GPT, every mention
82 scenes, the whole family · ← back to GPT
tap a year for its mentions
every year anyone Shawn Wang 17Jeremy Howard 4Vasek Mlejnsky 3Jack Morris 3Ethan He 3Carina Hong 3Alex Lupsasca 3Youssef Rizk 2Stephanie Palazzolo 2Stanislas Polu 2
Verbatim, from the transcripts: the passages where GPT comes up
⏭️ Forward Deployed: Voice AI on what works in 2026
- ▶ 19:18 unnamed speaker So on these cascaded pipelines, what, what models are you guys using, let's say, of the frontier models, of the, you know, GPTs, of the
Next 100x in AI: Inference, Networking, & Self-Optimizing Models — Philip Kiely & Ali Taha, Baseten
OpenAI’s Vision for the AI Super App — Akshay Nathan, OpenAI
- ▶ 40:16 Shawn Wang Presumably at OpenAI, people are obviously more open to being basically rated by GPT.
⚡️Making DeepSeek v4 outperform Opus 4.7 with Taste — @AhmadAwais , CommandCode.ai
- ▶ 6:35 Ahmad Awais If you spend a billion tokens on Claude, on GPT, on DeepSeq, you discover a lot of discrepancies across the stack and different full stack systems like DeepSeq is not that good at code code for some reason.
Scaling Past Informal AI - Carina Hong, Axiom Math
- ▶ 25:33 Carina Hong In fact, like, you know, GPT found a proof to an unsolved Erdos problem, and our competitor Harmonic, you know, Aristotle, um, you know, verified it.
- ▶ 29:05 Carina Hong It was, um, rolled out, I think, by, um, Berkeley and Meta researchers in 25, and they found, I think, whatever version of GPT they evaluated does, like, pass one, like, 3.6%.
- ▶ 47:11 Carina Hong And then I just try to, um, you know, use a non-deterministic LOM, uh, like GPT say, to try to get the full proof for that.
- ▶ 50:53 RJ Haneke I mean, the counter argument is, oh, you know, like we just do a lot of good RL and, you know, we've seen, uh, GPT, you know, solving, you know, I think some RDoS problem and like whatever.
Inside xAI: Building Grok Imagine in 3 Months, Videogen vs World Models, and Video Agents— Ethan He
Devin’s 80% Moment: Background Agents, 7x PRs, & End of Hand-Held Coding — Walden Yan & Cole Murray
- ▶ 3:52 Walden Yan But I also just think that a lot of the recent leaps, uh, especially, you know, you look at like models like Opus and latest GPT models, they are reaching levels of autonomy where people are actually fighting that they actually can't just…
- ▶ 6:44 Cole Murray Comparing GPT and Claude as both of them are going through it.
- ▶ 56:41 Cole Murray Kind of on the topic of like visible slop, a pattern that I see a lot of cross GPT models specifically is backwards compatibility, um, at all costs where it's doing these weird import exports so that it doesn't have to modify, uh, the…
🔬How GPT‑5 derived new results in theoretical physics and quantum gravity — Alex Lupsasca, OpenAI
- ▶ 57:50 Alex Lupsasca Actually, my experience as a trained professional physicist working on my own research using GPT now is that I would say there's two key ways in which my research has completely changed. 2 times in the scene
- ▶ 1:19:02 Alex Lupsasca So I think the models are smart enough now and have enough background knowledge that, you know, for this paper, I'd say GPT is about as good as me at finding the next thing to ask.
AI-Native Engineering: 100% adoption, 5x search throughput, unlimited tokens — Mikhail Parakhin
- ▶ 12:40 Mikhail Parakhin Like this, this for me, for me, actually, the important metric is the ratio of budget spent during code generation versus, uh, spent, uh, expensive tokens like GPT, uh, or, uh, DeepThink from Gemini, you know, checking on PR reviews.
⚡️ How to turn Documents into Knowledge: Graphs in Modern AI — Emil Eifrem, CEO Neo4J
- ▶ 15:41 Shawn Wang GPT as well, but, uh, no, no, every chart in both databases.
The Stove Guy: Sam D'Amico Shows New AI Cooking Features on America's Most Powerful Stove at Impulse
- ▶ 33:30 Sam D'Amico and, you know, as long as I don't get one-shotted by GPT, or, you know, other stuff, I avoided the whole four-oh.
Claude Code for Finance + The Global Memory Shortage: Doug O'Laughlin, SemiAnalysis
- ▶ 1:00:11 Shawn Wang And they, they had human experts do the tasks and, as well as GPTs, and here's the results, right?
