GPT-4, every mention
76 scenes, the whole family · ← back to GPT-4
tap a year for its mentions
every year anyone Matt Turck 23Stanislas Polu 11Aravind Srinivas 10Francois Chollet 6Des Traynor 5Akilesh Bapu 5Mitch Trojanowski 4Azeem Azhar 4Zico Kolter 3Sharon Zhou 3
Verbatim, from the transcripts: the passages where GPT-4 comes up
AI Could Take Over in 2029. Is It Already Too Late?
- ▶ 22:50 Ryan Greenblatt And like, yeah, the field is very different, and I think things have sort of, there's sort of a post GPT-IV era, and then there's a more recent, like, post wide adoption of coding agents era, and then probably soon there's going to be, you…
How to Build Autonomous, Long-Horizon AI Agents | Basis
- ▶ 14:51 Mitch Trojanowski Yeah, I think back then, I think that I believe, if I remember correctly, that was like GPT-IV, maybe IV turbo, um, at the time.
- ▶ 14:51 Mitch Trojanowski Yeah, I think back then, I think that I believe, if I remember correctly, that was like GPT-IV, maybe IV turbo, um, at the time. 2 times in the scene
- ▶ 16:29 Mitch Trojanowski Before, if you put anything in the, like, 80,000 tokens in the GPT-IV Turbo, it could not understand it.
OpenAI's Dan Roberts: Why AI Can Now Make Discoveries
- ▶ 7:58 Dan Roberts You could go and look at GPT-IV and, and see that there's, um, glimpses of that sort of useful behavior for, for the scientific process were, were already present.
- ▶ 39:33 Dan Roberts I wasn't at opening at the time, but if you imagine trying to like get small models to, to reason, you know, GPT one, GPT two, GPT three, and then GPT four to the again cartoon.
OpenAI Board Member Zico Kolter: Modern AI Is Just 200 Lines of Code
- ▶ 24:32 Zico Kolter It was GPT-IV. 3 times in the scene
State of LLMs 2026: RLVR, GRPO, Inference Scaling — Sebastian Raschka
- ▶ 16:58 Sebastian Raschka Uh, GPT 4.5 was rumored to be a bigger model, a bigger version of GPT four, but it was not very popular because it was too big, too expensive.
- ▶ 56:57 Sebastian Raschka Because I mean, back then it was GPT one, GPT two, GPT three, and now it's GPT four, 4.14 point, I think two or five, five, five, one, five, two, and all the models, they are iterating now, or even the same with deep seek.
- ▶ 1:06:16 Sebastian Raschka So what I use LLMs for is more like, Hey, I use actually the GPT-V Pro when I have written an article and put it in there.
Chasing Real AGI: Inside ARC Prize 2025 with Chollet & Knoop
- ▶ 3:51 Francois Chollet Um, that's, that's models like, you know, GPT-III, GPT-IV, GPT-IV. 3 times in the scene
- ▶ 11:44 Francois Chollet Because between, like, GPT-II and GPT-IV.
- ▶ 33:17 Mike Knoop And, um, this was sort of like the same consistent feedback, even as the models, uh, were getting better, you know, from GPT three to three and a half to four to four.
- ▶ 36:22 Francois Chollet It's like when GPT-IV was, uh, was released. 2 times in the scene
- ▶ 43:20 Mike Knoop Basically, since GPT-IV came out, there really was no open idea sharing at the frontier, um, and that felt like it was really setting progress back in the field to us, you know, um, and so we wanted to try and, uh, use this, like, prize as…
What You MUST Know About AI Engineering | Chip Huyen, Author of “AI Engineering”
- ▶ 1:03:38 Chip Huyen So we have, like, GPT-IV powered agent or, like, cloud powered agent.
Building the Easy Button for Generative AI | May Habib, CEO, Writer
Superintelligence, Bubbles And Big Bets: AI Investing in 2024 | Matt Turck & Aman Kabeer, FirstMark
- ▶ 0:46 Matt Turck Is GBD five going to be much better than GBD four?
- ▶ 21:46 Matt Turck Is, is GPT-V going to be much better than GPT-IV?
