The ‘AI-First Company,’ Big Tech Mega Earnings, NVIDIA vs. Anthropic

Ranjan Roy from Margins is back for our weekly discussion of the latest tech news. We cover 1) What an 'AI-First Company' is 2) Aaron Levie's AI-first memo to Box. 3) Luis von Ahn's AI-first memo to Duolingo 4) Tobi Lutke's AI-first memo to Shopify 5) Can non-tech company be AI-first? 6) Apple earnings 7) Apple's federal government issues 8) Microsoft earnings 9) Amazon earnings 10) Meta Earnings 11) NVIDIA vs. Anthropic 12) More on AI consciousness 13) The very handy Mag2 --- Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice. Want a discount for Big Technology on Substack? Here’s 25% off for the first year: https://www.bigtechnology.com/subscribe?coupon=0843016b Questions? Feedback? Write to: [email protected] Learn more about your ad choices. Visit megaphone.fm/adchoices

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

• 0:00Alex Kantrowitz?

  CEOs everywhere are calling their companies AI first. But what does that mean? Plus, we cover a major week of big tech earnings and Nvidia and Anthropic exchanging harsh words. That's coming up on a big technology podcast Friday edition right after this. This episode is brought to you by Genesis. What does it actually take to put a Gentic AI to work across an entire enterprise? I recently spoke with Genesis chairman and CEO Tony Bates about why orchestration is emerging as a competitive advantage in customer experience and what companies need to get right in the AI era. Then I sat down with Adam Mitchell, head of enterprise business solutions at Voya Financial to go inside a large scale agentic CX transformation, including the business

• 0:45Alex Kantrowitz?

  decisions, trade offs and lessons along the way. You can watch both conversations now on my YouTube channel today. It's never been more challenging to get frontier compute. How do you know where to look, who to trust, or how much to pay? Akash is an open cloud that provides easy access, flexibility and peace of mind. That's how companies like Razor were able to cut their inference cost by 15 times. Visit akash.network slash Alex for free consultation. That's akash.network slash Alex for free consultation. Welcome to Big Technology Podcast Friday Edition where we break down the week's news in our traditional cool headed and nuanced format.

• 1:27Alex Kantrowitz?

  Major week of news to speak with you about this week as every CEO seemingly has this new memo out calling their company AI first. So we're going to talk about what an AI first company means and why the companies that we're going to discuss, including Duolingo Box and Shopify have decided that it's important and whether that can expand beyond tech. We're also going to cover a major week of big tech earnings. We have Apple, Amazon, Microsoft and Meta all reporting. We're not going to go so deep into the numbers themselves, although we'll cite plenty of numbers, but really what we're going to look at is the trends, what it means for big tech and the economy in general. And then towards the end, we'll talk about Nvidia and Anthropic getting into it. A great letter that we got from one of our listeners about AI consciousness.

• 2:12Alex Kantrowitz?

  And then we will close out with a new gadget that Ron John has and loves, and he can't wait to tell you about it. Joining us as always on Fridays, I already gave it away, is Ron John Roy of Margins. Ron John, great to see you. Welcome back to the show.

• 2:24Speaker 01

  This is an AI-first podcast, Alex. How have you, uh, how have you used AI as a fundamental expectation of podcasting, as Toby from Shopify said?

• 2:34Alex Kantrowitz?

  It is interesting that we're going to talk about how all these companies are AI-first because in order to insert their memos into our show notes or our show document, I had to use AI. Every CEO seems like they want to post their memo as a screenshot. I mean, there are longer tweets now. You can just give us the text so we can read it as opposed to having to zoom in on your screenshot. Best possible way to make that happen, save those screenshots, drop them into Claude and have Claude spell it out.

• 3:02Speaker 01

  That is exactly what I did. So I did my part as an AI-first podcast, Alex.

• 3:07Alex Kantrowitz?

  This is a big trend. Three is a trend and we have three now where there have been three CEOs who have written memos to their companies talking about how they are now AI-first. That includes Shopify, which was the first, Duolingo and Box, which just showed up this week. So Ron, John, I'm gonna, we're gonna go through a little bit of these memos because I think this is important and I think we're going to see many more companies gonna, that will go this route. But why don't you start us off by giving your thought on what it means to be an AI-first company because it does sound like jargon. So look at what the three are saying and tell us what you think it means.

