Mark Gurman breaks down the Apple vs OpenAI lawsuit: smoking gun employee, Tang Tan's tactics, and what Apple already won

Speaker 1: The Germinator's back in the UltraDome getting us up to speed on what's happening in the thermonuclear war between Apple and OpenAI. And so he is here. Let's bring in Mark Gurman, the Gurmanator himself, to take us through his excellent reporting. As always, congratulations on the scoop. Take us through it. Friday, what was your day like? What happened? What's the headline?

Speaker 5: Friday was great, you know, end of the work week. But for me, there's really no end. It's Monday, Tuesday, it's Seven the whole days, twenty fourseven. Apple suing OpenAI. I didn't find it to be that big of a shock. I think things have been brewing for several months. This was a two year relationship that went south pretty much in the first year. Okay. I think it was doomed from the beginning. We can get into that.

Speaker 4: Yeah.

Speaker 5: But OpenAI had been pillaging Apple's hardware engineering ranks, cloud department, many other aspects of the company for over a year. 400 people hired by OpenAI, and people move between companies all the time. Yeah. But one company hiring 400 people from another company in Yeah. Such a short period of It's absolutely insane.

Speaker 2: And Apple and Apple even with all the the competition we talked about it from like Amazon and Microsoft and Meta has never gone through this kind of they've never been pillaged in this Yeah. It's not like it's not

Speaker 1: like Bezos got 400 people from Apple working on the Fire Phone.

Speaker 2: Everyone I've I've seen a million decks where we're like, we got this guy who worked on iOS at Apple and it's like says

Speaker 5: they worked on iOS at Apple.

Speaker 1: It's good

Speaker 5: for the resume. Yeah. Whether it's true or not, you know, every company likes to say, oh, we got this big wig from Apple, and then Yeah. You you ask people at Apple about this person, they're like, I've never heard of this guy. He didn't even work here. But that that that's besides the point. The the real story here is that Apple had known for a while now that there was probably some shady stuff going on when you have that many people. I mean the person running the hardware organization at OpenAI, a guy named Tang Tan, he was heavily respected when he was at Apple, but you know, once he left Apple it was wartime. This was a person who had been running the iPhone design team for many years, paired with Johnny Ive, who was the designer, you know, alongside Steve Jobs for decades. Ive's successor, Evan Tanke, who ran the ID studio, the industrial design studio for three years after Ive left, they set out to build the product to take down the iPhone. You know, you can ask yourself, is that even possible? But on the other hand, the fact that these people, you know, worked for Apple for so long and they wanted to create sort of the antidote to Apple, it's a scary thing for Apple. And that coincided with Apple facing its own AI struggles. Right now, their hardware engineering organization is going through the biggest transformation in decades with John Turnis stepping up to become CEO, and Johnny Srouji, the chips guy, now becoming in charge of hardware engineering himself. You know, he's a visionary when it comes to management and silicon, but the big question, is he a visionary when it comes to products? Obviously, Chernis will help with that as well. So there's some big questions at Apple right now. Obviously, it's been a very long time since they had a big new hit. The Vision Pro was a complete flop. And so this company combining the best AI, right, OpenAI, using some of Apple's best engineering talent, plus three ninety of their best friends from Apple. It's a scary sight for Apple.

Speaker 1: Yeah.

Speaker 5: You know, they maintain that this lawsuit was not about slowing down OpenAI. It's because they were able to findthey're not calling this, but I can't think of a better term. They knew some shady stuff was going on, and they found a smoking gun, right? Smoking gun was this young iPhone engineer who went from Apple to OpenAI at the beginning of this year, and it's a crazy story. He kept a MacBook, he had this relationship. They're not saying if it was like a romantic relationship or whatnot, but I have to imagine that probably had a component He was working with this girl who was still at Apple who eventually went over to OpenAI, and he had found some flaw in their cloud network and was working with her to get stuff even though he was still working at OpenAI. So, you know, you can make a Netflix special out of this, or you can make an Apple TV special out of this, but they'll never go for that.

Speaker 1: Maybe it'll be on Amazon A

Speaker 5: lot of interesting stuff. Yeah. And then they kept digging. They found a bunch of stuff lawsuit around this guy who worked at Apple for twenty five years. Everything that I'm told about Tang Tan is obviously a brilliant guy. He flew very close to the sun is what a few people told me. I'm not going to say he did shady things, but he did what he needed to do to help make the iPhone the most successful consumer gadget and its well designed piece of machinery, know, most successful consumer product ever, one could argue. So they knew some of his tactics. He helped build

Speaker 2: So you're saying there was an awareness at Apple that he was basically like always playing to win very aggressively Yes. Mhmm. Like potentially blurring some some lines.

