(Preview) Anduril and the Lattice Partner Program, A New Direction for Defense Tech, OpenAI Introduces ChatGPT Pro

Unpacking a quietly significant announcement from Anduril this week, the future of hardware and software in the U. S. defense industry, and checking in with OpenAI as the company announces a new subscription tier for ChatGPT.

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Andrew Sharp?

  Hello and welcome to a free preview of Sharp Tech.

• 0:08Ben Thompson?

  There's network of all this stuff on the edge.

• 0:10Andrew Sharp?

  So let me ask you, in terms of the implications, like the Lattice Partner Program and Lattice SDK, they had the Reagan Defense Forum out in LA last weekend, and there's obviously sort of a bustling ecosystem of defense tech startups out in LA. Is this the sort of thing that all of those startups would then be building on in the years to come as we build out our defense capabilities?

• 0:34Ben Thompson?

  Well, it's pretty interesting because there's going to be, you know, kind of low-key an ecosystem fight about this because Anduril is making a bid to be the dom, like the windows of like

• 0:46Andrew Sharp?

  Right, the middle layer that everyone uses, yeah.

• 0:48Ben Thompson?

  That's right, which where the most sort of profits do accrue. And so it's interesting because on one hand, it makes a lot of sense. It's also going to be like, you know, I think the timing is certainly good. Obviously Anduril is well-regarded with the incoming administration for lots of reasons, including, you know, Palmer Lucky famously being a Trump supporter previously.

• 1:10Andrew Sharp?

  Back before it was cool, yeah.

• 1:12Ben Thompson?

  For very, for very, still somewhat unclear reasons, losing his job at Facebook over it, which by the way, how insane is that? That you would lose your job for supporting a presidential candidate? Or that we had people at Google protesting against them joining Project Maven? Like the, well, Project Maven is actually an interesting angle. This is the high-level Defense Department idea of, look, we have to tie all this stuff together, right? Like, we have all, the Defense Department is a country of silos, like just on a weapon system level, on a service level, on all sorts of things. It's predicated on this hierarchical model where everyone's trying to protect their turf and sort of like move up on this, move up in the stack and all these sorts of things.

• 1:56Andrew Sharp?

  Right.

• 1:56Ben Thompson?

  So Project Maven was this idea that we need a common sort of data layer that sort of ties all this stuff together. Google won the bid, had all these protests, backed out, thought you were an American company, guess not. And so Palantir ends up winning that. A big question I had was, is there?

• 2:15Ben Thompson?

  A looming conflict, uh, conflict, maybe bad choice of words in this context. Is there a looming sort of, uh, ba- ba- another bad choice of words. But between Palantir and Androl, in that both are sort of making bids to be an operating system type layer, the common software layer that sort of manages all this sort of stuff. And it is interesting that, that Androl and Palantir did actually, uh, I believe it was at this forum, announce sort of a partnership, uh, you know, it was a consortium to accelerate AI capabilities for US national security. Um, they're also talking about, you know, partnering to deploy advanced AI systems. It's unclear what the specifics are. Actually, I do want to sort of dig into that a bit more, so, so stay tuned on that.

• 2:59Ben Thompson?

  But I think it's a, it's a good sign for both and it's a good sign for the US that this makes sense. Androl's focused on the edge, which is sort of like all the weapon systems that are out in the field. Palantir's focused on the center, like, like the commander who is trying to, to man- and it makes sense if you have lots of stuff plugged into Lattice and Lattice works seamlessly with sort of Palantir, that that's sort of a, a, a vertical stack that, that at least theoretically sort of makes sense. But it is a fundamental, what we're seeing is a, a fundamental shift in how you think about sort of weapon systems, that there need to be a common platform, there, it needs to be software at the center.

• 3:43Ben Thompson?

  The integration that matters is across weapon sys- and not just weapon system, also sensors and data processing and all these, these sorts of things. And it is a fundamental shift and it is notable.

• 3:55Andrew Sharp?

  Certainly sounds like a healthy shift because when you were explaining it earlier, I mean, it just seems insane to have our defense capacity full of these bespoke weapon systems that all run on unique software and are really, really expensive.

• 4:11Ben Thompson?

  Insanely expensive and all require like usually a person in it.

• 4:15Andrew Sharp?

  Yeah, exactly.

• 4:16Ben Thompson?

  Which, uh, from, from a just sort of a, a warfighting capability is very expensive in, you know, ways beyond money.

• 4:22Andrew Sharp?

  Mm-hmm.

• 4:23Ben Thompson?

  And so, uh,

• 4:24Andrew Sharp?

  And doesn't really scale with the modern environment and the way other countries are approaching these problems.

• 4:30Ben Thompson?

  Maybe. I mean, I don't know that other countries are this advanced on this particular point either, to be totally honest. What we do see is the explosion in drone warfare, in particular.

• 4:40Andrew Sharp?

  That's what I mean, is people are throwing out like really cheap drones that can then take down a $500 million X, Y, or Z, and the math doesn't really check out over time.

• 4:51Ben Thompson?

