OpenAI’s New Structure, Apple Replacing Google With AI?, Cheating With ChatGPT

0:00OpenAI shakes it up again with a new structure and a new top executive. Apple's Eddie Q says the company may replace Google with AI. And everyone in college is cheating with ChatGPT. That's coming up on a Big Technology Podcast Friday edition right after this.

0:19You're used to hearing my voice on the world bringing you interviews from around the globe. And you hear me reporting environment and climate news. I'm Carolyn Buehler. And I'm Marco Werman. We're now with you hosting The World Together. More global journalism with a fresh new sound. Listen to the world on your local public radio station and wherever you find your podcasts.

0:46Welcome 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. We're going to talk about this new structure at OpenAI and what it says about the company moving forward. Of course, you know, or maybe you don't, but this is new news coming out this week that OpenAI has abandoned its quest to become a for-profit, or at least have the for-profit make all the decisions, and the nonprofit will control the company moving forward. We'll talk about the implications of that. We also have this pretty fascinating statement from Apple's Eddie Q saying the company might substitute Google with AI, even though the company makes a lot of money from Google.

1:28And everyone, of course, is cheating. with chat GPT in college. Joining us as always on Friday is Ranjan Roy of Margins. Ranjan, great to see you. Welcome to the show. Everyone's cheating in college. It's just not worth that 120k anymore, Alex. Or is it even 120k now? That's probably more. Yeah. But of course, we should start with the big news that white smoke flowed from OpenAI headquarters this week as they named a new CEO of Applications. I guess I'm not introducing this segment very well, but it is pretty cool that there is an American Pope. Bobby Prevost from The Chi. He's really made it.

2:07He has. Did you see the Chicago Tribune headline? No. What was it? It was just Da Pope, like Da Bears. Ah, okay. Have you seen Conclave? Not yet. But his brother was saying, we play Wordle, we play Words with Friends, and I recommended that he watch the movie Conclave. So he's watched it. You've seen it? I watched it over a Christmas break with my family, and my parents and sister said it was really boring, and I was able to hold that over them this week as the Conclave began, and suddenly I was the most informed about how these things work. It was pretty good. It was slow, but it was very Oscar-y.

2:50okay so anyway i was personally excited about that news and i had to find my way to shoehorn it into this show but let's talk about fiji simo the instacart ceo she is joining open ai as the ceo of applications and i think this is interesting from an open ai standpoint because it is not just fiji simo joining as the ceo of applications which is very interesting because as we've spoken about in the past, this company is nothing but applications. I mean, of course, the models work, but their moat are their applications. But Sam Altman recently made Brad Lightcap into the head of day-to-day operations.

3:32And as he named Fiji the CEO of applications, he writes this message. In this new configuration, I will be able to increase my focus on research, compute, and safety. These are critical as we reach super intelligence. he also had to note that he will remain CEO of OpenAI. Kind of, we talk about how this company is very interesting, but also just filled with weird stuff. And this was definitely weird to me, to see that Sam is going to remain CEO, but he's starting to delegate a lot of his responsibilities to others, and like core company responsibilities. And he's also made CMO the person who is now the boss of Sarah Fryer, who was the OpenAI, who was the Nextdoor CEO and is now the OpenAI CFO.

4:29So Fryer comes over to OpenAI from Nextdoor to become the CFO. And now she's reporting not even to the CEO of OpenAI. It's weird. This is now an org chart podcast in addition to a podcast because we're trying to unpack this. It was interesting. And also to note from the blog post that OpenAI put out, Sam Altman will increase his focus on research, compute, and safety systems. So that's like his main area now. Brad's going to be heading day-to-day operations. And then Fiji is the head of CEO of applications. And so in some world, I actually see that as a good division of power and labor and skill set.

5:17So I think that actually does make sense. I think the most interesting part of this to me is actually on our recurring, is it the product or the model? This is kind of like cementing that they want to really focus more on the product side of things. Again, masters at building great UIs and great product experiences. but actually building a business out of it. Remember, their revenue forecasts are always just things of just incredulity. Like, I mean, on the current one, SoftBank is going to be a large, like massive part of their enterprise business. A year ago, or maybe it was only six months ago, they were going to be much more heavily into consumer revenue as opposed to enterprise revenue.