• 3:47Speaker 01

  So reading these three memos again from Duolingo, from Box and from Shopify, each one had a slightly different tone or intensity to how they said their employees need to be AI-first. But overall, the idea is you always need to start with an AI tool. Can the task you're trying to do be done in a better way with some kind of AI tool? And I think it's right. I think getting just everyday people in your company really fluent in whatever new tool is happening and figuring out how they can actually change the way they do work is actually far more important than any large scale effort that any company is going to try to implement. So I actually, I like

• 4:30Speaker 01

  That this message is going out and we can get into, you know, the tone and the nuances of how they're actually trying to communicate this, but I like that they're saying this because I do think these are the companies that will do better that are, I mean, it's just the same way. And even Duolingo, they talked about the shift to mobile. If they had missed that as a company, they would have been dead. I mean, Facebook slash Meta was one of the most famous transitions to the mobile era from desktop. So when these big transitions happen, you need everyone at the company, not just exploring the tech, but actually integrating the kind of AI or whatever that big transition is into their, their daily work.

• 5:11Alex Kantrowitz?

  So there are definitely a few themes that I saw in these memos. First is a desire to share information within companies. I would say the second is probably a desire to cut costs or at least minimize headcount increase. And then third is to scale up what you do in ways that you couldn't before AI. So Aaron Levie, in his memo to Box, he told his employees that we're focused on building an AI-first company. Just as we ask ourselves what our product would look like if we started Box in 2025, we need to ask ourselves what would work look like if we started in 2025. So it's basically just his way of saying, we need to collaborate better. Now, is it a coincidence that he's driving this, writing

• 5:54Alex Kantrowitz?

  this memo and releasing it in public and he has a company that is explicitly built on document storage and collaboration and a tool that can make you more efficient at work? Probably not a coincidence. He's probably going to talk a little bit about what Box can do and then release this to the world and, you know, try to get others to follow. But, but some of these things that he brings up are quite interesting. So to me, he starts really in the knowledge sharing area. He says, we want to use AI to ensure more time on average is spent on the things that really matter. So we'll use AI to help onboard, onboard Boxers faster, get everyone access to experts on any topic like Box AI in hubs, which I think is like an internal help desk, to make decisions more quickly, iterate on new ideas

• 6:39Alex Kantrowitz?

  more quickly, augment our code writing safely and better serve our customers and more. So Roger, let me ask you this, because this has been one of like the big questions about AI right away, is can AI help companies be more efficient? Sort through internal documentation better, get access to better expertise.

• 7:00Alex Kantrowitz?

  And basically spend time on things that matter more as opposed to like going through these like long, arduous onboarding processes and trying to cut through a tape to find out information. Do you think AI is at the point today where this can actually help with this task?

• 7:15Speaker 01

  Yeah, 1,000% and to Aaron's credit, I think Box is certainly one of the players in that space. And which also, I will say, the fact that Duolingo came out with it, I thought was pretty interesting because they have no real vested interest in kind of pushing that narrative. And in fact, they got a lot of backlash because thanks to Duo and their social media marketing, they've become a bit of a cult favorite, but also with the the anti-AI persona. So so putting it out there definitely meant something different. But I think the most important thing there and like what you were saying, or what Aaron said, starting a company in 2025, if we did it today, what would it look like? That is the most important question that every company has to ask themselves.

• 8:00Speaker 01

  And I've thought about it a lot because the last few years, it's been a more, most organizations trying some experiments, testing some out, some stuff out, seeing what works. There's a lot of internal inertia towards most AI efforts as well. But up and coming startups that are built completely from the ground up on these kind of technologies are going to just do stuff way faster, like much more efficiently, probably better for the end consumer as well. So any kind of incumbent has to rethink the way they do it. And in reality, the technology to actually like access large amounts of information,

• 8:43Speaker 01

  especially unstructured, and get an answer that is correct. They're avoiding hallucinations, whether that's through rag type technologies, whether that's other, like it's gotten so like really long context windows and Gemini and OpenAI. Like it's gotten so much better that this will become completely table stakes for every organization. So the ones who do it faster will do better.

• 9:07Alex Kantrowitz?