Speaker 5: Yes. Yes. So they knew their players.

Speaker 1: Got it.

Speaker 7: And so

Speaker 5: they knew this guy was going to do anything to get things done, but they weren't able to sue them until they found this kid. Mhmm. One thing I'll say, and it's kind of unfortunate, there's two people at OpenAI on LinkedIn with the same name. There's the guy who's named in the lawsuit, and there's another dude at OpenAI with the same name, and maybe he works for TBPN.

Speaker 7: No, I'm just joking.

Speaker 5: And I see a bunch of people posting on LinkedIn or on Twitter or whatnot a screenshot, and I'm like, You've got the wrong dude. And so you know whenever there's

Speaker 2: a That's interesting.

Speaker 5: Yeah. Whenever there's like a big tragedy or whatever, everyone goes out to try to find the guy's name, and everyone posts the wrong Facebook account. So this guy is getting blamed for getting blamed for it, has nothing to do with him.

Speaker 1: Yes. There there's also an an actor with the name Chang Liu and he has some very iconic photos on the Internet because he's like a basically a male model.

Speaker 5: Oh, that's funny. I think I saw that.

Speaker 1: Yeah. Yeah. Yeah. It's Fang Liu. He looks he looks like yeah. Just like a a model straight out of the runway. What does the timeline for this case look like? I was digging into XAI's lawsuit against OpenAI filed about a year ago. Just was dismissed, I guess, a month ago.

Speaker 2: Is that about App Store ranking?

Speaker 1: No. No. No. So XA SpaceX currently has an ongoing lawsuit with Apple about the App Store ranking. That has not been decided. But separately, XAI sued OpenAI for poaching employees and potentially getting those employees to share secrets of how Grok is trained. That's the allegation. And that case was just dismissed last month on the fifteenth. But it took a year to get there. And I'm wondering because we have that as an example, we also have the Samsung Apple battle that took eight years and ultimately ended in a settlement. And I'm just wondering about how you're feeling about the pace of things. There was sort of a quick back and forth on X, some comments from OpenAI leaders. But how will this story actually evolve and play out?

Speaker 5: It all depends. I mean, OpenAI is going to fight Apple like hell. Yeah. They'll have a law firm involved. Mhmm. In fact, you know, OpenAI hired

Speaker 4: a

Speaker 1: law They want

Speaker 2: to use chat, Kodak's bearish, you know, new models. Yeah.

Speaker 5: They hired they they they actually hired a law firm a couple months ago

Speaker 1: Okay.

Speaker 5: And they were gonna sue Apple over the relationship when it came to the Apple Intelligence Siri deal. So they've got lawyers looking into this and working on this. I wonder so Apple's saying that they reached out to OpenAI in February about their concerns over trade secrets and whatnot. I wonder if OpenAI started preparing its lawsuit against Apple as part of a down the line countersuit type of situation, or as in response to Apple's letter to every neither here nor there. I think this could take several months to years. Really depends. The discovery process is going to take a while. But if I'm going to be really honest with you, I don't think the courtroom timeline really matters. I mean, it matters to some extent, but the bigger picture already happened. If you're an Apple employee, and you're thinking about joining OpenAI right now

Speaker 1: Sure.

Speaker 5: First of you're going keep your mouth shut. You're not going to tell anyone. Yeah. Right? They're going to walk you out.

Speaker 1: Yeah.

Speaker 5: It is not a good thing to be someone on Apple right now who's thinking about joining OpenAI.

Speaker 1: Sure. Sure.

Speaker 5: Sure. You're gonna have Apple security all over you. Your manager's gonna despise you.

Speaker 1: And everything you do. They're gonna be Yeah. Watching every website you went to on your Yeah. Computer for the last month. So you got to know that you're going to be really clean if you're thinking about that. Yeah.

Speaker 5: Yeah. So if you're thinking about going to OpenAI and you're at Apple, the way Apple wrote this lawsuit, you have to be out of your mind to not think twice about it. And I think that Apple needed to stop the bleeding to OpenAI, and this lawsuit is certainly going to help with that. The other thing is, is you're going to have Tang Tan, all those guys caught up in depositions and discovery, and meetings with lawyers now. It's like, how are you supposed to launch a product? How are you supposed to do your job if you're stuck in the office all day with lawyers? Sure, sure. Everything's going to need to be cleaned up. They're going need to work in a clean room like environment now. They're going to investigate what kind of Apple IP is in there. And even if OpenAI did nothing wrong, all of the work that's going to need to be done to prove that, is just going to be a big waste of time for everyone. By just filing the lawsuit, Apple accomplished quite a bit.