6:06So my read on this is they have not been able to kind of tell a clear business story. And now hopefully Fiji Simo will try to wrap her head around that and actually come up with one. Yeah, but it's interesting because Sam, you would think that he'd want to be on the core as the CEO. But instead, he's going to be on if you're saying that this is that they're going to focus on applications. He's saying I'm going to focus on the non-core if you go with your logic. So you have, it's a very interesting moment where you have a CEO of one of the most promising tech companies in decades saying that my core focus is not going to be on the core competency of the company, which is applications.

6:49What do you make of that? Yeah, but that's, I mean, it's logical. It's accurate. That's him. That's her. Like, that's what they both have done separately their entire careers. Again, she is a masterful operator that's like helped build out Facebook's ad business, was at Instacart, actually built out Instacart's advertising business very well. Like that's what she does well. And maybe now does that mean ads are coming to OpenAI and ChatGPT sooner than we thought? We should we could definitely discuss that. But I think overall, it's clear that's the part of the business that needs help. And to their credit, they found someone with a strong track record.

7:33I'm curious what you mean by needing help, because like Sam has said in testimony this week, OpenAI's ChatGPT has 500 million weekly active users. Those are the most public numbers. I feel like the numbers are more, at least he's indicated that the numbers are higher. To me, I think what Sam is saying is that, you know, you could sort of delegate this type of work. And his focus on research is saying the models matter more. And notice how he says these are critical as we approach superintelligence and not AGI. Right. So he's already he's now leveling this up. From AGI to superintelligence. That's all that matters.

8:11Yeah. Okay. I agree that the idea of like, he does believe there will be one model to rule them all. It will be the all powerful model. And he wants to drive building that. But I think recognizing again, like when 500 million users is great, revenue is going to be what matters at a certain point. Like net income will start to matter. the i maybe this is a sign of maturity that he's starting to recognize these kind of things like if you think about on one hand they've had incredible product rollouts but from a pricing standpoint it's all over the place like again 200 bucks is deep research part of it now deep research is part of the 20 bucks and i get this stuff is all fluid and everyone like gemini and workspace was 20 a seat now it comes with it like this stuff is fluid at many levels but But this is the kind of area they're going to have to get their house in order if they're going to actually become the business they say they are.

9:13So I don't think this is like a recognition by him that this is kind of like a throwaway role and still it's going to be one business, one model to rule them all that's going to bring us to ASI and that's going to be the end all be all. I think he doesn't probably, I agree, I don't think he wants to do that stuff. But I also think he found someone who can do that stuff. What do you think about the fact that the CFO is going to report into Fiji SEMO? That seems more than just a CEO of applications. In terms of stuff he probably doesn't want to deal with. I think the operational financing, the line-by-line financing, like, come on.

9:54Do you think that's what Sam wakes up and wants to do? Well, no. I don't think so. So I think I should take a moment to talk about Fiji because I've gotten the chance to interact with her a few times when she was at Facebook. So she's going to come in. She's going to be the CEO of applications. This is, of course, we talked about how OpenAI is a consumer business. This is going to be her tripling down on OpenAI's consumer business. You need someone that understands consumer applications if you have a lead and you really want to build on it. And she is that person. In my conversations with Fiji, I found a few things to be true.

10:32Number one, she's a pretty clear communicator and a very clear thinker about the role of product. We talked about on the Blue app on Facebook, even though it's had its ups and downs, the role of groups there. And I believe that groups has really reanimated the Facebook product in a way that few tweaks have in a long history since it's languished after the friend sharing moment. So I would say credit to her for that. The other thing that I have known about Fiji is that she is not like a Game of Thrones style corporate operator. She won't make a run for the CEO job. She will be very loyal and supportive of Altman.

11:13She's been on the board for a while, just as she was an incredibly loyal executive for Zuckerberg. And obviously, she was able to parlay that into the CEO of Instacart. And she, of course, ran the Facebook app for a while. The one thing I'll say is that Sam at OpenAI has really struggled to keep a number two. We know that Ilya left. We know Miro Murati left. We know that Greg Brockman had to take time off. There's that meme of the picture of these are the founders of OpenAI and they all disappear and Sam Altman has left. And he really needs stability in his lieutenants. And of course, he's been able to keep Lightcap for a while and Fryer and now Simo.

12:00And I think this is actually quite a positive sign that he's decided to bring Fiji in, because if anyone seems like they can ride the chaos of that company out, it's her. Yeah, I agree. I mean, I think it shows a level of stability. it shows. Again, Sarah Fryer, very senior executive at some pretty large companies. But I mean, Fiji Simo has been a very powerful executive at the largest consumer companies in the world. So I think I agree on that. It's going to be very interesting to watch how it plays out, how she even communicates her own role. Because again, the announcement came out in a kind of amazing way that she sent an email, I believe, internally at Instacart because it was going to be leaked into the press and publish.

