Shawn Wang: belief

Transcript context

…campaigns, um, maybe in tech, maybe this decade. Um, and, and then I think, you know, there was a lot of backlash as to, like, what specifically was real in the, in the videos that they launched with. And then they took 9 months to ship to GA, and now you can buy it [01:26:00] for 500 a month and form your own opinion. I think some people are happy, some people less so, but it's very hard to live up to the promises that they made. And the fact that some of them, for some of them, they do, which is interesting. I think the main thing I would caution out for Devon, and I think people call me a Devon show sometimes, because I say nice things, like one nice thing doesn't mean I'm a show. Um, Basically, it is that like a lot of the ideas can be copied and this is the always the threat of Quote unquote GPT wrappers that you achieve product market fit with one feature It's gonna be copied by a hundred other people So, of course you gotta compete with branding and better products and better engineering and all that sort of stuff Which Devin has in spades, so we'll see. April, we actually talked to Yurio and Suno Um, we talked to Suno specifically, but UDL I also got a beta access to, and like, um, AI music generation. We, we played with that on the podcast. I loved it. Some of our friends at the pod like play in their [01:27:00] cars, like I rode in their cars while they played our Suno intro songs and I freaking loved using O1 to craft the lyrics and Suno to, and Yudioh to make the songs. But ultimately, like a lot of people, you know, some people were skipping them. I don't know what, Exact percentages, but those, you know, 10 percent of you that skipped it, you're, you're the reason why we cut the intro songs. Um, we also had Lama 3 released. So, you know, I think people always want to see, uh, you know, like a, a good frontier, uh, open source model. And Lama 3 obviously delivered on that with the 8B and 70B. The 400B came later. Then, um, May, GPC 4. 0 released, um, we, uh, and it was like kind of a model efficiency thing, but also I think just a really good demo of all the, uh, the things that 4. 0 was capable of. Like, this is where the messaging of OmniModel really started kicking in. You know, previously, 4 and 4. 0 Turbo were all text. Um, and not natively, uh, sort of vision. I mean, they had vision, but not [01:28:00] natively voice. And, you know, that, uh, I think everyone was, fell in love immediately with the SkyVoice and SkyVoice got taken away, um, before the public release, and, um, I think it's probably self inflicted. Um, I think that the, the version of events that has Sam Altman basically putting a foot in his mouth with a three letter tweet, you know, Um, causing decent grounds for a lawsuit where there was no grounds to be had because they actually just used a voice actress that sounded like Scarlett Johansson. Um, uh, is unfortunate because we could have had it and we, we don't. lawsuit where there was no grounds to be had because they actually just used a voice actress that sounded like Scarlett Johansson. Um, uh, is unfortunate because we could have had it and we, we don't. So that's what it is and that's what the consensus seems to be from the people I talk to. Uh, people be pining for the Scarlett Johansson voice. In June, Apple announced Apple Intelligence at WWDC. Um, and, um, we haven't, most of us, if you update your phones, have it now if you're on an iPhone. And I would say it's, like, decent. You know, like, I think it wasn't the game [01:29:00] changer thing that caused the Apple stock to rise, like, 20%. And just because everyone was, like, going to upgrade their iPhones just to get Apple Intelligence, it did not become that. But, um, Um, it, it is the, uh, probably the largest scale rollout of transformers yet, um, after Google rolled out BERT for search and, um, and people are using it and it's a 3B, you know, foundation model that's running locally on your phone with Loras that are hot swaps and we have papers for it. Honestly, Apple did a fantastic job of doing the best that they can. They're not the most transparent company in the world and nobody expects them to be, but, um, they gave us. More than I think we normally get for Apple tech, and that's very nice for the research community as well. NVIDIA, I think we continue to talk about, I think I was at the Taiwanese trade show, Comtex, and saw him signing, you know, You know, women body [01:30:00] parts. And I think that was maybe a sign of the times, maybe a sign that things have peaked, but things are clearly not peaked because they continued going. Ilya, and then, and then that bridges us back into the episode recording. I'm going to stop now and stop yapping. But, uh, Yeah, we, you know, we recorded a whole bunch of stuff. We lost it and we're scrambling to re record it for you, but also we're trying to close the chapter on 2024. So, uh, now I'm going to cut back to the recording where we talk about the rest of June, July, August, September, and the second half of 2024 is news. And we'll end the episode there. Ilya came out from the woodwork, raised a billion dollars. Dan Gross seems to have now become full time CEO of the company, which is interesting. I thought he was going to be an investor for life, but now he's operating. He was an investor for a short amount of time. What else can we say about Ilya? I think [01:31:00] this idea that you only ship one product and it's a straight shot at superintelligence seems like a really good focusing mission, but then it runs counter to basically both Tesla and OpenAI in terms of the ship intermediate products that get you to that vision. OpenAI now needs then more money because they need to support those products and I think maybe their bet is like 1 billion we can get to the thing. Like we don't want to have to have intermediate steps, like we're just making it clear that like this is what…

Stored transcript either side of the excerpt. The highlighted words are the published quote; the surrounding text is unedited source, never generated.