  There's a real math problem facing the US military in general. That is correct. And so, when you back up, and the thing that's interesting about this is Palantir and Andro, which I think are sort of on the vanguard of this, the reason they are on the vanguard and could do this is because they did it outside of the DoD. They did it with venture capital. They did, like, they built it themselves. They're like, this is the way it should be done. We can't wait for the US to tell us to do this. We're going to build it. And then they will hopefully wake up at some point and realize this is sort of the way, the way it should be done. And it is reminiscent of sort of Bob Noe saying, I'm not going

• 5:36Ben Thompson?

  to take defense contracts because it's too limiting. We're going to build chips. And that is the way to make chips faster and more capable. And then, it worked out then. It was also easier then because there was a consumer market to be built for.

• 5:50Andrew Sharp?

  I was gonna say, there's not really a consumer market for Andro, as they, like, take this leap.

• 5:55Ben Thompson?

  Yeah, for Andro in particular, it's tough because it's not clear what, like, I mean, I guess theoretically you could see the consumer application of like drone swarms and things like that. But Palantir obviously has a very thriving, yeah, Palantir obviously has a very thriving commercial business, an exceptionally thriving stock, that's for sure. But, but I do think they're, you know, their overall approach is, it makes a lot of sense. And I don't, I don't know, like, I just think it's, it's, it's cool, it's exciting, this, these two pieces that they're taking the next step. An operating system is, is nice when it's all your stuff. It's a different thing when you're creating a platform for other companies. We talk about, you know, it's, it's time to build. Like,

• 6:40Ben Thompson?

  what, what is there to build? Well, what about building weapon systems? And now suddenly you have the hardware, or the, sorry, the software capability to know exactly what to build for. Again, this is sort of setting up Andro very, very well to have like real, meaningful, long-term lock-in for, you know, for, for sort of better or worse. But, you know, I, it's obviously pushing us in the right direction that we need to go from a national security perspective.

• 7:08Andrew Sharp?

  Yeah, I mean, it seems to be superior to what we were working with in the past. How much resistance was there within the defense industry to something like this? I mean, like Raytheon, for instance, or Northrop Grumman, like, has that been part of the Andro story is running into?

• 7:24Ben Thompson?

  Yeah, I mean, Andro, like, they, so Gregory Allen noticed this, noted this on a, an interview we did a couple of months ago, that there's a new, there was a new defense contract that Andro just won, which was sort of meaningful, that they're, they're sort of very much, you know, that it was the first meaningful kind of, and we're still talking on the order of a few billion dollars, it's not like massive, but, you know, it's been, I think, a long, a long haul to get there, and there's still a long ways to go. There's a bit of a revolving door between the Defense Department and these sort of defense contractors, it's sort of its own Galapagos Islands sort of bit, where they're off doing their own thing, in part because that's what they were told to do in the late 90s, like consolidate or die, and we ended up with all these monopolies that, that, that are, that are incompetent.

• 8:09Andrew Sharp?

  Slowly atrophying, yeah, exactly.

• 8:11Ben Thompson?

  Yeah, and I mean, we see what's happened with Boeing and, and, and things along those lines. So I, I think the, the answer frankly is probably going to be, you know, they're gonna have to be dragged kicking and screaming sort of into this. One thing that I can't, I, I did interview the, the Palantir CTO, I can't remember if this came up in the interview, but it was company conversation, where one thing they've, they've realized is Palantir really got its start with the military in Iraq, not like in the later years, when they were, you know, trying to clean, clean up the mess, and was very effective in sort of on the ground, sort of managing data, figuring out where things were. And what sort of happened

• 8:54Ben Thompson?

  for them is that was a junior level of officer that was sort of on the ground, and was finding real utility from this software that they were using. And as that sort of level of officer has moved up, and has become more and more senior, suddenly their advocates at sort of the high levels have, have increased.

• 9:13Andrew Sharp?

  That makes sense.

• 9:14Ben Thompson?

  You know, this is, it's funny. This is like a career thing, right? You, you realize this. Everyone, you know, young people are like, oh, I need a good network. I need a good network. Um, you know, I have one of the most incredible networks in the world. Not because they went to network events, but because, you know, did good work. And that as you, you like, you have two parts. Number one, people want to know you. So you reach out, but number two, there's a lot of people I knew as a young person that suddenly are like.

• 9:38Andrew Sharp?

  Powerful now.

• 9:39Ben Thompson?

  Yeah, it's like, hey.

• 9:41Andrew Sharp?

  It takes time.

• 9:42Ben Thompson?

  I don't even mean this person. It does take time. And I think there's, there's a bit about, you know, there's certainly a generational shift that is sort of necessary. But I mean, I just, I appreciate these companies doing this. I appreciate Andro starting in the late, you know, the, the late 2010s. When you were vilified for doing a defense contractor. And the reality is we have this capability now and we have this SDK announcement that is downstream of doing the work for years to sort of get to this position. And it's, it's certainly a direction that we need to go.

• 10:14Andrew Sharp?

  All right. And that is the end of the free preview. If you'd like to hear more from Ben and I, there are links to subscribe in the show notes, or you can also go to sharp tech.fm. Either option will get you access to a personalized feed that has all the shows we do every week. Plus lots more great content from Stratechery and the Stratechery Plus bundle. Check it out. And if you've got feedback, please email us at email at sharp tech.fm.