12:47So she wanted to get ahead of the story, but she still didn't give a lot of clarity. We have Sam Altman's blog post on the OpenAI blog that starts to explain this a bit, but it will be very interesting kind of how she talks about her role with OpenAI publicly. Right. And so I gave all the spiritual reasons why Fiji might make sense for this job, but Ranjan, I'm actually curious to hear your perspective as, you know thinking about someone who was the instacart ceo or who still is who will be the instacart ceo for a moment is it interesting and of course she was on the board but is it interesting that somebody coming from a partner that open ai would have to instagram would have to integrate if it was going to be like the catch-all app for all computing she's going to run applications like it almost seems like fiji will say to the door dashes of the world and any other company that chat GPT might gobble up.

13:44Look, I was on the other side. As the Instacart CEO, I worried about traffic to my app and all this other stuff. But here's how I would think about it as someone in your shoes. You know, if this this to me is one of the biggest signs ever that OpenAI wants to be effectively the everything app with chat GPT or even the new internet. I mean, Altman, again, was in Washington this week testifying that he thinks AI could be bigger than the internet, which I guess is something you would say if you're doing full-time marketing, now that you don't have to have the CFO reporting to you. But it just seems to me like that is another signal as to where this is going, and I'm curious what you read into it.

14:25See, I disagree. I don't think this is WeChat-style everything app. My call on what this means for me, and it's the same thing. When she went over to Instacart, the big conversation was around. Instacart had started to build out their advertising business. DoorDash had, Uber had. Basically, all these consumer apps with thousands, or I mean, sorry, hundreds of millions of users that had commercial intent within them realized, we can start to inject advertising. So they started to build out these businesses. Do you know how much Instacart's ad revenue is right now? i don't have it right here do you yeah of nine uh billion dollars like they got up to one billion dollars so like she built that out from over the last few years it had started growing the foundation was there she scaled it so my call on this ads are coming to chat gpt in some way and in reality like i do believe there is going to be some kind of advertising model that enters this whole world.

15:33Perplexity has, I think, started to launch some ads, but they certainly talked about it. And they have like, they've been pushing an advertising model. Sam has said he doesn't want ads like Larry and Sergey once did. It's coming, I think. And if someone can, again, figure out a creative way to do it, I think she could be the one to crack what does advertising look like in an AI chat? Well, if this is the case, this has made me tremendously depressed because I don't know if you caught my conversation with Gary Marcus this week, but towards the end on Wednesday, we talked about how these things, people are opening up to them in a way they never did to Facebook.

16:18And what, like you're going to now start telling it all your intimate secrets and all of a sudden this AI that every night, we're going to talk about Mark Zuckerberg talking about how ai could be your friend or your therapist like if you use it for these purposes you're going to share more with chat gpt than you would with any service on the internet and that data is going to be used for ads i mean it's it's like maybe i'm naive and being upset about this because this was obvious we have a headline from the financial times in december 2024 saying that openai is exploring ad revenue but it just feels like ad revenue from a chatbot that many people are going to view as a friend, a companion, a therapist, is an extremely tricky road to go down because of the amount of data and intimate thoughts people are sharing with these bots.

17:07So I'm curious what you think about this. Am I right to be depressed? You're right to be depressed on it, but I'll also go with the naive to be depressed because come on, like, come on, in terms of the idea that OpenAI would not go down that road, given the sensitivity of a lot of the chat, I think. I mean, they almost cannot do that given their valuation, but I'm not sure ethically they would actually be against it anyways. But I think overall, it is going to be interesting to me that people are more personal because there are not ads. So then you get into a, would they become less personal if they're worried about ads?

17:54Like, I think maybe that does start to degrade the overall intimacy of the product a bit. But to me, I guess first, it feels like, I mean, that is inevitable or the ethical boundaries will never have stopped that and has never stopped advertising. And then also, even though we're not, like Google search has a more intimate knowledge of you than Facebook posts. But in the end, maybe I'm too cynical, but I feel advert. The ad ecosystem knows everything and anything about all of us. But even still, a lot of the ads that we get don't feel so tailored to us, don't feel personal. And wouldn't you agree that if they do this the wrong way, that this is a potentially catastrophic hit to their brand of course i mean definitely like and first of all like i feel meta to their credit even after ios 14.5 and like identity tracking was supposed to be more difficult it's like ungodly how good their ad algorithm is twitter and slash x worst ad algorithm imaginable so like people are able to do it and figure it out but i do agree that it could be a risk where okay let's say they start to get too weird and sensitive in terms of the ads that they show you given the type of data that they'll have on you in a way they have to be extra careful because one that can affect their overall like the utilization of the product but two in terms of overall privacy concerns remember at the enterprise level like it's always the starting point that people worry open ai is going to train on your data or doesn't respect privacy that they're branding on that and we've talked about this a lot, is not great.