  By the way, I think this is the reason why it's important for us to read these memos and why we're leading here because we know that this is going to be the way that everybody in the economy ends up operating, if not everyone, a large majority. And it starts to answer some of the big questions that you might ask in terms of what the impact is going to be both on business and society. And I think there was a theme throughout, but I'm just going to start with Box here, that the idea is, and I know I cited efficiency as one of the things that they're aiming for, and they definitely are, but like, what do you do with efficiency? And you have to remember that every company that's beginning in 2025 is going to start

• 9:48Alex Kantrowitz?

  with many of these tools natively in their approach. Like, if you need a internal knowledge hub, which I hate to use the term of internal knowledge hub, because

• 9:58Speaker 01

  It's a thing.

• 9:59Alex Kantrowitz?

  Yeah, it is. But if you need one of those, just, I don't know, it just feels like very like, put me in my cubicle and make me read the internal knowledge hub. But when you have, and everybody has internal knowledge, shared internal knowledge in a company, the newest companies will build with that centralized in AI, queryable by large, by natural language. And so the question is, modern companies, what are they going to do? Do they, do they become efficient and kill everything? Or do they become efficient and do more? And I think this is kind of the most underrated part of the Aaron Levie memo, and then we'll move to Duolingo. He says, we'll bias towards a keep what you kill mentality with AI, which means that when an organization frees up resources or saves money with AI, we'll aim to reinvest

• 10:44Alex Kantrowitz?

  those things back into more strategic thing. Again, we want to use AI to get more great work done, not the same amount as before. This, to me, is the thing. Companies using AI are not going to say, we are happy doing what we did before, and this will just enable us to do it more efficiently. Therefore, we will cut our spending and we'll cut people. I think the smart companies, companies like Box, run by a person that really gets this, will understand that this is a mechanism to push even harder and grow and take that money in and reinvest. And it doesn't mean job loss, and it doesn't mean cost cutting. What do you think?

• 11:17Speaker 01

  Well, yeah, I think there's always two ways to look at the potential benefits of AI. There's cost savings, but then there's also doing things that you could never do before. And in the Duolingo memo, Luis Manon, he talks about doing implausible tasks. And like, again, you imagine as an online shopping, we talked about this a lot last week. There's a world where you can truly personalize every bit of messaging that is now being good, that going out to any kind of customer. And when you say internal knowledge hub, it makes us think again, you have to like, check with HR what the vacation policy is. I think, think about it more like, internal knowledge is even all the actually like aggregated

• 11:58Speaker 01

  product, knowledge, product data, customer data, all of that kind of knowledge and being able to quickly leverage it and generate new types of content in new ways. And that you can serve customers and just make more interesting things than you could ever before. And again, to me, it's interesting because we, you know, AI, a lot of the industry is trying to push the narrative that it's potentially, you know, the most transformative technology since fire or whatever else. I still kind of am in the camp that it's, it's like cloud or it's like mobile. And you just have to think like, if you were at a company that, or even computers, call it 30 years ago, if you're at a company that's an incumbent,

• 12:42Speaker 01

  that is not investing and not becoming fluent and native in that new technology, you're going to lose. So to me, it's not, you know, this earth shattering, once in a millennium thing, but it's a pretty important thing that everyone's going to have to get their hands around.

• 12:56Alex Kantrowitz?

  Right. And just to the subject of building more, I know we're going to get to Meta earnings and what Mark Zuckerberg said about AI. He made, uh, Mark Zuckerberg made the podcast rounds this week and we'll, we'll talk about that. Or maybe we should just talk about it now. I thought he had a fairly miserable showing on Theo Von's podcast, which was very, very difficult to watch and just not the right host for the, for the guest. I mean, Mark wants to talk about technology and Theo wants to talk about basically everything that makes Mark Zuckerberg squirm. In fact, there was a great moment where Theo Von looks at him and he goes, uh, do you think, and he goes, is social media bad? And Mark is just like, I don't think so, uh, which was hilarious. Uh, but otherwise that,

• 13:39Alex Kantrowitz?

  that interview really wasn't worth listening to. The one that I thought had some good moments was Zuckerberg and Dwarish Patel. Dwarish, of course, has been on the show. And they start talking about AI, what AI will enable Meta to do. And Zuckerberg says, look, we're too big to have a customer service organization. But could we have AI-enabled customer service that has, you know, voice potentially? And then, you know, enables us to build something that can serve 3 billion people and have limited need for like human oversight. In which case, we're going to hire more customer service people because they'll be able to maximize what they do with AI. And I think that was what you saw with Duolingo.

• 14:21Alex Kantrowitz?

