OpenAI's Big Week, The Pin Is Here, Cruise's Major Recall

0:01OpenAI hosts its Developer Day, narrows GPT use cases and promises to defend copyright claims. Cruise does a major recall. Have we reached peak subscription? And the pin is here. All that and more coming up right after this. Welcome to Big Technology Podcast Friday Edition, where we break down the news in our traditional, cool-headed, and nuanced format. We have a great show for you today. A quick show on a Friday. We're joined, as always, by Ronjon Roy. Ronjon, welcome. I'm ready to make my own GPT. Let's go. It's time, finally. Let's make our own GPTs. I'm excited to hear what you think, whether this is actually going to be a real thing or not, and how this actually changes OpenAI's competitive position after we talked about all the problems they might have last week.

0:47But first, I want to make a quick announcement. I'm pleased to announce the Big Technology Podcast, after three and a half years of being on the market, has finally hit 1 million downloads. So I want to say thank you to all of you, the listeners, who come here week after week and tune in to us. One Million is a Major Milestone, started this podcast in 2020, effectively as an excuse to speak with interesting people. And because you've been here week after week, we've just gotten better and better in terms of the access. And hopefully it feels more like a real show now, now that we're so deep into it.

1:20We have lots of really great CEO and other executive interviews coming up in the next few weeks. Maybe we'll tease them at some point this show. But again, this couldn't happen without you, the listeners. and not only have we reached the million download mark, but we had a gathering last night, a small gathering at the home of Campbell Brown, who's leaving Meta along with her and Joe Marchese of VC, who's been on this show before, helped us host it. We had about 50 people joining us to celebrate and it was amazing and we made a small announcement. It took three years to hit a million downloads and we're here to celebrate that.

1:57But I can announce that as of last month, We're now on track to do a million every year, which is so exciting. Thank you. Thank you. Yes, that announcement is that not only have we hit the million listener or million download mark, but we're now on schedule to do a million downloads, if not more, every year. And again, couldn't happen without you. So I want to say thanks to all of you, the listeners. And also a big thank you to Ranjan. Ranjan, you've been a major, major part of this. And it's great to be able to speak with you week after week about everything going on in the tech world. So thank you.

2:30million downloads this is exciting definitely exciting okay let's get into the news um so first of all open ai held its first developer day this week as sam altman came on stage was doing the whole steve jobs thing with all the apple stuff uh apple seeming stuff in the background i'm curious if you saw the parallel but the major news of the day and this is coming from cnet and it's actually it's quite pertinent from seeing that open ai looks for its iphone moment with custom gpt chatbot apps basically what open ai introduced was the ability to build chatbot apps uh with effectively um natural language where you say you're now a bot that's gonna you know the example sam altman gave was you're gonna give advice to startups in the mode of sam altman and just like loaded a bunch of his uh speech transcripts in there next thing you know it sounds like him and it can be done without any programming.

3:26And this is how CNET put it. It'll let you build special purpose AI apps using its technology. And with a new app store coming, that'll let you find or share these GPTs as the company is calling these customized artificial intelligence tools. OpenAI looks like it's hoping to have something of an iPhone moment. Ranjan, what was your perspective watching this? Do you think that this is starting to alleviate some of the concerns we've had about OpenAI or does it bring up new ones? What was your reaction? All right. So the first thing, when we say it was Apple-like, my favorite part of this week's keynote was this was Apple-like, but the old Apple.

4:04Because for anyone who's watched any recent Apple event nowadays, they're pre-recorded, super high gloss, that kind of developer community energy of people in a room excited is completely gone. They've purposefully taken it away. This This felt like one of those old events. This felt like everyone in the room genuinely excited live demos that actually worked and were cool. So all of those things, I think, really good for open AI, really interesting, good for technology in general. Specific to the announcement around these custom GPTs, I think this is a pretty big deal because being able to create narrow sets of information that solve narrow problems is actually one of the biggest limitations ChatGPT or other generalized tools have had so far.

4:52And then you get into, whenever you're trying to solve the entire universe of information, that's when it leads to hallucinations. That's when it leads to other problems. Now the idea that anyone can easily just upload some information and from that body of information, there can be a chatbot that only queries that body of information. That can be potentially huge. This is something that I've done manually and built tools and products myself. And it's really time consuming and difficult and you need developers. Now anyone can go in and do this. So I think this is pretty big and pretty exciting.

