GPT-4, every mention

mentionsepisodesper episode tap a year for its mentions

Verbatim, from the transcripts: the passages where GPT-4 comes up

loading…

▶ 24:17 Karine Malata Um, I think GPT-IV came out during the batch, which was really interesting, and as we were running this pilot for this first customer, sort of OpenAI was coming out with new models in the middle of the pilot, which was changing, one, our…

▶ 6:10 Mukund Jha GPT-IV was the, the first one that we sort of started looking at.

▶ 4:10 Ian Fisher And then GPT four four comes out and it just blows you out of the water.

▶ 20:37 Garry Tan and they were doing fine-tuning with RL, but then, ah, yeah, GPT-IV 2 times in the scene

▶ 2:08 Jake Heller Because we were so focused on large language models and were researching deeply in this space, we got really early access to GPT-IV.
▶ 11:45 Jake Heller Uh, as soon as we got access to GPT-IV, it was like the first thing that we did.

▶ 38:34 Nick Joseph I think, like, the original, like, GPT-IV had, like, I think it was 86.4% was its MLU score.

▶ 0:36 Harj Taggar Bob was an early engineer at PayPal, an early executive at Palantir, and was recently chief research officer at OpenAI, where he led the development of ChatGPT, GPT-IV, and the O-One reasoning model.

▶ 16:07 Tom Blomfield I tried GPT-IV just a couple of days ago, and honestly, I wasn't yet as impressed. 2 times in the scene

▶ 4:34 Bob McGrew And then, uh, after GPT-ONE seemed successful, we brought in the ideas from Dota and from the robot hand of training at, you know, larger and larger amounts of scale and training on a really diverse set of data, uh, and looking for…

▶ 5:29 Diana Hu Up to GPT-IV, which is this company had built, it actually put some constraints and was able to automate a lot of the schematic design that you as a human had to design what components it needed to go on the design, and to some extent the…
▶ 9:24 Jared Friedman The, the other thing I think is interesting about this example is like during the batch before a one came out, diode had tried to do this with GPT four.

▶ 26:15 Garry Tan I guess one of the things that Jake Heller from Case Text, uh, you know, I think of him as maybe, I mean, not surprisingly, a YC alum who got access to both, uh, three, 3.5 and four. 6 times in the scene

▶ 10:36 Garry Tan There was sort of this moment there where we thought that, um, you know, people might just go to GPT-IV and distill out all the weights. 3 times in the scene
▶ 18:02 Jared Friedman Just by swapping out GPT for, for, for a one.

▶ 3:08 Diana Hu There's been a lot of ways for AI and multiple AI winters, but this one actually, GPT, 3.5 and then four blew out of the water a lot of tasks.
▶ 18:35 Garry Tan And what he was saying is the best way to use whatever GPT-IV, the, you know, latest closed source models that are most expensive and have the most parameters, uh, just think of it as a prototyping tool. 2 times in the scene

▶ 6:45 Jake Heller So we were really lucky to get early access to GPT-IV before it come out. 3 times in the scene
▶ 12:34 Jake Heller So this is gonna sound probably insane to most people who hear it, but I think the GPT technology is under hyped because what we see in GPT-IV
▶ 15:25 Jake Heller Of course, like GPT-IV itself is really incredible, but how do you engineer it so that you can have thousands of users on the same time and it's reviewing millions of pages of text all at the same time?

Turn any episode into a week of clips.