Public transcript

The Real Reason Anthropic's CEO Walked Away From Sam Altman

By: Bloomberg Originals
Length: 3 min
Language: English
Text from: the video's captions
Transcribed: 13 October 2026

Read the full transcript Open on YouTube

Notes in English.

Summary

In this short Bloomberg interview, Anthropic CEO Dario Amodei explains the viral moment at India's AI summit where he and Sam Altman appeared not to hold hands, attributing it to summit disorganization rather than a deliberate snub. He discusses how trustworthy AI actors should band together to force higher standards on the rest of the industry. Most notably, he says his decision to leave OpenAI came not from safety disagreements but from a loss of trust in the company's leadership, their honesty and their stated values, and that he's at peace with each company going its own way.

Key points

Dario Amodei says the India AI summit hand-holding moment was due to disorganization, not a deliberate snub of Sam Altman.
He argues the trustworthy actors in AI need to band together so untrustworthy players are forced to adopt the same standards.
He says safety disagreements alone were not enough to make him leave OpenAI, since disagreements are normal.
He says he left because he felt he could not trust the company's leadership — that their values, honesty and motives were not what they claimed.
He says he is at peace with Anthropic and OpenAI doing things their own way, and the market and public opinion will decide who wins.

Questions it answers

00:00The hand-holding moment

What happened at the India AI summit when Dario Amodei and Sam Altman didn't hold hands?

He says the summit was extremely disorganized, the standing order was changed last minute, and everyone was suddenly told to hold hands for a photo — it was not a deliberate snub.

The hand-holding was ordered on the spot for a photo.
He says international summits with heads of state are generally super disorganized.

00:38Trust and industry standards

How can AI leaders cooperate on existential risk if they can't even hold hands on stage?

He says some players are more trustworthy than others, and the trustworthy actors need to band together so untrustworthy players are forced to adopt the same standards.

Some actors won't do the right thing on their own.
If a majority of the industry does the right thing, the rest have little choice but to come along.

01:24Why he left OpenAI

What really made Dario Amodei leave OpenAI?

He says safety disagreements alone weren't enough to leave; the real issue was feeling he couldn't trust the company — that its values, honesty and motives weren't what they claimed, plus disturbing patterns of behavior.

Disagreements on safety are normal and not sufficient to leave.
Distrust, dishonesty and misaligned values were the deciding factors.

02:44Moving on

How does he feel about Anthropic and OpenAI going their separate ways?

He says the way to resolve it is for each side to go off and do their own thing, and he's completely at peace with that — the market and public opinion will decide who wins.

He sees no point arguing without shared vision or trust.
He's fully at peace with the split.

Notes

The India AI summit hand-holding moment

Dario Amodei explains that the summit was extremely disorganized; the order of people on stage was changed at the last minute.
Everyone was suddenly ordered to hold hands for a photo, including Narendra Modi telling people to do so.
He insists it was not a deliberate refusal to hold Sam Altman's hand, and that these international summits with heads of state are generally chaotic.

Trust and cooperation on existential risk

Asked how AI leaders can cooperate on existential risk if they can't hold hands on stage, he says some players are more trustworthy than others.
His view: trustworthy actors need to get together and put untrustworthy actors in a position where they have to adopt the same standards.
He notes that from experience, some people won't do the right thing on their own, but if a majority of the industry does, the rest have little choice but to come along.

Why he really left OpenAI

He says there were valid disagreements on safety, but disagreements alone are not sufficient reason to leave — people disagree with him and with each other all the time.
The real issue: he felt he could not trust the company's leadership — that their values were not what they said, that they were not honest, and were not in it for the reasons they claimed.
He describes disturbing patterns of behavior and dishonesty that made it very hard to keep working with or trusting the company.

Where things stand now

His resolution: go off and do your own thing rather than argue with someone you don't share a vision with or trust.
He is completely at peace with Anthropic doing things their way and OpenAI doing things theirs.
Ultimately, he says, we'll see who wins in the market and in the court of public opinion.

Text from the video's own captions. Summary and notes written by AI from the transcript, so check anything important against the recording.