Inside AI’s $10B+ Capital Flywheel — Martin Casado & Sarah Wang of a16z
- ▶ 36:15 Shawn Wang Uh, minor disagree only because I, I'm pretty, like, have pretty high confidence that basically OpenAI will always release a GPT-V and a GPT-V codex. 2 times in the scene
⚡️ Reverse Engineering OpenAI's Training Data — Pratyush Maini, Datology
- ▶ 2:56 unnamed speaker No, we're going to go into like the GPT training data stuff, but I'm just kind of curious what, what, what gets posted there. 2 times in the scene
- ▶ 6:06 unnamed speaker Okay, so let's, uh, tell us about the seahorse emoji and GPT. 3 times in the scene
- ▶ 15:26 Pratyush Maini And so this was, like, quite interesting for me, because what this suggests about the GPT training data is that the self-reflection data has now actually become pretty much core to the training of all frontier models, because we're seeing…
🔬 From Red Teaming GPT-4 to Automating Drug Discovery: The Future of AI in Science — Andrew White
- ▶ 17:53 Andrew White Basically the, the best model, um, whatever, Opus seven or GPT 10, like, um, it really can only propose the first experiment, maybe slightly more clever, but at a certain point you just need information, right? 2 times in the scene
Captaining IMO Gold, Deep Think, On-Policy RL, Feeling the AGI in Singapore — Yi Tay
[State of RL/Reasoning] IMO/IOI Gold, OpenAI o3/GPT-5, and Cursor Composer — Ashvin Nair, Cursor
- ▶ 24:17 Ashvin Nair I think a lot of people didn't really think of GPT or GPT-II as something that was like super compelling probably.
Steve Yegge's Vibe Coding Manifesto: Why Claude Code Isn't It & What Comes After the IDE
- ▶ 36:09 unnamed speaker You don't get any respect if you're a GPT rapper, but like you start, like people are starting to be more productive and like actually develop
The Future of Email: Superhuman CTO on Your Inbox As the Real AI Agent (Not ChatGPT) — Loïc Houssier
- ▶ 27:09 Shawn Wang If you, if you think about GPT wrappers,
Terminal-Bench 2.0: the most impt coding agent benchmark of 2025 gets a v2! Launch + Q&A w/ founders
- ▶ 17:15 unnamed speaker So let's say I want to run GPT or quad on this task and train on it. 2 times in the scene
Why RL Won — Kyle Corbitt, OpenPipe (acq. CoreWeave)
- ▶ 40:32 Alessio Fanelli Is GPT-V the first model that had a prompt optimizer by one of the larger labs?
DevDay 2025: Apps SDK, Agent Kit, MCP, Codex and why Prompting is More Important than Ever
- ▶ 4:56 Sherwin Wu Candidly, we've actually been trying to do this, uh, you know, a couple of times with, uh, last dev day with GPT, two dev days ago with, um, I'm sorry, two devs ago with GPTs and plugins, uh, which was, I think, not tied to a dev day.
The antidote to AI fatigue — Answer.ai Solveit
- ▶ 46:08 Jeremy Howard And so, we were like, this is such a great way to turn, um, discussions into, into text, that we took on Andre Capathy's challenge, and his challenge was, hey, can anybody take my, let's build the GPT tokenizer video, and turn it into a…
Better Data is All You Need — Ari Morcos, Datology
- ▶ 27:01 Ari Morcos Oh, you know, GPTN is going to cost, you know, a trillion dollars to train.
Greg Brockman on OpenAI's Road to AGI
- ▶ 1:04:05 unnamed speaker So go ahead and write the GPT wrapper.
The AI Agenda: GPT5 leaks and the business of AI News — Steph Palazzolo, The Information
- ▶ 28:11 Stephanie Palazzolo You know, the past year we've seen a lot of progress, especially in reasoning models, but I think, like, under all that has been the fact that, like, GPT-IV has been kind of the, like, leading GPT model for, for a very long time, and… 2 times in the scene
🕰️ The Oral History of Windsurf (ft. Varun Mohan, Scott Wu, Jeff Wang, Kevin Hou, Anshul R)
- ▶ 49:41 Alessio Fanelli But Connor was the person and basically open AI at the time was like, we cannot release GPT because it's like too good and so bad.
- ▶ 3:28:28 Kevin Hou You could say, all right, we'll throw GPT, we'll throw Gemini at this timeline problem.
Cline: The Collaborative AI Coder
- ▶ 42:51 unnamed speaker It was like, I forget, like GPT they were using under the hood at the time wasn't very good at formulating these search blocks perfectly and it would fail oftentimes.
⚡️Ranking Agentic LLMs — Pratik Bhavsar, Galileo
- ▶ 5:26 Pratik Bhavsar Historically, what we have seen from our previous initiatives is that, okay, maybe the best GPT or best cloud is the top model, but there might be very small gap with the model just below it.