19:43So I think they could blow any goodwill they have up if they go down the wrong path with this. And I also think it's worth pointing out that there are various degrees of targeting that you can do with each degree higher becoming more lucrative for you. So let's just talk this out. The bottom level here is just, you know, sort of broad targeting the type that you do in tv so that could be like you know targeting for basically location maybe age gender but really audience and frequency right the reach and the frequency that's the basic level of targeting and chat gpt can definitely do that with 500 to let's say 800 million users that it has now the second level is intent-based targeting right and this is similar to Google.

20:31So that would be OpenAI potentially using the chats that you are engaged in at the moment to surface things to you. For instance, like let's say you say, I need some help researching a car that I'm interested in buying. Well, in the moment, OpenAI could say, all right, well, here's like Hyundai and Toyota, but you might also be interested in Nissan, which is a partner of But I found a good deal for you. Okay, so that wouldn't be too different from Google And I don't I think it could get away with it, but there's this third type of targeting That's the most lucrative that is also I think the most dangerous to open AI And our and maybe I'll even say our society if they end up going this route And that is building and it's totally possible to do this and advertisers will want you to do this building a psychological or some sort of persona profile of the individuals based off of the entire history of chats that they have with you.

21:31Now, of course, that's tough to scale, but there are shortcuts you can take and algorithmic groupings you can do to put people in cohorts. But if you start to advertise to someone based off of the entirety of the personality and the history of the chats that they've shared with you, that to me is the red line, even though you can make a lot of money from doing that open AI. I just want to say right now to you, please don't do that. That is definitely not the move. Intent, fine. Building these like psychological profiles of your users, it's tempting. We know other companies have done it. If you do that, it's going to be a disaster.

22:10So please don't do it. I like listeners cannot see Alex is currently in Paris. He has a big bookshelf behind him, and I feel you're getting very existentialist today in the conversation. Well, what's not true about that, though? I agree. Well, on one hand, Meta can actually, I mean, extract what is in the image that you post and understand your emotional well-being based on what the captions you're posting are, what kind of messages you're sending. And all that data is already there and can advertisers certainly target based on that kind of like psychographic data and well-being. So I agree that almost by default, people's public posts are always going to be slightly more censored, hopefully, than what they're really entering to ChatGPT.

23:05But I don't know, between what people search on Google, between what people actually post publicly and that advertisers already have access to. I don't know. Maybe I'm too cynical on this one. But I think, yeah, I think you are. I think this is a degree deeper that you can go with chatbots. You know, if someone's telling ChatGPT, I'm depressed, I don't have plans this weekend. They know that you live in a certain area. they know what you typically search and you say hey I'm feeling sad because we know again that therapy is one of the top uses and it says well here's an ad for a movie or a restaurant that you can go to and uses that data to sort of exploit your current state I feel like that's that is dangerous territory to get into it's like you're there you're therapist you're talking to them and suddenly they're like, well, let me hold up this product.

24:03I think this can help you right now. You know what real self-care is? Hims. It's because you're bald. Hims. Hims. Hims doesn't sponsor the show, by the way. Affiliate model therapy. That's what the world needs right now. No, but I totally hear you that the companies already know a lot about us to begin with. But anyway, I'm kind of getting creeped out just thinking through these possibilities because it could get really dark. And now we know that OpenAI is going to move to a more traditional company structure or maybe not. This is from Bloomberg. OpenAI walks back its for-profit plan and the nonprofit is going to keep control.

24:48So OpenAI is backtracking on its plans to become a more conventional for-profit company after facing mounting pressure from former employees, academics, and rivals, including the billionaire Elon Musk. The Chattachapi team maker said Monday that it's moving forward with an effort to restructure its for-profit division as a public benefit corporation, but the overall business will instead remain under the control of its nonprofit, a major shift in its plans that will effectively maintain the contours of how OpenAI is currently set up. So the nonprofit continues to control the company. This is from Brett Taylor.