- ▶ 29:26 Matt Turck The, I think the price per token of, uh, GPT-IV dropped, uh, something like 90%.
- ▶ 39:17 Matt Turck In addition to having the best chips, they just released a model that was actually better than GPT-IV.
- ▶ 1:02:50 Matt Turck If, um, ChatGPT goes from ChatGPT-IV to ChatGPT-V, I guess there were rumors at some point that GPT-V was going to come out, uh, in the next few weeks, GPT-V slash Orion, but apparently that's not happening.
Making AI Work: Fine-Tuning, Inference, Memory | Sharon Zhou, CEO, Lamini
- ▶ 27:33 Sharon Zhou I think there's been a trend of mixture of experts, GPT-IV, uh, Mixtral are all a mixture of experts.
An inside look at “Mastering AI” | Jeremy Kahn, Author & AI Editor, Fortune
- ▶ 45:21 Jeremy Kahn There, I find it, you know, you can go to, uh, to Claude, or you can go to GPT-IV, and, um, it's pretty good at finding alternatives for that. 2 times in the scene
From Business to Warfare: How AI Affects the Modern World | Azeem Azhar
- ▶ 13:11 Azeem Azhar And so GPT-IV was definitely, in a sense, better employed.
- ▶ 23:11 Azeem Azhar The exponential increment from GPT-IV quality to GPT-V quality.
- ▶ 40:50 Azeem Azhar So we're using GPT-IV to look at GPT-III to figure out better explainability, which we can then hopefully apply to bigger models later on. 2 times in the scene
From Tiny Romanian Startup to Global AI Automation Leader | Daniel Dines, CEO of UIPath
- ▶ 48:33 Daniel Dines But in the end, we will beat, you know, any performance of GPT-IV or whatever, and they would be extremely reliable.
- ▶ 51:54 Daniel Dines So maybe they start with GPT-IV, but who knows, maybe code is better for a particular task or not. 2 times in the scene
How Airtable is Ushering 500,000 Organizations Into The Era of AI | Howie Liu, CEO of Airtable
- ▶ 8:56 Howie Liu Like you try, you know, chat GPT with GPT four, you try a cloud three Opus, like whatever you want, and you can ask it like human like questions. 2 times in the scene
- ▶ 30:38 Howie Liu To the model, maybe GP four or GP five will be good enough to like really be the one that kind of breaks out.
AI, Data and Blockchain: a VC perspective | Tomasz Tunguz, Founder of Theory Ventures
- ▶ 14:35 Tomasz Tunguz Uh, the, the, we ran this analysis, I mean, if you look at the pricing page for, like, OpenAI, and you look at the 4.5 turbo, and the cost per inference, and compare it to the next most recent model, there's a 160 X difference in pricing.
Navigating the AI Landscape: A Survival Guide | 2024 MAD Landscape with Matt Turck and Aman Kabeer
- ▶ 19:56 Matt Turck And, um, how much of a jump GPT-V is going to be in terms of overall performance over GPT-V, I think is going to be, uh, super interesting, very much like a bellwether of like where the industry might be going.
Is AI a platform shift or a paradigm shift? With Benedict Evans
- ▶ 28:10 Benedict Evans And chat GPT three was there and four is here.
2024 will be the year of ENTERPRISE AI | Florian Douetteau, CEO of Dataiku
- ▶ 20:46 Florian Douetteau These kind of things that are important for the enterprise when moving things to production, we see quite a bit of usage patterns where people would start GPT-IV for design and then decide to move to
How Intercom transitioned to being AI-first | Des Traynor, Co-Founder of Intercom
- ▶ 3:04 Des Traynor And then when ChatGPT four, when we got our first access to that in March, that was when we started to realize shit, this thing,
- ▶ 18:40 Des Traynor So Finn's built on GPT-IV, by the way, we've, we tried, we wanted to build on a three, 3.5, but, uh, it didn't, it's still, I remember back when we used to talk about hallucinations, like four was the, was the sort of the perceptual change…
- ▶ 21:48 Des Traynor We use four because it's best, but like the very second there's a better model out there, we'll do it.
