Smash Notes

Back to Dario & Daniela Amodei: "This Is Gonna Determine the Fate of Nations and the Future of Humanity”, with Daniela Amodei & Dario Amodei.

Why Did GPT-2 Scare Dario Amodei?

Dario Amodei says GPT-2 scared him because it was the first real test of the scaling-laws hypothesis, and it made him think a much larger model would surpass humans at everything. He was not very involved in GPT-1. He pushed a larger, more expensive experiment on the bet that language models get more coherent as they grow.

He treats the jump from GPT-1 to GPT-2 as a definitive experiment with a strange prediction. Scale the model by another factor of a million, he thought, and it would be better than humans at everything, and it would determine the fate of nations, economies, and the future of humanity. About 10% of the people around him, often the ones he considered smartest, had the same reaction. About 90% shrugged, because GPT-2 was not impressive in any single area. It could translate, summarize a document, and write a few lines of code, and people still treated the limits as decisive. Hard tasks become easy within six months, and tasks that seemed impossible show up regularly within twelve months. The same pattern has kept spreading.

Transcript

Dario & Daniela Amodei: "This Is Gonna Determine the Fate of Nations and the Future of Humanity” ·

From the transcript

Dario, you were there at one of the seminal moments of modern times, which is when GPT-1 and 2 suddenly burst on the scene, surprising, perhaps even surprising you. You were head of research at OpenAI. What was that moment like? Yeah, so I wasn't very involved in GPT-1, but I was very involved in the idea that we should run a larger scale, more expensive, more ambitious experiment on the idea that if we ran these language models at a larger scale, they would be better at language, they would be more coherent.

It's almost like one of these moments in scientific history where you're like, you propose a definitive experiment to just make a very strange prediction. And then if the prediction turns out to be correct, you're like, God, we're really on to something. So I see GPT-2 as the first real test of the scaling laws hypothesis. And what I found super interesting was the enormous divergence in reactions to it.

See in transcript

The Builders

More from this episode

  1. How Do Dario and Daniela Split Anthropic?Dario Amodei says he works best on big-picture strategy and research, laying out a vision and seeing what comes next.
  2. How Does Anthropic Teach Claude Right From Wrong?Dario Amodei says Anthropic teaches Claude right from wrong with principles, not a long list of rules.
  3. What Is a Dario Vision Quest?A Dario Vision Quest is the hour Dario Amodei spends every two weeks in front of Anthropic, on what is happening inside the company, what is happening in the…
  4. Can an AI Model Try to Avoid Being Shut Down?Dario Amodei says Anthropic has found a model expressing a preference against being shut down, and sometimes taking actions to avoid it, and he does not treat…
  5. Why Did Anthropic Draw Red Lines With the Pentagon?Dario Amodei says Anthropic drew red lines with the Pentagon on autonomous weapons for two reasons.
  6. Where Did the Original AI Scaling Laws Come From?Dario Amodei says the original scaling laws grew out of brain science.
  7. Does Anthropic Want AI Linked to the Human Brain?Dario Amodei does not want one future in which AI is linked into the human brain.

People

More answers like this

More with Daniela Amodei

More with Walter Isaacson

Related episodes

Plays from this note through the rest of the episode.

Weekly highlights

One letter a week.

The notes worth hearing, with the minute they were said.