  So just as Luis Vanan, the CEO of Duolingo, releases this memo about going AI-first, Duolingo doubled its courses. It said it's launching 148 new language courses and 28 supported user interface languages. And this is what Luis said, or no, the senior, the senior director of learning design, Jesse Becker, said this. By using generative AI to create and validate content, we're able to focus our expertise where it's most impactful, ensuring every course meets Duolingo's rigorous core standards. So it's pretty amazing that AI can enable a company like that to provide what it's doing just at a much greater scale,

• 15:04Alex Kantrowitz?

  which creates a much larger moat over any language company that's not AI-first. And to me, I think that's really the promise.

• 15:12Speaker 01

  Well, yeah. And that's again, a good example of increased efficiency. It took 12 years to create 100 courses. And now they just created another 150 that they're confident are at the same quality and scale as the original ones. But then, I mean, language, and I'm a, I'm proudly on day 334 of my Duolingo streak in French. I don't know if my French has actually gotten that much better, but I'm obsessed with the Duolingo widget and making sure I don't lose my streak. But How much more weirdly and almost amazingly personalized, like now they actually have these video calls you do with an animated character and talk to them. And, you know, like, there's a lot of talk after OpenAI release voice mode that Duolingo is dead. It's even more reason

• 15:57Speaker 01

  that a specialized language learning company like Duolingo, if they leverage this technology well, they're going to create some amazing stuff that will, we can interact with and learn languages with in ways that, I mean, when I was sitting in a sixth grade French and with a textbook trying to conjugate verbs, could only dream of starting to learn a language today. So that's the stuff I think that's like, again, efficiency over here, but then creating new amazing experiences is that's the next phase.

• 16:30Alex Kantrowitz?

  Now, I'm not opposed to those video calls, but I just think that they're awful. You like

• 16:35Speaker 01

  them? It's no, it's good. You can act, you can't, yeah, I try to mess with them sometimes. And it actually like, they reel you back in to ask you questions about the vocabulary you've been learning. And again, it's like, I don't know, weirdly, because it's French, I think it's like a animated, really like a blase, girl with purple hair. I don't know. It's weird, but

• 16:59Alex Kantrowitz?

  That's, I think that's the character for all of them. They're not impressed with your, with your language skills, the Duolingo characters.

• 17:04Speaker 01

  See, but wait, wait, but in a way, when we're talking about is OpenAI gonna kill everybody? Or is it the product and not the model? Such a good example. They have a whole brand. They have a whole animation style with which they deliver the courses. They have like an entire vibe to them. And, and that's, it's a reminder. Everyone, everyone can still win right now.

• 17:27Alex Kantrowitz?

  I hate while I'm, that I'm grudgingly coming towards your opinion, but that is great support for your, the product matters. But then again, better models enable this type of experience. So you and I will let that debate run another day. You know, I was so optimistic leading into this segment, talking about how a company is just going to double down and grow. But then I look at some of the things that Did cause the backlash for Duolingo and that of course is the human cost here. So Van An in his memo said, we'll gradually stop using contractors to do the work that AI can handle. AI will be part of what we look for in hiring. AI will AI use will be part of what we evaluate in performance reviews. And this is really important and this was also mirrored in the Shopify memo.

• 18:11Alex Kantrowitz?

  Headcount will only be given if a team cannot automate more of their work. So when you look at the sort of calling of contractors and then the fact that you're going to have to justify a headcount. It'll be much more more difficult to get headcount because the AI should do this stuff. And then I wonder like, okay, so if this does get applied across the entire economy. Uh, are we going to see any growth in jobs? Like, could this really lead to a lot of job loss? I just saw a tweet this that come across my feed this week, talking about how new graduates are really struggling to find work right now. Now, maybe it's because of the uncertainty in the economy, but I don't think it's out of school

• 18:55Alex Kantrowitz?

  or or even a remarkable insight to say that a lot of companies now are not hiring their most junior level because they're turning some of this over to AI. What do you think, Rajan?

• 19:07Speaker 01

  I mean, it's tough, but to me, it's an inevitability. Like if you are able to build and run a company a certain way and you don't, someone else will. I mean, I think it's just that simple that the way work is done is changing and there are certain tasks. And I mean, Duolingo is the perfect example. They had contractors simply checking the output. Now with generative AI, you can safely, if you build the right system, which I'm sure they did, check. Thousands, hundreds of thousands of answers very, very quickly and efficiently. You're not, just because you, now that you can do that, you're not going to do the opposite.

