PPO

4 statements across 3 episodes · 3 bullish · 1 bearish · 3 people on the record · first statement Jan 11, 2024 by Nathan Lambert · said 1 times in 1 episodes since 2025 · across every show →

Mentions by year

brought up most by Benjamin Eysenbach (1)

tap a year for its mentions
0011112025episodesmentions
0112025episodes it came up in
000.50.5112025episodesmentions per episode

every mention, scene by scene, with the transcript →

Everything said about PPO, oldest first

Jan 11, 2024 negative
Opinion
Lambert: Scaling PPO in RLHF is a Nightmare
“PPO is kind of a nightmare to scale.”
Nathan Lambert Jan 11, 2024 ▶ 1:01:41 The Origin and Future of RLHF: the secret ingredient for ChatGPT - with Nathan Lambert
Jan 11, 2024 positive
Assertion Supported
Lambert: Meta Used Rejection Sampling to Bootstrapping Llama 2 RLHF
“Llama started their RLHF process with this to get some signal out of preference data. That preference data went into a reward model, and then the reward model did a good enough ranking that it was, like, essentially superpowered instruction tuning based on rew…”
Nathan Lambert Jan 11, 2024 ▶ 1:03:01 The Origin and Future of RLHF: the secret ingredient for ChatGPT - with Nathan Lambert
Oct 16, 2025 positive
Insight
Corbitt: PPO enables training purely on real production traces without simulation
“And PPO, now in practice, a lot of times when you're training with PPO, you also will use an environment like that because it lets you do a bunch of runs and be more data efficient. But at least in principle, you have the option with PPO, you can actually, lik…”
Kyle Corbitt Oct 16, 2025 ▶ 23:39 Why RL Won — Kyle Corbitt, OpenPipe (acq. CoreWeave)
Dec 31, 2025 positive
Insight
Eysenbach: 1,000-layer RL requires reward-free objectives, not just architectural tricks
“I think the main conclusion is that using big networks not only requires these architectural tricks, but also, as Kevin mentioned before, it requires using a different objective. This objective doesn't actually use rewards in it, and so there's another word in…”
Benjamin Eysenbach Dec 31, 2025 ▶ 8:08 [NeurIPS Best Paper] 1000 Layer Networks for Self-Supervised RL — Kevin Wang et al, Princeton
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 200 episodes transcribed, diarized, checked and made playable, runs on the StarZero media pipeline. Drop in your own episode and the podcast clipper finds the moments worth sharing, cuts them, captions them, and reframes them for every feed.