- ▶ 24:08 Matt Turck So where are you in that exploration of, um, GPT-IV versus Claude versus Llama versus building your own thing, and how do you think about it? 3 times in the scene
- ▶ 35:48 Des Traynor It's a, the latency is a little bit of a problem in a phone call, and like as in a good GPT-IV query might take three, 4:05 seconds, and that just makes for an awkward phone call.
- ▶ 39:44 Des Traynor So there's a, there's just a bit more, like I would say business leadership needed from a product manager working in this space, uh, to think about like the how's, when's and where's of the problems we tackle, how much money is currently…
How Nomic AI Is Driving The Open Source Revolution
- ▶ 9:06 Brandon Duderstadt Yeah, so it started as basically just a reaction to GPT for becoming, you know, increasingly closed source.
How Moody’s Analytics Is Using AI to Transform Credit Risk | Cristina Pieretti & Yimei Fan
- ▶ 8:14 Cristina Pieretti It's called Research Assistant, and basically what it does, it's, ah, we use, um, ChatGPT-IV as the LLM, although we've tried a lot of, we talked about in our prep, we, we, we have tried different, ah, large language models, and we've also…
- ▶ 11:10 Matt Turck So you chose GPT-IV, but you mentioned you looked at others. 2 times in the scene
The Race to Build the Ultimate AI Programmer | Poolside CTO Eiso Kant
AI vs. Call Centers: ASAPP CEO Gustavo Sapoznik on the Future of Customer Service
- ▶ 15:38 Gustavo Sapoznik You can say, well, couldn't GPT-IV, for example, make those predictions?
How to Ship Reliable GenAI Apps: Humanloop CEO on LLM Observability, RAG & Rapid Iteration
- ▶ 0:18 unnamed speaker So, you know, at its core, Humanloop's helping product and engineering teams build reliable applications on top of large language models like GPT-IV.
- ▶ 16:36 unnamed speaker I would say that like, it's still the case in terms of like pure, like raw performance, GPT-IV and OpenAI's models still have a little bit of an edge over their closest competitors. 2 times in the scene
- ▶ 30:16 unnamed speaker And then GPT-III, and now we're on GPT-IV.
AI Medical Scribe Breakthrough: DeepScribe CEO Akilesh Bapu on Healing Doctor Burnout
- ▶ 1:50 Akilesh Bapu Um, now with the release of GPD-IV, um, every human in healthcare is now taking AI super seriously, and everybody knows the practical applications of it, uh, which is a first, and that's something that, as startups, we weren't able to do,…
- ▶ 16:10 Akilesh Bapu And then we have our, uh, we have GPT-IV that's also used. 3 times in the scene
- ▶ 20:30 Akilesh Bapu And generate a node with GPT-IV.
Custom LLMs at Scale: Lamini CEO Sharon Zhou’s Playbook for Enterprise AI
- ▶ 13:06 Sharon Zhou Uh, GBD-IV.
- ▶ 35:18 Sharon Zhou Um, one of the earliest, you know, examples of that, um, his team has now built out all the major foundation models, including GPD three, four, Claude, uh, Lama, Lama two, et cetera.
Perplexity AI CEO on Dethroning Google & Redefining Search
- ▶ 6:07 Aravind Srinivas GPT 3.5 was being developed and, and, and, uh, GPT four was not even there. 3 times in the scene
- ▶ 24:52 Aravind Srinivas Um, as for like perplexity labs and the API, um, everyone's been talking about us as a rapper over chat GPT, GPT, GPT, GPT for models for open AI.
- ▶ 28:07 Aravind Srinivas Like what is people's incentive to come and use these models in, in a world where GPT 3.5 turbo and GPT four already exist. 2 times in the scene
- ▶ 33:35 Aravind Srinivas Um, everyone's going behind training GPT-IV. 3 times in the scene
- ▶ 36:17 Aravind Srinivas Today, it looks like, you know, they're, like, lagging behind GPT-IV.
Secure, Private, Powerful: Dust’s Vision for Enterprise AI Agents | Stanislas Polu
- ▶ 16:20 Matt Turck Like, is there like a group that works on, uh, whatever Deli and another group that works on, uh, GBD three or four?