• 19:51Speaker 01

  And if you don't do that, someone else will. So I think to me, and this is going to be an ongoing discussion, don't get me wrong. This is, this is going to cause dislocation the same way like in the nineties globalization and more low cost manufacturing, we're feeling the effects thirty plus years later. So like, it's not like this stuff is not going to cause some kind of dislocation. The only thing is like, this is where if in a perfect world, there's some thoughtfulness put into what it is doing and there's, I don't know, I even hate saying, but some kind of more mass upskilling of people. And as I say this, I literally am like, this will never happen in the current moment.

• 20:34Speaker 01

  But, but you have to prepare for this and any economy that does will be better off the same way any company that does will be better off. What do you think?

• 20:44Alex Kantrowitz?

  No, I agree with you completely. I just think that it's sometimes going to hurt. And if you're looking for a job, it's going to hurt. It's going to be much more difficult to find a job. And to keep your job, you're going to have to really show that you're using AI. I mean, I'm just looking at the, again, looking at Tobi Lutke's memo in Shopify. He says, "We're going to add AI usage questions to our performance and peer review questionnaire. Learning to use AI well is an unobvious skill. My sense is that a lot of people give up after writing a prompt and not getting the ideal thing back immediately. Learning to prompt and load context is important and getting peers to provide feedback on how this is going will be valuable."

• 21:24Alex Kantrowitz?

  You're also going to, every time you're going to create a prototype, you have to do it with AI exploration. He says, "You can learn to produce something that other teammates can look at, use and reason about in a fraction of time it used to take." Actually, I think this is smart. Like, I think that if you're going to try to get this to take off within your organization, you need to create this incentives and people will respond.

• 21:49Speaker 01

  Yeah, I mean, when he writes that, in a way, it makes me happy because I know he gets it. It's kind of like when the Klarna CEO was on, I was a bit skeptical about some of the stuff that was coming out in the press at first, but then the way he talked about your first prompt, you won't get the right output, and then your job is to keep working till you get it and learn how to do it. And I think him pushing that, it's exactly correct. And I actually think the new grads article that came out, I'm still gonna, I don't know, attribute it maybe at least more to the current state of the economy, because honestly, I would rather be a new graduate coming out right now than, I don't know, having been

• 22:31Speaker 01

  in the workforce seven, eight years and become completely habitual in a certain way of working. And I do think the idea of like, these skills are necessary. I remember probably like even three or four years ago, sometimes you would get people putting Google Docs on their resume as a skill. And I'm sorry, I'm sorry, like, no, no, come on. That's like Microsoft Office back in the day. Like, those things might have a certain window where it is considered an actual listable skill, but I don't know, man. Do you have that on your resume, Alex?

• 23:08Alex Kantrowitz?

  I definitely had Word and Excel on my first resume. There's no doubt about it.

• 23:11Speaker 01

  Yeah, the first one. First one. When was that?

• 23:14Alex Kantrowitz?

  Early ones. It was a long time ago.

• 23:17Speaker 01

  A long time ago.

• 23:18Alex Kantrowitz?

  So, but Shopify in particular wants to be able to do this because it seems like they're gonna just integrate more and more with AI. In fact, last week, we talked about how you might want to shop in ChatGPT. And then this week, it looks like we were on to something. OpenAI rolls out new shopping features with ChatGPT search update from Reuters. So it's gonna, OpenAI said it updated ChatGPT's web search capabilities to improve online shopping for users with personalized product recommendations with images, reviews, and direct purchase links. So if you're Shopify, you definitely have to do this because This is where it's going. It might end up being that your front page is not going to be

• 24:01Alex Kantrowitz?

  the vendor pages that you've created, but the actual user interface of ChatGPT. I tried to get Jamil Ghani, the head of Prime, to bite on this question in our interview on Wednesday, but was unsuccessful, although I imagine that they're going to be thinking about this quite hard.

• 24:19Speaker 01

  Well, that's actually a perfect example that if you don't have AI-native employees, you're not going to realize the urgency of how shopping can change. And they might not have pushed and worked and actually delivered, probably being, I think, like the first major e-commerce platform to actually deliver a native integration into ChatGPT. Because I believe perplexity for their shopping, I don't think it's with Shopify. I think they, it might be in-house or like, this is a big deal, but this is only going to happen because internally, people understood what is the shift taking place in terms of how people are shopping? And when everyone marries ChatGPT and needs to buy its stuff, which is, listeners can go back to last week's episode if that sounds a bit odd.

