GPT-2

2 statements across 2 episodes · 0 bullish · 1 bearish · 2 people on the record · first statement Apr 3, 2025 by Francois Chollet · said 28 times in 16 episodes since 2019 · across every show →

Mentions by year

brought up most by Sebastian Raschka (3), Gary Marcus (3), Francois Chollet (3), Sean Gourley (2), Gideon Mann (2), Clement Delangue (2), Amjad Masad (2), Stephen Balaban (1)

tap a year for its mentions
00326420192020202120222023202420252026episodesmentions
02420192020202120222023202420252026episodes it came up in
001.523420192020202120222023202420252026episodesmentions per episode
2026 5 mentions in 3 episodes 2 per episode
2025 6 mentions in 2 episodes 3 per episode
2024 4 mentions in 4 episodes 1 per episode
2023 6 mentions in 3 episodes 2 per episode
2020 1 mention in 1 episode
2019 6 mentions in 3 episodes 2 per episode

every mention, scene by scene, with the transcript →

Everything said about GPT-2, oldest first

Apr 3, 2025 bearish
Assertion Supported
Chollet: 50,000x LLM scaling yielded flat progress on ARC benchmark
“Because between, like, GPT-II and GPT-IV. There's been this 50,000 X scale up of base models that has resulted in, in, in basically a flat curve. On Arc.”
Francois Chollet Apr 3, 2025 ▶ 11:44 Chasing Real AGI: Inside ARC Prize 2025 with Chollet & Knoop
Jan 29, 2026 neutral
Assertion Not checkable as stated
DeepSeek's architecture is still built on a GPT-2 scaffold
“And you can actually, in fact, Take a GPT one or two model and with a few, I mean, few lines of code almost, you can transform it into the latest let's say deep seek version, 3.2 architecture. It's not like a big leap. It's still the same as scaffold.”
Sebastian Raschka Jan 29, 2026 ▶ 2:54 State of LLMs 2026: RLVR, GRPO, Inference Scaling — Sebastian Raschka
Made with StarZero

Turn any episode into a week of clips.

This entire site, over 400 conversations 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.