Why Anthropic Broke With Sam Altman: Dario Amodei Says Trust Collapsed
TL;DR · What you'll learn
- 1 Asked about the moment in India when he appeared to refuse a handshake with Sam Altman on stage, Amodei said the summit was highly chaotic and everyone was suddenly told to link hands after last-minute staging changes.
- 2 When asked whether leaders building the world's most important technology can really cooperate on existential risk if they cannot even shake hands, he said trusted actors should align first.
- 3 He described an industry order in which trusted players act together and force less trusted actors to adopt the same standards.
- 4 On leaving OpenAI, he said ordinary disagreement over safety is not enough to make someone quit a company, but that the decisive factor was lack of trust.
- 5 He said that once you feel you cannot trust someone, their stated values diverge from reality, they lack integrity, or they show disturbing patterns of behavior, continuing to work together becomes extremely difficult.
- 6 He concluded that there is little point fighting with someone who shares neither your vision nor your trust, and that both sides should go their separate ways and let the market and public opinion decide.
Read as slides
3 slides total
What Really Happened in the Handshake Incident
In Bloomberg's interview, Amodei is asked about the moment at an AI summit in India when he appeared to refuse a handshake with Altman. His explanation is almost comically literal: the summit was run in a state of extreme confusion, speaker positions and sequence changed at the last minute, and after the photos they were suddenly told to join hands.
He invokes the image of Prime Minister Narendra Modi unexpectedly urging everyone to hold hands, emphasizing that international summits with heads of state often descend into this kind of staging chaos. While careful not to single out India for criticism, he gently pushes back on media readings that tried to assign deeper meaning to the moment.
A Worldview Built on Trusted Actors Coordinating First
Amodei says that some players in the industry are more trustworthy than others, and that Anthropic is not the only one he would put in that category. His ideal outcome is a system in which trusted actors first align with one another and then push less trusted players into a position where they have little choice but to adopt the same standards.
Even if some participants would not choose the right behavior on their own, a trustworthy majority can structure the field so the rest have no real alternative. It is a pragmatic view of existential-risk coordination, one that relies less on goodwill than on incentives and industry structure.
The Core of His OpenAI Exit Was "Trust"
On the question that has become Silicon Valley folklore, why he left OpenAI, Amodei says that disagreement over safety alone is not enough to justify leaving a company. Internal differences of opinion are normal.
What proved decisive, he says, was a collapse of trust. If you feel you cannot trust the other party, that their professed values do not match reality, that they are not acting with integrity, or that they show unsettling behavioral patterns, then continuing to work at that company becomes very difficult. If two sides share neither vision nor trust, his conclusion is that they should stop fighting, go separate ways, and let the market and public opinion render the verdict.
Editor's Take
The value here is Amodei stating plainly that the core of leaving OpenAI was 'a lack of trust.' Disagreement over safety alone wouldn't make him quit, but sensing a gap between stated values and reality, and an absence of integrity, makes collaboration impossible — an organizational and human problem, not a technical one. His order-of-trust view — that if a trustworthy majority acts rightly, others are pressured to adopt the same standard — offers one hypothesis for how self-governance in AI might form.
Source
Bloomberg Originals
The Real Reason Anthropic's CEO Walked Away From Sam Altman
This article auto-summarizes the YouTube video's transcript with Claude. Please refer to the original video for nuance and exact wording.
Watch on YouTube →Related
3 articles
Bloomberg Originals
Inside the Mind of Anthropic CEO Dario Amodei: The Weight of Exponential Growth and a Bet on 'Doing the Right Thing'
Asked about sleep, Amodei said he is 'learning the skill of resting and sleeping under abnormal pressure,' candidly acknowledging the mental load of leading the company.
Bloomberg Originals
Inside Anthropic, the $965 Billion AI Juggernaut | The Circuit
Dario Amodei is presented as an unusual AI celebrity who has long warned the world about AI risk, while his company Anthropic has grown into a frontier AI firm valued at nearly $1 trillion.
THE SECRET SAUCE
The 'Pope of AI' Dario Amodei's Four Signals: What Anthropic's Clash With State Power Really Means
The AI world has its own 'Pope' -- not Sam Altman, not Jensen Huang, but Anthropic founder and CEO Dario Amodei.