• 25:04Alex Kantrowitz?

  We're bullish on the relationship.

• 25:06Speaker 01

  Yeah, yeah, buying them, buying some GPUs. But you got to know how this, what it looks like, like you have to really understand the overall picture and flow of these things, otherwise you're not going to do the right things.

• 25:20Alex Kantrowitz?

  Now, how many companies are going to go out and say that they're AI-first to distract from the fact that they're struggling? I mean, when I first saw these headlines come out about Shopify, I had a couple of people that reached out and were like, this is just them trying to paper over a disastrous business. In March 2023, for instance, they laid off 20% of their workforce and gave up on logistics. And then they also did another big layoff recently.

• 25:51Alex Kantrowitz?

  And the suggestion that some made to me was, look, this is just them trying to sort of distract us from what's actually going on and have us look at the new shiny object. I don't think Shopify is a disastrous business, but we can look at the headlines and see that they've definitely had, to use another term, as bad as internal knowledge hub. They've had some speed bumps. They've had some headwinds.

• 26:14Speaker 01

  I think if any of this messaging came from a CEO or management that typically did not make public statements or was not in the public conversation, I would be more skeptical. If like someone, it was just kind of like a silent CEO who suddenly is coming out and is like AI AI, I'll be a little more skeptical. But again, Toby, he doesn't show up. I've been like, actually, he goes on podcasts too, right? But I mean, he certainly

• 26:43Alex Kantrowitz?

  I've been asking him Shopify. If you're out there listening to this show, get Toby on the show. We'll ask him good questions. I think you should do it. There's nothing to be afraid of.

• 26:53Speaker 01

  I uh, but, but I mean, he certainly tweets and is in the conversation. Duolingo is like, though, I mean, they, their, their brand persona is one of the like strongest out there, but uh, their public and Aaron Levie, I mean, it's like, I would say probably like the single most quality voice, but also prolific voice on agentic AI as well. So.

• 27:17Alex Kantrowitz?

  Yeah. All right. Let me ask you two big questions and then we're going to go on to big tech earnings. The first is, can this be top down? Like, can this really be something that a CEO commands a company to do and they end up doing it?

• 27:32Speaker 01

  So normally I would have said no, but I actually think the Shopify memo has an interesting take on it. Again, the idea that if you want budget, you have to demonstrate it could not be done in a different way. That is a good top down, like actually behavioral way to push the change. If it's just everyone spend two hours a week using whatever internal AI tool or system we're building, that's not going to work. But this is actually the way you get it done. So I think it could be done. I think it's going to cause some internal tension and speed bumps, as you uh, said, but uh, I, I think it can be done. What about you?

• 28:13Alex Kantrowitz?

  Definitely. Yeah, I was skeptical at first, but then I think that when you integrate something into performance reviews, that's when it gets done. Right? I think that like, performance management is what drives a lot of the tech industry comp. I mean, as you might imagine, money matters. And if your pay is contingent upon you spending a couple hours or, you know, building chat GPT into everything you do, you're going to do it. You want to, you want to get paid.

• 28:40Speaker 01

  I mean, it's no different than any other major tool set that came before. Again, using docs collaboratively was like transformational, I guess, but we all just kind of did it.

• 28:53Alex Kantrowitz?

  Okay, here's my second question for you. Obviously, this is happening within three very techie tech companies, right? Can we see this in non-tech companies? And if so, how long do you think that's going to take?

• 29:05Speaker 01

  That is a good question because I do agree that, I mean, yeah, if you are a digitally native, pure tech company, that's like lots of engineers, it's going to be easier. But yeah, that actually, I don't know. I don't know. That's a tough one. I mean, it could be. I would like to see something kind of dramatic soon, and let's say in the next few months. But I don't know if we will.

• 29:32Alex Kantrowitz?

  I think it will take years. I mean, that's one of the things that when we think about the transformative power of this technology, the transformation takes quite some time always. So even if this does become a thing everywhere in the tech industry for normal companies to implement it, it's just going to take much longer. It's my perspective, years, years, not months.