5:28These things are going to be extremely easy to build. I think the real question here is, is anybody actually going to want to use them? Like, why would I go use a Sam Altman startup bot? Or why would I go use a Canva bot, for instance, to create some graphic design material? if I could just go watch the Sam Altman talk on YouTube or just go to Canva and build exactly what I need with a broader menu of things? What's the use case? Yes. So this is from a technology perspective. I think this is really interesting. From a consumer and business standpoint, I'm a little confused. And that's because right now I think every business has this problem that they have all this information.

6:08So the idea of being able to just query only that information, avoid hallucinations, I think that's actually a huge opportunity. This idea of a consumer-facing app store where I will create my own GPT and then put it out in this market and someone will use it, I'm not 100 % sure exactly what the use cases are. but I still think the flip side of that is it will potentially unleash a whole new wave of creativity and people doing cool ideas and creating these use cases. Because again, I remember the first app I ever downloaded on the App Store was a thing that made it like you were drinking beer.

6:49Do you remember this app in the iPhone? No, I don't. I might have drank too much in this app. I'm just forgotten. No, no, no. Seriously, you would hold your iPhone up and it would use the accelerometer. And as you turn the iPhone around, it would look like the beer was disappearing. The most ridiculous thing. So the idea that that would lead to a multi-billion dollar business in transform technology clearly wasn't easy to forecast. So I think this idea of opening this up, making the, and again, the other thing OpenAI is so good at, separate from the underlying models and big technology products that they build, is they're great at user interface and experience.

7:30That's why ChatGPT was the revolution it was. It's just easy to use. It's good. It's nice to use. It's delightful. So I think if they can do that and really let anyone experiment and play and come up with these use cases, I think it gets pretty interesting. Okay, so here's the real question. Well, one of the main questions I came away thinking about, which was that there's been this long debate of whether it even made sense to build a startup in AI, because effectively what you were was this thin wrapper on top of OpenAI software. and people said well if you're a thin wrapper on amazon web services you can build at a pretty effective sass company right and so that was the argument to say go and build away on these on these platforms but the thing is open ai just seems like it has an interest in getting more and more deep into what people are using its software to build trying to get deeper into their territory and something like this really makes it easy for, you know, more competitors to come and compete with the startups that are out there today.

8:36And okay, maybe that creates more companies, but like, let me just give you this example, right? There's this company character AI that I like to pick on a lot because it seems like it's the least defensible business in the entire AI world. And it was like, you know, given a$1 billion valuation led by Andreessen Horowitz, of course. And you can like talk to historical figures, right? So if you want to go speak to Abe Lincoln, you can go into Character AI and speak with him. I bet with this new GPT creator, you can just upload like, you know, all the open Abraham Lincoln speeches and you can make an amazing Abraham Lincoln bot.

9:14And what are you going to go to Character AI? You could probably build a better Abraham Lincoln through this stuff. So what do you think? Does this sort of settle the question of like whether thin wrapper startups are actually valuable and tell us no, they are not? Thin wrapper startups are dead. That's it. I mean, I think they, Sam Altman even said that, you know, this is just the best way to build the technology and to scale overall consumer utilization. So, I mean, it's whether they want to or not, this is the way the industry is going to go. And I would say for a while, there has been that feeling in the entire market that these thin wrapper startups will die.

9:50And yeah, we can go make our Abe Lincoln bot right away. We can take all the big technology transcripts and make Alex bot pretty quickly. And yeah, I think, but that's the big question that I thought made this announcement really exciting. We have talked a lot about what are the business models around generative AI. There's so much excitement and so many high valuations, but it's still so unclear around what's going to be a sustainable business model. And I think this was a big moment in defining what the next few years are going to look like. And to all the thin wrapper startups, I say good riddance.

10:31Let more people build with this stuff. But the question is going to be how they're going to build. And that leads us to the second core part of this announcement, which is that OpenAI has effectively guaranteed to take on your defense costs if you get sued for copyright while building with some of its tools. So this is from TechCrunch. OpenAI promises to defend business customers against copyright claims. As part of a new program called Copyright Shield, OpenAI says that it'll pay legal costs incurred by customers, specifically of its developer platform and chat GPT enterprise, who face lawsuits over IP claims against work generated by an OpenAI tool.

11:07I mean, doesn't that just incentivize people to just go steal reams of copyrighted content and then train these custom gpts or anything effectively using open ai software and not give a damn about the fact that you know you're using material that has been protected legally from being stolen so i'm curious what you think about this it's sort of concerning to me to see that this program is just not only not only that it exists but it's marketed as a way to get people to build an open ai software i'm going to say yes to all of the above, because I do think it does create that incentive. And I think it's going to create some really complex, convoluted, murky copyright situations.

