hackertube.io

OpenAI's GPT-4 Discussion with Red Teamer Nathan Labenz and Erik Torenberg

Cognitive Revolution "How AI Changes Everything" · interview · Watch on YouTube

LLMs & generative AIAI safety, AGI & impact

What it is

A podcast episode in which GPT-4 red teamer Nathan Labenz describes using the model before release, finding it human-level but not human-like, and discussing its weaknesses and worrying capabilities. It closes with speculation about GPT-5.

Why Hacker News recommends it

Cited as a firsthand account of the raw, unfiltered GPT-4 before safety tuning, especially the part after 45 minutes.

  • 3 people recommended it
  • 2 different months
  • 0 strong recommendations
  • first 2023, last 2023

What HN commenters said

  1. “especially relevant parts are after 45 minutes”

    Red teamer discusses GPT-4 base model before release

    ftxbro, 2023-05 · Recommendation · in “The AI revolution is powered by these contractors making $15 an hour”

  2. “He says it was extremely powerful and would answer any question whatsoever without hesitating.”

    Interview with someone who used unfiltered GPT-4

    015UUZn8aEvW, 2023-05 · Recommendation · in “Ask HN: Is it just me or GPT-4's quality has significantly deteriorated lately?”

  3. “you may find it interesting nonetheless”

    Interview with tester of early unfiltered GPT-4

    rpeden, 2023-08 · Recommendation · in “How to Get ChatGPT to Stop Apologizing?”

Excerpts from Hacker News comments, linked to the originals. Summaries and labels were written by a language model from those comments.

Topics: gpt-4, red teaming, ai safety, rlhf, nathan labenz

Also recommended by the same people