- ▶ 6:56 Pratik Bhavsar So first and foremost, I would say that I was expecting GPTs to perform the best because they kind of pioneered the tool calling, function calling, support, and they have been working on that for so long.
Information Theory for Language Models: Jack Morris
- ▶ 35:05 Shawn Wang Uh, I think we did an episode with Nicholas Carlini where he had an extraction attack, uh, on, on one of the GPT models and, uh, they got it fixed.
- ▶ 54:37 Jack Morris It's kind of like in, in physics, you know, when they try to measure these constants, like gravity, people tried to measure the rate of acceleration of gravity for a long time, or like those Greek guys, like back in, in the BC era, when… 3 times in the scene
Scaling Test Time Compute to Multi-Agent Civilizations — Noam Brown, OpenAI
- ▶ 29:05 unnamed speaker I feel like GPD models are kind of similar, right?
⚡️Launching AI Diplomacy: the hardest LLM Game Benchmark yet - Alex Duffy
- ▶ 26:07 unnamed speaker You have the GPT creative writing model?
⚡️Factorio Learning Environment: the ultimate Game Agent Eval — Jack Hopkins
- ▶ 10:09 unnamed speaker It was kind of on the level of GPT
Why Every Agent needs Open Source Cloud Sandboxes
- ▶ 15:39 Vasek Mlejnsky A lot of people are saying, like, if you are a GPT wrapper in 23, you were in a really bad position, uh, because, like, all the value will be captured by the AI labs. 3 times in the scene
The new OpenAI Agents Platform: CUA, Web Search, Responses API, Agents SDK!!
- ▶ 3:44 Shawn Wang So it's kind of like how GPT and the old series models are also unifying.
Beating Google at Search with Neural PageRank and $5M of H200s — with Will Bryk of Exa.ai
- ▶ 46:19 unnamed speaker How do you think about older GPT models then?
2024 Year in Review: The Big Scaling Debate, the Four Wars of AI, Top Themes and the Rise of Agents
- ▶ 3:45 Shawn Wang Even if it's never as prestigious, like it'll always be low status because at the end of the day you're manipulating APIs or whatever, but you're wrapping GPTs, but there's going to be an increasing stack and an art to doing these, these…
- ▶ 1:23:29 Shawn Wang In February, uh, you know, people were, uh, we were still sort of thinking about last year's Dev Day, and this is a three months on from Dev Day, uh, people were kind of losing confidence in GPTs, um, and I, I feel like that hasn't super…
Best of 2024 in Agents (from #1 on SWE-Bench Full, Prof. Graham Neubig of OpenHands/AllHands)
- ▶ 14:25 Graham Neubig So, like, GPT doesn't have very good, uh, air recovery ability.
Best of 2024: Synthetic Data / Smol Models, Loubna Ben Allal, HuggingFace [LS Live! @ NeurIPS 2024]
- ▶ 19:29 Loubna Ben Allal But, uh, and we can see this, for example, in the GPT family of models, how we went from just a hundred million parameters to more than a trillion parameters.
Best of 2024: Open Models [LS LIVE! at NeurIPS 2024]
- ▶ 18:34 Luca Soldani And what they found is, as a reaction to, like, the close, uh, like, of the existence of closed models, like OpenAI or Cloud, um, GPT or Cloud, uh, a lot of content owners have blanket blocked any type of crawling to their website.
[Paper Club] BERT: Bidirectional Encoder Representations from Transformers
- ▶ 4:03 unnamed speaker Whereas models like BERT and GPT were more, uh, use the fine tuning approach where they just, uh, it's more general purpose. 2 times in the scene
- ▶ 28:41 unnamed speaker Uh, it's GPT.
- ▶ 45:19 unnamed speaker Um, Eric, someone has asked, does BERT pre-training objective MLM, like mass language modeling, follow the same LLM scaling laws as GPTs?
Agents @ Work: Dust.tt — with Stanislas Polu
- ▶ 9:56 Stanislas Polu I think, uh, at OpenAI, there's always been like a large chunk of the compute that was reserved to train the GPTs, which makes sense.
- ▶ 14:38 Stanislas Polu Uh, he was, uh, he's always been, uh, mostly focused on, on training the GPTs, and rightfully so.
[Paper Club] Upcycling Large Language Models into Mixture of Experts
[Paper Club] Molmo + Pixmo + Whisper 3 Turbo - with Vibhu Sapra, Nathan Lambert, Amgadoz
- ▶ 1:09:45 unnamed speaker And the final one is my favorite, like, the state of GPT by Karpathy in, I think, twenty-twenty-three as well.