Speaker 2: Yeah. So how do you think a company like Apple who notoriously is, you know, very protective about internal secrets, process, supplier, you know, all these different things, is in the middle of of getting, you know, you know, the craziest talent raid in consumer electronics history and then just like has someone go out the door and and they're not even, you know, we were reading Ben Thompson's piece earlier just and he was kind of mocking Apple's security and and as I was reading the complaint, I was thinking back to the deal sort of rippling drama that happened last year and how Yeah. The guy was kind of set up in in some way. And and I was thinking like

Speaker 1: 2.5.

Speaker 2: Is Apple Apple savvy enough to to like try to create a smoking gun by being like, yeah, let's not let's just let it yeah.

Speaker 1: This is full tinfoil. Yeah.

Speaker 5: You think Apple just made this whole thing up?

Speaker 2: No. No. I'm not saying

Speaker 1: that I'm

Speaker 2: not saying I'm not saying that at all. I'm just saying like, it would be crazy to me if you get you're getting people post and you're not like, oh, yeah. We should make sure all the

Speaker 1: Oh, you're leaving the company? Sure.

Speaker 2: Yeah. Just keep your laptop kind of thing. Oh, yeah.

Speaker 1: Yeah. That'd wild wild twist. Well, mean, I I'm sure more information will come out and

Speaker 5: The discovery process is going to be very exciting, very interesting.

Speaker 1: Totally.

Speaker 5: I think Apple has a very high burden of proof here from a legal standpoint. I'm not a lawyer, but being able to prove that its IP is within certain products is going to be it's to be difficult to prove. Yeah. So we'll see if we're able

Speaker 1: to do it. Also the narrowing of, like, what is a trade secret because, like, it's been this, like, awkward reality in the AI race that when a AI researcher leaves for another lab, they basically take the architecture that isn't necessarily patented in this sort of a trade secret. Like, they take that with them and they just say, oh, yeah. Like, we should probably train a model of this size or we should probably include some of this training data. And that's why we've seen the frontier. But that's very different from hardware stuff with patents and

Speaker 6: Right.

Speaker 5: I would like to know what trade secrets or what Apple will be able to claim as a trade secret

Speaker 1: Yep.

Speaker 5: Ends up in these products. Because as you know, and everyone listening to this knows, hardware is incredibly commoditized. There are hundreds of companies that make electronics in the same space as Apple all over the world. And being able to prove that you're using Apple technology to release your phone, and build your phone, and develop your phone, that's going to be very difficult. I don't see this being a clean victory for Apple if it goes to a jury trial. They don't My guess is they're not going to have Remember the Samsung trial? This is what Samsung phones looked like before the iPhone. This is what the iPhone looked like. Yep. This is what Samsung phones looked like, right? They broke it down so a third grader would understand it. It was very easy to follow. Short of that, I just don't see how Apple could win a jury case here. And if you read the complaint, over and over again, acknowledges that they don't really know what is really happening here. They know enough to know there's something shady going on, and we need discovery to figure out what is going on. So they'll do that, but I'd be very surprised if they find something. The other part of this is there's a few things that you can really make a trade secret case on right now. That's silicon, And Apple has very proprietary silicon code, the way they develop chips and design chips, and they went after this company called Rebos that Meta ended up buying, and they got major changes made to how Rebos operates. The other one you mentioned is AI. You know, there's no AI thing that OpenAI is stealing from Apple. And from the chip side, everything we know is that OpenAI is using Qualcomm and third party chip providers. So I'm very curious to see what Apple ends up finding. Interesting. I'm sure they'll find some good text messages and emails, though.

Speaker 1: Yeah. Can you take us back in time to, Johnny Ive's departure, the birth of IO, how that was perceived within Apple? Because it feels like when the acquisition happened, was 55 employees at IO at the time. A lot of them ex Apple. And Ben Thompson pointed out that he said, I think Apple feels betrayed not just by Tan, but also Johnny Ive. Start with the fact that Ive framed his partnership with OpenAI as an attempt to undo the harms he thought were caused by the smartphone, not exactly a ringing endorsement. So you see these, you know, people leave and then they're like, oh, like, you know, we need we need an antidote to the social media thing. This is a common a common thread that that people pull on in their second act of their career. But what was the IO journey? Was it pretty smooth when that happened? Was this something like, if if you play out a different scenario and IO is just consulting for, you know, a Samsung and Ferrari. Know, Ferrari and a bunch of other, you know, companies that are not, like, directly at each other's throats, new hyperscalers, a disruptive AI take off, all of that. How what how does the story look then?