11:48One of the use cases that I already had been playing with before it was this easy was if you upload a bunch of analyst reports around specific companies from different banks and then try to query them, you know that's something very personally useful but am i allowed to build something like that if these are you know am i allowed to take all the knowledge these other entities create and then just create this layer that other people can use and then potentially monetize it if i'm putting it on the gpt store i think this stuff is going to be really problematic and and i think they come out with this announcement today because micro or adobe had come out with a similar announcement i believe Microsoft itself had I think they're just doing this assuming they're gonna potentially have to assume some cost here but otherwise people don't build in here and then it never takes off so they're just hoping that the GPT store becomes so big that they can just pay whatever they need to on the side and hopefully it won't be enough to actually disrupt the business this is good they're good this is one of those things that they announce and like when the rubber meets the road, we're really going to find out.

13:01It's not like one of those things that's like, aha, that was a nice feature. Like they're going to have to use this. And one thing I'll say is they've raised a lot of money, but you can't pay lawyers in Azure credits, at least not in my part of the world. Maybe, maybe that's the next thing, paying lawyers in Azure credits, paying all of your vendors in Azure credits. And by the way, this is a broader thing that's happening across the tech world right now. We have not only this, but there was a headline this week from Andreessen Horowitz who apparently made this submission to the U.S. copyright agencies, saying that billions of dollars in AI investments could be worth a lot less if companies developing the technology are forced to pay for the copyrighted data that make it work.

13:41This is from Insider, and this is from the A16Z writing that's quoted in the story. The bottom line is this, imposing the cost of actual or potential copyright liability on the creators of AI models will either kill or significantly hamper their development. Now, listen, I like using this technology. I've generally been totally okay with them using my work to train. And as this industry has gotten bigger, it just seems to me that like for them to come out and say we would lose potentially billions in value if we paid copyright, it's totally shifting my opinion on this because it means that the copyrighted material is providing billions of value.

14:22And if that is the case, there should be a deal. Don't you think? I thought this is one of the most interesting communication strategies that I think we're going to be able to look back at and think about, was it the right one or not? Because they're obviously trying to toe the line that this industry, this technology will not develop if they're hampered by copyright. But in doing so, you're right. They are making the admission that so much of the value created is based on copyrighted material. They're saying it out loud. So I think this was pretty clumsy. I think trying to scare us into thinking we're not going to get the technological development and promise because of copyright and liability, but then admitting it, I think this was pretty clumsy.

15:09So I think it's going to be, yeah, this is going to be one of the most, the legal side. If you're a lawyer, get involved in this stuff right now. Start learning about it because honestly, this is going to be the most interesting space in all of law for the next few years. So I will make a quick plug that I just got off another recording with Kathy Edwards, who's the VP of Search at Google. And we talked about them training on copyrighted material and whether publishers should opt in or opt out. So that's going to drop on Monday. We talked a lot about generative search, which I know that we like to discuss.

15:44Ronja, and I won't give it away, but we definitely got into the nitty gritty on that stuff. And we also talked about the big SEO article that The Verge had. I still have it up. It's an article is the people who ruined the internet about SEO folks. It was fun to ask Google about a story with that type of headline. So as we keep going on this, there was another AI story that caught your attention, which is there's this thing called the pin. And the pin is effectively an AI phone. It's from a company called UMaine. And this is a headline, Silicon Valley's big, bold sci-fi bet on the device that comes after the smartphone.

16:24This is what the time says. UMaine is billing the pin as the first artificially intelligent device. It can be controlled by speaking aloud, tapping a touchpad, or projecting a laser display onto the palm of a hand. In an instant, the device's virtual assistant can send a text message, play a song, snap a photo, make a call, or translate a real-time conversation into another language. I mean, wow, you kind of just like wear it on your, on your, like a lapel pin, but it can do all these things. Ranjan, talk a little bit about your perspective on this device, whether you think it's cool. Are you going to buy one?

16:59What's your perspective? I have so many conflicting feelings on this right now. This is, I think this is going to be one of the more interesting stories in consumer technology because on one side, I have been hoping that the form factor of a phone changes because I think over the last number of years, I actually remember I am now an Apple ecosystem locked in person. But for a brief period, I had, I think it was the first Pixel and the Pixel Buds. And Google Assistant is light years ahead of Siri, which Siri somehow still is absolutely terrible. And I remember I had the Pixel Bud earphones and it actually works so well where you can call up the write song on Spotify, you can actually do things and take actions just with voice without pulling your phone out.

