a document. Any team shipping an assistant could make it.
Perspective Coverage
5 publishers
- Builder
- Builder 36%
- Operator
- Operator 37%
- Investor
- Investor 27%
Reality
- Evidence45
- Adoption5
- Hype gap+35
- Incentives70
- Confidence62
Mustafa Suleyman's essay and BBC interview criticise how Anthropic trains Claude to act human. The four capabilities he warns about are granted separately, by different people, in every agent rollout.
Perspective Coverage
7 publishers
- Builder
- Builder 37%
- Operator
- Operator 44%
- Investor
- Investor 19%
Reality
- Evidence70
- Adoption
- Not scored
- Hype gap+30
- Incentives60
- Confidence68
buildOne report1 publisher DeepMind's interpretability lead found Anthropic's J-space again in an open 27B model
Anthropic screened Claude for one property, verbalizability, and found two more in the same representations. Neel Nanda reproduced the structure in open weights, and the outside commentators Anthropic invited disagree about what it is.
Reality
- Evidence40
- Adoption22
- Hype gap+12
- Incentives62
- Confidence42
Microsoft's conduct code puts its MAI models on different rules from the Claude it resells
Microsoft's new code of conduct binds only the MAI models it builds in house, and it rejects the model welfare research at Anthropic, where Microsoft has $5bn invested and whose Claude models it sells inside Copilot.
Reality
- Evidence64
- Adoption15
- Hype gap+30
- Incentives80
- Confidence58