- ▶ 25:24 Stanislas Polu And you can not only have a generic frontier model, Ghibli four and cloud, but also custom assistants that you define, and that will be, uh, focused on certain tasks with access to your company data, which makes them
- ▶ 30:35 Stanislas Polu As an example, it could be something that just creates an answer from a knowledge base, and maybe somebody will show that person that, oh, but you can also use GPT-Flow to translate it in Spanish, and they're like, oh, wow, incredible, and… 2 times in the scene
- ▶ 36:54 Stanislas Polu So GPT-FOR, so it's two K, Claude, a hundred K.
- ▶ 38:34 Stanislas Polu So the ones you mentioned, if you're paying customer, you have access to GP four and Claude, and we foot the bill, uh, really want to be in a business where you're not thinking about tokens. 7 times in the scene
Guardrails AI: The Playbook for Safer, Hallucination-Free LLMs — Shreya Rajpal Explains
- ▶ 20:09 Matt Turck Maybe, maybe I'm wrong, but like given the above, um, if I'm an enterprise and I'm thinking whether I want to, you know, work with, uh, GPT for whichever other proprietary sort of, um, you know, API kind of system versus, uh, grabbing… 4 times in the scene
From NLP Start-Up to Generative-AI Platform: Milos Rusic (deepset) Unpacks Product-Market Fit
- ▶ 13:15 Milos Rusic If you want to use an open source model from Hugging Face, or if you want to use, uh, GPT-free for,
Hippocratic AI’s Munjal Shah: Building the First Safety-First LLM for Healthcare
- ▶ 20:09 Matt Turck Um, so, uh, it's, uh, presented as a state-of-the-art model that, um, has outperformed GPT-IV on a hundred, uh, oh, five out of the 114 healthcare exams and certifications. 3 times in the scene
- ▶ 32:38 Munjal Shah You know, it is encouraging that GPT-IV was able to get a lower hallucination rate, even while the model got bigger than, than three and a half.
Reinventing Search with AI: Richard Socher on Building You.com & the Future of Google
- ▶ 15:15 Richard Socher Uh, so one of the other many, uh, advantages of Vue Pro is that you can actually use GPT-IV.
Startup to Industry Standard: Lukas Biewald Explains How W&B Scaled MLOps for OpenAI, NVIDIA & More
- ▶ 12:27 Lukas Biewald But then, um, we even, we were, like, shocked by how much adoption they've been getting lately, and, and, um, you know, people, I think, want to compare, um, you know, asking GPT-IV to, um, training your own model, and we, of course, want…
Entering the Data-Centric Era of Foundation Models with Alex Ratner, Co-Founder & CEO of Snorkel AI
- ▶ 4:01 Alex Ratner This is basically the idea that, you know, most of these foundation models, like a GPT-IV, five, six, seven, are going to need to be, um, customized on your specific data and knowledge and workloads.
- ▶ 8:11 Alex Ratner Um, I think, you know, if I had to take a bet and hedge a little bit, closed source models, like a GPT-IV, five, six, seven, whatever comes, are going to be very hard to match in terms of generalist capability for, say, consumer use cases…
- ▶ 13:54 Alex Ratner So we view this, this data development layer as critical, you know, on top of these kind of base LLMs, whether they're, you know, GPT five or four or whatever, you know, or they're an open source base, um, you have to use this data…
Long Term Memory for AI with Pinecone Founder & CEO, Edo Liberty
- ▶ 20:03 Matt Turck I want to deploy, um, GPT for in, uh, the enterprise, but for mission critical application, like chat, where I don't want it to like say stupid stuff to my customers. 2 times in the scene
A Conversation with Sarah Catanzaro, Amplify Partners
- ▶ 10:32 Sarah Catanzaro I mean, GPT four was released yesterday.
The 2023 MAD (Machine Learning, Artificial Intelligence & Data) Landscape
- ▶ 11:07 Matt Turck You know, all the things, right, like ChatGPT is basically the fastest growing product ever, and the GPT-IV that was released yesterday is just even more powerful now. 2 times in the scene