17:50So I started thinking, I remember at the time, like a voice only phone, not having to carry around your phone, this could totally change the way. Like right now, I feel the idea that we're all walking around with this block in of tech in our hand is not going to be how things look in 10 years or 15. I hope it's not how they look in 10 or 15 years. So what is that next form factor? Is it glasses, which people have tried? So the pin idea is definitely a new one. So yeah, I am overall very excited about someone pushing very hard on changing what a mobile phone is and means in our mobile computing device.

18:28I just want to say, I cannot wait for the but here. The what? But the but the but is here. The but is here. The whole time you were talking that up. I'm like, there's a butt coming. There's a butt. There's a butt. The founders of Humane AI, and again, Humane has raised$230 million before even launching this first product. They raised from the likes of Tiger Global and others. So, you know, like the who's who of big money fundraising. And the launch video, it's this 10-minute video, was the weirdest thing I've ever – Speaking of keynotes or Apple-like launching and OpenAI and Sam Altman capturing the spirit of Apple, this was like whatever Apple ever did right in the past, take the opposite of that.

19:20It's the two founders are very somberly and weirdly in front of the camera. And then they start holding up the pin. They don't even say what you're going to use it for. They go right into the colorways of the pin. they start talking about how you can get a battery add-on for the pin already indicating that the battery life is going to suck like rather than just telling you why the hell would i wear a pin and i again they start to touch on the use cases there's a shorter one minute kind of promotional video that gets into it a little better where again you're like walking around instantly translating conversations you can take a video or a photo but just by telling it to you can play music.

20:02Basically, again, you can take actions that you normally would on your phone and just do it in this totally different form factor. But I strongly recommend people to go watch at least some of the 10-minute video. You don't have to watch the whole thing because you will be bored to tears. But I just don't understand. They've been working on this for five years. They've raised hundreds of millions of dollars. And just the video they came up with was the most energy-sapping thing imaginable. There was a meme of the guy in the video, the founder, co-founder tapping the pin and saying, make me look more excited.

20:35But, but it did, Ranjan, it's so like you to be like, listen, like, you know, the product looks cool, but let's take a little deeper look into the marketing to show, you know, basically take a look into the mental space of this company and see whether we think it's going to be successful or not. Because I think that actually matters. You know, if you're not able to get that across, it generally indicates that there's something else going on there. And it is interesting to, you know, this is the natural successor to the smartphone. It does the exact, and you're right, it does the exact opposite of the, it starts with why approach that Apple has always used with Simon Sinek has pointed out that if you're getting a marketing message, you always talk about why, not what.

21:14And they talked about what before why, and I don't think they even mentioned the why. No, and the original iPhone, aside from my beer app, which I think you all should look up and go see this classic history of technology moment. Again, I remember Steve Jobs, it was like some of the elements where it's not like the internet, it is the internet. It was showing that web browsing could be possible on the phone. Taking a New York Times article and actually showing that it's possible to read it comfortably on the phone. The snap to zoom, pinch to zoom, showing this thing that just you can instantly, magically enlarge a photo and zoom in on it these were all the exactly the why and they made you want it and it made it look cool whereas whereas like if i were to just read text about this pin i probably was more interested until i saw them not be excited about it at all and then uh yeah now it's it's definitely not going to be my main phone i don't even know if i'm going to buy it $699, by the way,$699 and$24 a month for the service that's, I think, routed through T-Mobile.

22:22But there were a few cool things that I liked. For instance, this screen that you can project on your hand. Just like basically you hold your hand out and this like laser thing projects a screen on your hand, which, you know, you tilt your finger one way or the other, and it allows you to answer a phone call or scroll through different music, which is cool. I just think maybe wearing a pin like that is a little bit doofy. I don't know if that's exactly what I want to be doing. That's the other problem too. Again, as you started this conversation, you said lapel pin. Who wears a lapel pin other than a politician during a primary series of like the American flag?

22:59So it's so different that it really is going to require people to be so excited about it that they're willing to put their coolness and reputation on the line to wander around with this thing. So, yeah, I think I'm so I'm just so confused. But again, the projector thing, if it works, it looks very cool. Yeah. This is the future. This is the future I want. But this is what, like, I'm ready for some massive transformation in consumer technology. It's been a while. But, yeah, I cannot definitively say yet whether I will buy it or not. I might wait to see one, hopefully, from someone else. But I might buy it.