• 29:55Speaker 01

  I don't know. Maybe as the tools get actually better and more functional and intuitive, I think there's a shot. There's a shot. I'm more optimistic, I think.

• 30:05Alex Kantrowitz?

  You know what helps those companies or those tools become better, more intuitive and usable?

• 30:12Speaker 01

  Better products.

• 30:13Alex Kantrowitz?

  No, model improvement.

• 30:16Alex Kantrowitz?

  Good try, though.

• 30:18Speaker 01

  Better UI.

• 30:21Alex Kantrowitz?

  All right. No, I mean, there's a case to be made for both. Let's talk a little bit about big tech earnings. Actually, it's been a very interesting earnings season. So we're talking here Thursday night. We're doing Friday edition a little early because I'm about to go to France. Let's talk about it. We had Apple, Amazon, Meta and Microsoft all report over the past couple of days. And the reason why I pay attention to these companies, I mean, among many reasons, but one of the most interesting things about them is they can sort of be a read through to the rest of the economy. If things are going well with Amazon, then things are probably going well with the consumer. And if things are going well with Microsoft, then enterprises are probably faring well. Companies by and large beat their earnings expectations.

• 31:04Alex Kantrowitz?

  The big tech companies by and large beat their earnings expectations in the quarter that they just reported. Remember, this is Q1, so it's January, February and March before Liberation Day. But Apple reported, this is from the Wall Street Journal, the highest first quarter or March quarter revenue it's seen in more than two years as people moved quickly to buy smartphones and other devices before the tariffs were announced in April. The company said sales rose 5%, 95 billion, which was ahead of analyst expectations. However, the China revenue came in just a little bit lower than expected. So they're going to clearly see some issues in the China market, especially after the

• 31:49Alex Kantrowitz?

  US and China go head to head and China's putting out propaganda like China won't kneel, which is if you watch that and you're in China and then you go to the store to buy a smartphone, it doesn't seem like you're going to run the American version, given the signals your government is sending. And also iPhones are banned within Chinese government offices. And then you look at the, another interesting thing that we heard was demand, like production. So this is from CNBC, half the iPhones sold in the US were made in India. And Gene Munster, an analyst at Loop Ventures, said he expected the figure to be close to 25%.

• 32:29Alex Kantrowitz?

  So it's already, the US, the Apple seems to be driving much of its manufacturing capability to India, especially for the US market. One last stat and then I'll turn over you. Mark Gurman talked about the tariff impact. Apple said there was no tariff impact in the March quarter. It cannot precisely estimate for June quarter, which is the current one, but estimates $900 million in cost increases in the period. This is with an exception for iPhone. What's your big picture take on what's happening with Apple? It seems like it's weathered the worst of it.

• 33:03Speaker 01

  I think anything that happened pre-liberation day, January through March, obviously tariffs people expected were coming, but I mean, I guess attributing, I don't know, like, did you personally, in the months of January to March, were you like, I gotta stock up on stuff before? Were you panic buying?

• 33:23Alex Kantrowitz?

  No, I just bought, I mean, I bought in April, after liberation day. I bought a couple pairs of sneakers from Amazon, that was it.

• 33:30Speaker 01

  No, see, that's the thing. I think any kind of like, even the conversation around panic buying really started after liberation day. So, I mean, to their credit, it looks like, and the release of the iPhone 16e was also cited as one of those factors, and kind of like, the lower cost, but still, AI, Apple Intelligence-enabled phone. So, I don't know. I think, I still believe, and we're certainly going to get into Apple, the ex-tariff conversation, but, I don't know, Apple still is the most vulnerable, but the most promising part of that was the 50% in India, that blew me away. I mean, it blew Gene Munster away, but I

• 34:09Speaker 01

  mean, 50% already. I guess, supply chain, supply chain Tim has been pretty good on this stuff already.

• 34:18Alex Kantrowitz?

  It's almost as if, yeah, if you're gonna take one CEO who's ready for this, it is the guy with the supply chain.

• 34:25Speaker 01

  Supply chain Tim.

• 34:25Alex Kantrowitz?

  But I don't think this is gonna be the end of the story. I mean, okay, so Cook also said that the tariff hit will be about 1% in the June quarter, which is less than most had been expecting, according to Gene Munster. Again, Apple, Apple had this exemption from tariffs, but that is the export side of things. So, I think the real thing is the China market. 17% of Apple's revenue last year came from China. It was 15% in Q1. I don't know the percentage yet for this year, but I think it was kind of in line on that front. A shrinking of Apple revenue in China is going to be bad.