Speaker 5: So the story well, let's I'm not going to get into the Johnny I stuff right now, but

Speaker 2: Sure.

Speaker 5: You know, Tang Tan, he tells John Turnis. By the way, the other side of this thing is him and Turnis hate each other. Mhmm. Happy to go into that. Mhmm. But Turnis, you know, he was up for the SVP of hardware job in 2020. Tang Tan and a lot of the people at Apple who now are at IO and at OpenAI, they wanted Tang to be the new head of hardware engineering because they enjoyed working with him on the product design side of things, and the ID studio and product design worked very closely together. Tang had run the iPhone side of things. Ternis for a number of years was running the iPad and the Mac side of things. He had felt that a lot of the core technologies that Tang was working on, you know, they got iPhone priority, and he didn't really want to share that with the iPad and the Mac team. You know, now Ternist and Cook are sort of getting the last laugh, sort of, you know, giving Tang that reputation that he's not going to ever be able to draw of being someone who's sort of, they didn't say the words, but a ringleader around this. So what happened. He tells Apple at the end of twenty three he's leaving. Ternus lets him stick around through February '24 because Tang's role at Apple was so big. He had his tentacles in so many different things. They actually had to replace him by splitting up his responsibilities, if I remember correctly, across five or six or seven different people. Wow. So it took a lot of time to unwind. He behind the scenes was already working with Johnny Ive and Sam Altman and some of these other folks on creating this company. I think the origin story of IO was they knew all along that they were going to get acquired by OpenAI. The other part, and this sort of got buried when OpenAI acquired IO, But like OpenAI already owned 25% of IO, if I remember correctly, for a year or maybe more before the acquisition of OpenAI buying IO was announced. So that's pretty interesting. That's very interesting. Neither here nor there, IO, they're building some of the best talent who's ever worked at Apple. Combining the best talent that's ever worked at Apple with the AI acumen of OpenAI as a scary sight, I would say.

Speaker 1: Mhmm. Oh, no. They're gonna have to pivot to a VR headset. My dream. I just want another VR headset. That's all I want. Yeah. Me and me only. I'm the only one that's gonna get it. It's it's rough. What else is what else is new in in Apple world? What else are you covering outside of this?

Speaker 5: Lots of stuff lots of stuff to come. Yeah. Apple's gonna start its biggest new product period in its history

Speaker 1: Yeah.

Speaker 5: This fall.

Speaker 1: And and at higher prices. Right? Like, the the the memory shortage is very real. It was sort of like a delayed announcement post WWDC. But, you know, the the the the strategy there was like rip the band aid off after WWDC, but with enough time for everyone to absorb the news before the products are launched. Is that generally the theory there?

Speaker 5: Well, you know, they hit the point where they couldn't impact their margins anymore. Obviously, they're reporting earnings at the end of this month, so they'll talk more about that. Yeah. But the big picture here is the iPhone is going to get its price hike in September. Yeah. Everything I'm told is that the cost to build because of the memory makes these phones $150 to $200 more expensive a unit.

Speaker 2: Yeah.

Speaker 5: So I think they're going to raise the prices by 150 to $200. Yeah. They don't want to get up that margin. I mean, fact that nobody's buying the Vision Pro and Apple hyped the price up by $200 anyways because of the memory tells you everything you need to know. This is not a company that's going to compromise on its margins.

Speaker 1: Sure. Is there any world where vertical integration around memory is on the table? Like, is going deeper there locked? Even like the Southwest bought all those, like, futures on oil and had cheap gas and cheap flights for a long time. Like, you know, even if you're not AGI filled and you don't build the foundation model, there's still a world where you go lock up really long memory contracts and keep your prices down.

Speaker 5: I don't see it happening. Mhmm.

Speaker 1: Just not culturally the value?

Speaker 5: I don't yeah. I don't see it happening. How is I think they're trying to get the suppliers to expand.

Speaker 2: Sure. Yeah. How has Apple's relationship with TSMC evolved or changed across the AI boom? Mhmm.

Speaker 5: Oh. Well, I think the long and the short of it is that, you know, Apple is still the preferred customer of TSMC

Speaker 7: Mhmm.

Speaker 5: But there are plenty of other customers now that are a big piece of the pie. And so I think the relationship still remains very strong. But I don't think this is exclusively because of the AI boom. And you can see Apple looking at secondary potential third suppliers on silicon. So you've seen that with Intel. They've talked to Samsung as well. But the TSMC relationship, from everything I understand, is still as strong as ever. It's just Apple's not the only game in town as a customer at this point.