23:46I might buy it. I'm not decided yet. I don't think I'm going to buy it. It's expensive,$6.99, and I just bought the new iPhone. So, but I do think that just like, I don't know if they sent me a demo, I would absolutely try it out. I'll put it that way. Yeah. I think the only thing I was thinking about is I got the Apple watch ultra recently and it was the first time I actually got the, with T-Mobile, it's like five bucks a month, actual service, cell service on the watch. and I actually will leave the house now sometimes without my phone and only the watch and in a way maybe the Apple Watch Ultra kind of is like the pin because you can do things with it you can talk to it you can play music on it so so and which is good which is which reminded me that there is a future after the phone but maybe maybe it is the pin maybe it's the pin I think it's the brain chip obviously it's the brain chip all right brain chip so speaking of Elon Musk um and technological breakthroughs the long-awaited grok ai which is the ai that Musk has been developing came out or is in limited preview right now and the major criticism of it is that it's a bit cringy and has lots of boomer humor and it's not actually funny though it tries to be and Sam Altman open eye AI CEO decided to do two things at once.

25:15First of all, he tried to demo his new GPT bot builder, and two, tried to take down the co-founder of the company that he is now CEO of, aka Elon Musk. So he does this tweet that says GPTs can save a lot of effort. And he puts this instruction into GPT Builder, be a chatbot that answers questions with cringy boomer humor in a sort of awkward shock to get laughs sort of way. And then GPT Builder responds, great, the chatbot is set up. Its name is Grok. How do you like the name? Or would you prefer something else? I mean, just an absolutely ruthless takedown. And then Elon Musk responded. And he goes like this, he goes,

26:25Okay, that was Musk's response. There's one more line of Musk's response. I'm going to ask you this. Do you think that Musk wrote that response? That's a good, I was at, just as you were saying it, I was like, so you can taste the bark. That's good. That's actually pretty good. If Grok wrote that, maybe I'm going to have to give Grok a little bit of credit. Cause I, I, it's rare. I say this on this podcast. I agree with Elon Musk. One of the, the most annoying things about the way Chachi BT, Bard, even Claude everything. it's like everything is so defensive and hedging itself and saying, but I'm just a chat bot.

27:05I do not do these things. I don't know humor. I don't know, you know, so on that side, I think there's a space for something that is a little funnier, but again, maybe that could just be a custom GPT and you can actually just train it on actually funny things and that'll be enough. um yeah so so i think there's room for improvement in terms of all of these chatbots the way they've been so defensively programmed to date but again i what grok is and what it's going to be is it going to just kind of be like a rumble or something that's kind of like the anti-woke uh anti-woke technology version of some other platform i i have a feeling it could go in that direction more than any kind of like big transformative technology.

27:55Right. And so, by the way, that tech comeback was written by GPT, but was written by Grok. Musk admitted it. So here's the thing. Or a team of copywriters that came up with it and then attributed it to Grok. Well, in which case they should all be fired. It wasn't that funny. Here's my thing. If you want to build a piece of technology that's a counterweight to what they're doing in the mainstream and wants to not be more edgy, but just be able to go places that the current technology won't, then you should build an equally good piece of tech and just have it go there. And it doesn't mean program it with political views, but just like, you know, when it gets topics where there's room for debate on it, like actually decide to like lay out the debate as opposed to ChatGPT, which so often will take one side.

28:46So to me, Elon Musk did hire lots of great people to build this. But if his answer is this sort of edgelord that is supposed to, it's like cringy funny as opposed to what I just laid out, I don't see it working. Yeah, I think that's a really good point because we have talked about this a lot, that the team he's assembled for this X.AI, he's gotten some serious talent. But this is, again, a reminder that you're some leader in the field and whatever you have been working on, it's turning into a joke about a stick up someone's ass and tasting bark. Like, I mean, all the serious work you've done in your entire life, and this is what it's coming to.

29:30I'm curious how long people will stick around. Right. Okay, before we go to break, I want to address something that we brought up last week. So I was making all the cases for why OpenAI might be at a disadvantage. And I heard something this week that was actually quite interesting and makes the case that one of the disadvantages that I talked about is actually an advantage. And I'm curious what you think, Ranjan. So I said that OpenAI is limited to Microsoft and that will keep it out of, potentially keep it out of places like Google and Amazon where a lot of people are gonna be looking to build and it's the opposite of the anthropic model.