OpenAI o3, every mention
32 scenes · ← back to OpenAI o3
tap a year for its mentions
every year anyone Shawn Wang 6Noam Brown 5Nathan Lambert 5Will Brown 4Alex Lupsasca 3Sujay Jayakar 2Quentin Anthony 2Olivia Watkins 2Josh Ma 2Ashvin Nair 2
Verbatim, from the transcripts: the passages where OpenAI o3 comes up
🔬How GPT‑5 derived new results in theoretical physics and quantum gravity — Alex Lupsasca, OpenAI
- ▶ 2:15 Alex Lupsasca Um, in particular, ChadGPT-O-III was the first really strong reasoning model that could do actual math that was useful for my research and could save me a lot of time. 2 times in the scene
- ▶ 1:10:43 Alex Lupsasca Oh, three, which was the really first, really strong reasoning model came out and was able to do a calculation for me that would have taken me days and did it, 11 minutes.
The End of SWE-Bench Verified — Mia Glaese & Olivia Watkins, OpenAI Frontier Evals
- ▶ 6:37 Olivia Watkins And so this happened by first, um, taking all the problems that O-three couldn't solve reliably, and then again, uh, getting a lot of humans to do basically another pass of, uh, kind of digging into, you know, what's wrong. 2 times in the scene
[State of RL/Reasoning] IMO/IOI Gold, OpenAI o3/GPT-5, and Cursor Composer — Ashvin Nair, Cursor
- ▶ 0:18 unnamed speaker Before that, you were opening, I worked in O.M.O.N.O.T.E. 2 times in the scene
- ▶ 17:33 unnamed speaker And there, we merged the four O and O-one O-three line into five, and now we're splitting it out into five and five codecs again.
- ▶ 21:32 Ashvin Nair Uh, it's, uh, it's kind of like, you know, now that, um, like, you know, when O.T.E. was, like, kind of, structured as a product, like, I think it just, like, kind of gets, like, larger and larger how many people worked on it, so. 2 times in the scene
How Zyphra went all-in on AMD + Why Devs feel faster with AI but are slower — with Quentin Anthony
- ▶ 34:29 Quentin Anthony Um, if I need something a bit more heady, like, um, writing me, for example, the, the setup for kernels, like we were talking about earlier, I might use a thinking model, like an extended thinking on GPT five or an O three kind of prompt. 2 times in the scene
Amp: The Emperor Has No Clothes
- ▶ 25:32 Alessio Fanelli So you have Sonic for, for the agent, you have O three for the Oracle.
Greg Brockman on OpenAI's Road to AGI
- ▶ 19:45 unnamed speaker Four is multi-modality and all these different low latency, long thinking with a three, what's going to be the five flagship thing? 3 times in the scene
The RLVR Revolution — with Nathan Lambert (AI2, Interconnects.ai)
- ▶ 6:51 Nathan Lambert Thinking like what is the right diagram to encapsulate how O-three is trained, which in action, they take multiple actions because the next sequence depends on the feedback from the environment, which is some sort of information store. 2 times in the scene
- ▶ 22:11 Shawn Wang You, you, you seem to assert that O three, 4 times in the scene
- ▶ 31:13 Nathan Lambert So if you were to train an open model, that is going to be a good reason or like oh three, but on.
- ▶ 39:44 Nathan Lambert Like if O three just infinite loops itself for a bunch of people, like that's not good.
⚡️Using RFT to Build Clinical Superintelligence
- ▶ 23:11 Brendan Fortuna Similar to like, you know, if you ask, you know, O-three to go on the internet and find the right travel kind of destinations, it's just going to kind of get like tricked by whatever has the best SEO.
AI is Eating Search
- ▶ 17:21 unnamed speaker If you're looking at, like, AI mode or some of the more recent, like, you know, O-three search integrations that ChatDBT has,
⚡️Anthropic vs Cognition on Multi-Agents: A Breakdown with Dylan Davis
- ▶ 22:43 Dylan Davis So O three was released, they had this reasoning thing, and it's basically a visualization of the improved capability of a model if it can reason and think over time.
Scaling Test Time Compute to Multi-Agent Civilizations — Noam Brown, OpenAI
- ▶ 6:13 Noam Brown And I think that that's true even today, that we, we saw that with like going from O-one preview to O-one to O-three, consistent progress. 2 times in the scene
- ▶ 11:04 Spooks (Swyx) So let's say we have, now we have the four O like natively omni model type of thing, then that also makes O three really good at GeoGuessr.
- ▶ 14:12 Noam Brown And you could ask, you know, you could just ask O-three. 3 times in the scene
ChatGPT Codex: The Missing Manual
⚡️Open Questions in Agentic RL — Will Brown (Prime Intellect)
- ▶ 1:38 Will Brown Um, that's kind of the O-Tree magic, is lots and lots and lots of multi-turn tool use with multimodal input in very general settings beyond just mapping code. 4 times in the scene
What is an RL environment? w/ Nous Research's Roger Jin
Fullstack-Bench: The Eval for Coding Agents — with Sujay Jayakar, Chief Scientist, Convex
- ▶ 17:54 Sujay Jayakar Same holds with O three. 2 times in the scene
smol agents are all you need
- ▶ 11:09 Swyx (Marcos Swix) so one of the things that people are talking about in surprisingly O-Tree adopted, or actually not even O-Tree, Deep Research adopted was Gaia.
Outlasting Noam Shazeer, Crowdsourcing Chai AI w/ 1.4m DAU — with William Beauchamp, Chai Research
- ▶ 1:10:49 William Beauchamp And that's what opening I've done with O-one and O-three.
The Unreasonable Effectiveness of Reasoning Distillation: using DeepSeek R1 to beat OpenAI o1
- ▶ 11:05 Shawn Wang Like the only way you, you reach the frontier with the full size models of O-one and O-three is with that stuff.
OpenAI o1 isn’t a chat model (and that’s the point)
- ▶ 31:45 Alessio Fanelli And I'm sure we're going to have an update when O-Tree is made available and see how, how the model evolves.
The State of Reasoning — from Nathan Lambert, Interconnects/AI2 [LS Live @ NeurIPS 2024]
- ▶ 0:27 Nathan Lambert This was before Oh three was announced by open AI.
2024 Year in Review: The Big Scaling Debate, the Four Wars of AI, Top Themes and the Rise of Agents
- ▶ 1:23:23 Shawn Wang It's actually, we got O-one and O-three.
Building AGI in Real Time (OpenAI Dev Day 2024)
- ▶ 1:11:42 Alistair Pullen Um, and also because, fundamentally, like, we are, I think, fairly clearly in a position now where we don't have to worry about what happens when O-two comes out, what happens when O-three comes out.
- ▶ 1:22:38 unnamed speaker Agents and AI employees beyond level three, and his projections of the intelligence of O-one, O-two, and O-three models in future.
- ▶ 1:42:22 Sam Altman The cost, you know, I think would have been manageable with O-one, but by the time of O-three or whatever, like, maybe it would be pretty unacceptable.