Public evidence record

John Schulman

Published podcast speaker

Claims: 16
Episodes: 1
Shows: 1
Named items: 0

Claim ledger

What John said.

3 transcript-backed records

01 / prediction

15 May 2024

I expect that models will be able to use websites that are designed for humans just by using vision, after the vision capabilities get a bit better.

“I expect that models will be able to use websites that are designed for humans just by using vision, after the vision capabilities get a bit better.”

Speaker: John Schulman
Episode: John Schulman (OpenAI Cofounder) — Reasoning, RLHF, & plan for 2027 AGI
Publisher: Dwarkesh Podcast

02 / prediction

15 May 2024

I expect the mental model of a helpful assistant or helpful colleague to become more real.

“I expect the mental model of a helpful assistant or helpful colleague to become more real.”

Speaker: John Schulman
Episode: John Schulman (OpenAI Cofounder) — Reasoning, RLHF, & plan for 2027 AGI
Publisher: Dwarkesh Podcast

03 / prediction

15 May 2024

I realized a lot of the things that people thought were flaws in language models, like blatant hallucination, could be not completely fixed but things that you could make a lot of progress on with pretty straightforward methods.

“I realized a lot of the things that people thought were flaws in language models, like blatant hallucination, could be not completely fixed but things that you could make a lot of progress on with pretty straightforward methods.”

Speaker: John Schulman
Episode: John Schulman (OpenAI Cofounder) — Reasoning, RLHF, & plan for 2027 AGI
Publisher: Dwarkesh Podcast