Speaker 4: Mhmm.

Speaker 1: Well, thank you so much for taking the time to come break it down for us. We appreciate you

Speaker 5: You got it.

Speaker 2: Hopping on. Thank you. Congrats again on on a major scoop on a major scoop. How much how much time did did you were you able to were you able to, like, prewrite your article before the actual complaint was live? Or did you did it go live and then you wrote it?

Speaker 5: Did I write my article before the complaint was live?

Speaker 2: Like, did you know it was

Speaker 5: a Good bad question. I'll I'll tell you next time.

Speaker 1: Thank you so much See for coming on the Mark. See guys.

Speaker 5: See you soon.

Speaker 1: Let me tell you about Public. Whether you're long short long Apple, short Apple, Public has investing for those who take it seriously. Stocks, options, bonds, crypto, treasuries, and more with great customer service. Let me also tell you about Shopify. Shopify is the commerce platform that grows with your with your business and lets you sell in seconds online, in store, on mobile, on social, on marketplaces, and now with AI agents. Yes. Yeah.

Speaker 2: I can see why he wouldn't actually wanna answer that question

Speaker 1: Of course not.

Speaker 2: After I set

Speaker 1: it up. He knows everything. Nothing happens in Cupertino without Mark Gurman knowing first. There's an interesting AI threshold that was broken over the weekend. I didn't know that I wanted goalpost in this particular area, but I like this. So Ethan Moloch shared this. He says, this was one of those impressive AI thresholds for me. I gave GPT 5.6 Soul in Codex control over my computer and asked it to win the daily challenge for the game Slay the Spire two randomized factors so it can't cheat. You can't just look up the answer. And he said it worked for five hours making complex game choices and won. And I've been a huge fan of Slay the Spire. I really enjoy this game. I don't know that I didn't know that there was a sequel. I guess there's a sequel out already. But it is sort of a it's a slow paced game. It's turn based. So you can take as long as you want to make every decision. And so it's not real time, so it's somewhat suitable for this. But it makes me very, very clear that these are great benchmarks for computer use because you can RL on certain Pokemon benches that the data will flow in once it becomes a big trend. But if you can just fire up some random game that you find personally interesting and watch a coding agent sort of win that game just through computer use, that's gonna be a very visceral feel the AGI moment for a lot of people. What do

Speaker 3: you think? Mean, yeah. I think the current coding models are are not good at this because they're like a little bit too slow. Yeah. But if you remember, standard intelligence had that computer agent model where it was like playing games in the browser. Yeah. So I that's like a very different architecture from like what, you know, Codex is using right now

Speaker 1: or whatever, Cloud Code. But Yeah. I think it

Speaker 3: should be very possible.

Speaker 1: So, yeah. Mean, there's been a lot of advancements in AI that can win games, but it needs a very specific harness. Like, it always needs to have some sort of like the output is basically JSON or something and that's translated into actions in the game. What's so cool about this demo is that there is no harness. It's just moving the mouse and clicking the buttons and it learns all of that on the fly, I think is really cool, which means that we need to move the goalpost one more time because we are coining and we are creating Rustbench, which is not the programming language Rust. That's right. It is the level in Call of Modern Warfare two. Right? You need to be able to beat Jordy on Rust if you're building a coding agent. Wait. That's oh, it's

Speaker 3: Is that COD four?

Speaker 1: Oh, is it called

Speaker 2: m w two. Wait. Which one is it? We play Modern Warfarers.

Speaker 1: We play COD and we play Rust. Rust. And you and

Speaker 2: And the fact that you guys can't even get it straight

Speaker 1: We are going to set up Taking it seriously. Codex on Rust and we are going to one v one it and hopefully and hopefully defeat it for a very long time because humans must stay in control. But I have a feeling that the AI will win eventually. But it might take the Cerebrus chips. It's going to have to think pretty quickly, think on its feet because five hours to win Slay the Spire, I think it it takes a normal person like maybe thirty minutes, forty five minutes to win one of those challenges. We're getting called not real gamers. It's really rough right now. We're getting absolutely destroyed.

Speaker 2: Damn. Young John would be just so disappointed. It's been

Speaker 1: a while. Let me tell you about Codex. Codex is a powerful workspace for getting work done with AI agents and playing video games now. Whether you're writing code, analyzing data, creating content, or automating business workflows, Codex helps you move projects forward from start to finish. And we have our next guest with us in the team.