25:22We made the decision for the nonprofit to retain control of OpenAI after hearing from civic leaders and engaging in constructive dialogue with the offices of the Attorney General of Delaware and the Attorney General of California. The one good thing, I think, is that the company structure is going to resemble a much more normal structure so that there's no more of this capped profit sharing deal. It's just going to be like you have normal shares in the company, but the nonprofit will continue to control it. What's your view on this, Rajan? Good? This one makes my head spin. Not quite as in a dire way as ChatGPT creating an advertising profile based on your deepest, darkest secrets.

26:05But, I mean, it just gets more and more convoluted again. Like, Matt Levine had a good, he was like, Note the nonprofit will continue to control the public benefit corporation and will become a big shareholder. it suggests that the nonprofit will be a minority shareholder. The nonprofit might own, say, 30 % of the economic value of the for-profit company, but will have super voting stock giving it control of the board. Like, basically, it's just nonprofit-y enough to be a nonprofit legally, but it's not really a nonprofit, but it's also not quite a for-profit. I think, to me, all of these decisions, the only thing I think I try to think about, because otherwise, again, I get a headache, is what does it mean for their funding and what does it mean for potential exits and outcomes?

26:56Because to me, they don't really operate by normal rules anyways. And remember, SoftBank, in that$40 billion funding round, they committed 30, but supposedly it can be reduced to 20 if they don't become a truly for-profit corporation. So I don't know, but it still wasn't even clear. did this new structure mean that SoftBand can renege on$10 billion? Well, this was definitely the week of great Sam Altman quotes. So first of all, here is the first thing he says about the structure. By the way, remember, he talks as if somebody else tried to make this structure transition. He says, I won't pretend that it wouldn't maybe be easier if we were a fully normal company, but the mission comes first.

27:41We believe this is well over the bar of what we need to be able to fundraise. So, you know, the mission comes first, but we were trying to go fully for profit, but the mission comes first. This is bizarre. I mean, I actually, I love it. I think that quote should be on a poster and that captures everything about OpenAI better than anything either of us could have said. So thank you, Sam, for clarifying everything, at least emotionally, even if not logically. Sure. Sure. I mean, I'm telling you, the mission comes first, but this helps us in our fundraising. That's it. That's OpenAI in a nutshell.

28:24I guess so. So speaking of the SoftBank stuff, he also said that SoftBank will not freeze the rest of its funding with the new structure. And then OpenAI, like we said before, announced it will remove a cap on the financial returns its investors can earn, a move likely to appeal to current and future backers. So yes, the fundraising will continue. Now let me say the however part. However, what is going to happen with Microsoft? This is from Newcomer. Microsoft wants to retain its rights to have access and ownership rights to OpenAI's model. Whereas OpenAI is trying to claw that back, the topic has cropped up numerous times between the companies.

29:05So this is still not settled, even though the announcement has been made. finally newcomer continues to write microsoft wants to amend an agi provision that gives open ai an out once it develops artificial intelligence general intelligence there are these issues were an obstacle to altman's planned for-profit conversion and they remain an obstacle in the new plan to make the business a public benefit corporation that's still controlled by the non-profit parent. No restructuring process can move forward until this is figured out. So what Newcomer is saying is this ain't over. It never is. It never is.

29:51He says the fact is that Microsoft has OpenAI in a type spot and knows it. At a high level, relations are still strong. Altman welcomed Microsoft CEO Satya Nadella to their offices the other week. Beneath the surface though, the competition is rising. Microsoft is no longer just a major investor in the company, but increasingly a competitor, especially in enterprise software. A source close to the matter said the changing dynamic is influencing the proceedings. Some at OpenAI worry that Microsoft is setting up such a high bar barriers to a deal as to make one impossible. So this is the only place I've heard of this.

30:29But assuming that this is correct, is it possible that Elon Musk wasn't the biggest barrier to this OpenAI restructuring? It's Satya Nadella and Microsoft who realize that they have a big say into what the future of OpenAI looks like, and they will hold it back from being able to move forward, which means basically holding it back from being to from being able to do anything else because of course the soft bank money will be diminished unless its terms are met what do you think that's actually i like this because the idea that masa is coming in 30 billion strong and like one would assume in any normal company that would give a very clear decision making capacity and very clear ownership and in reality because everything is so murky and muddled that Microsoft is still able to legally pull the strings in weird ways to benefit themselves.