Building AGI in Real Time (OpenAI Dev Day 2024)
- ▶ 1:50:56 Kevin Weil Um, it's kind of a, a different take on what we did with GPTs, and GPTs are a little bit more long-lived. 2 times in the scene
llm.c's Origin and the Future of LLM Compilers - Andrej Karpathy at CUDA MODE
- ▶ 1:51 Andrej Karpathy So, roughly a year ago, I was trying to add a video to my YouTube series, and I was trying to teach people, um, LLM training, GPT training, and so on, and I was basically hacking on that, and GPT trying to get it to work. 2 times in the scene
Building AGI with OpenAI's Structured Outputs API
- ▶ 39:26 Shawn Wang I have a very simple one, which is a lot of people try to do, use GPT as judge, right?
Answer.ai & AI Magic with Jeremy Howard
- ▶ 36:55 Jeremy Howard Um, so that should be a good commercial opportunity for them if they can figure out what those tasks are.
Breaking down the OG GPT Paper by Alec Radford
- ▶ 12:42 unnamed speaker Uh, and this is like, this is kind of a big, big gap in this work that, uh, GPT folks are going to fill.
Supervise the Process of AI Research — with Jungwon Byun and Andreas Stuhlmüller of Elicit
- ▶ 26:11 unnamed speaker It's, it's funny, like, I guess every major GPT version, you, you have, like, some big insight.
Why Google failed to make GPT-3 -- with David Luan of Adept
- ▶ 7:36 Shawn Wang The history of GPT, as far as you know, you know, according to you. 3 times in the scene
A Comprehensive Overview of Large Language Models - Latent Space Paper Club
- ▶ 1:08 unnamed speaker In fact, it's when the GPT era started. 2 times in the scene
- ▶ 27:36 unnamed speaker So, examples of this can be things like, if let's say, I ask GPT to explain the moon landing to a six-year-old in a few sentences. 2 times in the scene
- ▶ 35:39 unnamed speaker That means the same adapter is being used in, let's say, a general model, and also, let's say, a, uh, GPT model.
- ▶ 43:08 unnamed speaker So, if let's say I've got a GPT model, and I, I type in a particular prompt, um, and this GPT model sees some email, and then it outputs some sort of phone number that is supposed to be private, um, and let's say a user takes this and does… 2 times in the scene
- ▶ 49:53 unnamed speaker This is, this is of course the encoder decoder, uh, phase of things, uh, beyond the GPT stuff.
Making Transformers Sing - with Mikey Shulman of Suno
- ▶ 5:26 Mikey Shulman And so the same way you don't program into GPT, you don't say this is a noun and this is a verb, but it, it has implicitly learned all of those things. 2 times in the scene
Open Source AI is AI we can Trust — with Soumith Chintala of Meta AI
- ▶ 1:12:31 Soumith Chintala Maybe open source models are being as used as GPT is at this point in, like, all kinds of, in a very fragmented way. 2 times in the scene
The State of AI in production — with David Hsu of Retool
The Four Wars of the AI Stack - Dec 2023 Recap
- ▶ 1:12:57 unnamed speaker I thought, Riley, I thought, like, you know, in the early days of GPDs, 5 times in the scene
The "Normsky" architecture for AI coding agents — with Beyang Liu + Steve Yegge of SourceGraph
- ▶ 19:27 Shawn Wang So, so the context is that, um, I think a lot of research shows that like what one context utilization matters, um, based on model, like GPT uses the top of the context window and then apparently cloud uses the bottom better.
- ▶ 1:09:41 Steve Yegge Can you guys believe how amazing it is that the open source models are, like, competitive with, you know, GPT and Anthropic?
Powering your Copilot for Data - with Artem Keydunov from Cube.dev
- ▶ 33:14 Artem Keydunov I think I saw people use all kinds of models to be honest, it's usually GPT.
The End of Finetuning — with Jeremy Howard of Fast.ai
- ▶ 19:00 Jeremy Howard Um, but yeah, there was a bit of this stuff going on, and the problem was everybody was doing, and particularly after GPT came out then, everybody wanted to focus on zero shot and few shot learning. 2 times in the scene
RAG is a hack - with Jerry Liu of LlamaIndex
- ▶ 8:07 Shawn Wang It uses GPT to build a knowledge tree in a bottom-up fashion by applying, applying a summarization prompts for each node. 2 times in the scene
Generating your AI Media Empire - with Youssef Rizk of Wondercraft.ai
- ▶ 1:26:12 Youssef Rizk The, the GPT, like the fact that basically you can do anything. 2 times in the scene
Ep 18: Petaflops to the People — with George Hotz of tinycorp
- ▶ 57:52 George Hotz But now of course, you know, four inch screws and orange juice is in, is in GPT's training corp.