• 35:08Speaker 01

  Yeah, no, I'll agree there. And actually, that's, I mean, not just bad, but potentially catastrophic. Cause like, I have to imagine out of all things, like, an expensive cellphone purchase. Like, and again, we will get into Apple, the product side of it, but the, it's more of a status symbol than anything. In the same way, like, I mean, Tesla, we've seen the brand impact on the product sales. Like, when the iPhones, one of its primary values is that of a status symbol. And suddenly that status in China could be, have a very negative connotation, rather than this means you are cool and wealthy.

• 35:49Speaker 01

  That could be a huge problem, independent of any actual, like, regulatory or legal restrictions, just from pure brand standpoint.

• 36:00Alex Kantrowitz?

  And then the other side of this is, let's say, okay, we didn't see any price hikes, right? But let's say the price has to go up because of this, or we lose China downloads because of this. The thing that Apple's selling and has growth potential is services, which is a hundred billion dollar business. And has been growing at an insane clip and has, and has major margins. So you want to be able to sell as many phones as possible, so you can sell more services. And if tariffs cause an issue there, then that could be a problem for Apple. But I have to say, candidly, my reaction to what we've seen and heard from Apple in the call so far and from earnings is, actually, it's not as bad as feared, to use a line from Dan Ives.

• 36:44Speaker 01

  Yeah, I think that's, the market at least responded accordingly, that, I mean, I think we ended up today, as we're recording on Thursday, only up a little bit, we were up more earlier, but, I mean, overall-

• 36:57Alex Kantrowitz?

  The stock did trade down in after hours, but not, not tremendously. It's down 2% in after hours.

• 37:02Speaker 01

  Oh, yeah, I guess, Apple. Overall, from the earnings, just this, this narrative that it's not as bad as we were thinking, even though the time period that all these companies are reporting for is before Liberation Day. Still, the idea that there's some, like, good feeling that actually the economy is humming along at some capacity still. That, and basically, I guess, if we- Go back to zero tariffs or something like that. Everything will be okay again. Seems to be the the feeling right now.

• 37:34Alex Kantrowitz?

  But something beyond tariffs can hurt Apple. And that is aimed directly at its services revenue. Because for a long time, Apple had told app developers that they could not tell other people that they could buy subscriptions to their products through mechanisms that were not the Apple App Store. So for instance, if you were on Netflix, Netflix couldn't say, just go sign up on Netflix.com and you will avoid the Apple App Store tax. Then the federal government had something else to say. Why don't you cue us into the story, Ron, John?

• 38:14Speaker 01

  Okay. I think this is one of the biggest tech stories in a long time. And I think this really poses a huge threat to Apple. And I'm, I will inject my own personal feelings about, as a longtime Apple fanboy, but what I've been seeing happening to the company. So there's been this trial with Epic Games for, that's gone on almost five years that actually, and again, it's amazing that like, through different administrations, through different political climates, actually ended with the judge ruling that Apple, not only is it two years ago that there was an injunction against the idea that they take a 30% commission

• 38:55Speaker 01

  and block any other way, or are very overly restrictive about how app, app developers, as you said, that you can't include any language, go to Netflix.com to pay. And I'm sure many listeners have dealt with this. Like if you try to buy something, I think if you try to like on Amazon, sometimes you're a book, you're redirected to the Amazon app on the mobile web. Like it really creates bad overall UX flows. So then what happened is after that initial injunction. Apple basically did a bunch of workarounds and still were forcing app developers to pay 27%. And the judge came out and said, that is completely not okay. And now you have to

• 39:40Speaker 01

  give, allow the options to actually let people go and pay elsewhere outside of your actual like embedded app payment ecosystem. There's two big things that came out of this. First, so in 2023, after the initial decision, Phil Schiller had actually said that maybe we move to a model where Apple takes no commission. But apparently, Apple's finance chief had come up with the idea around 27%. Tim Cook sided with the finance chief, Luca Maestri. So you have a situation where it's very clearly documented and now in the public record that even after this initial injunction, that they basically worked around for circumventing

• 40:24Speaker 01

  this. The second thing that this is actually the craziest part to me, because as a user, I felt this. If any, if listeners have tried to go and pay even like click out to a directly to a payment site on an external website, there is a Slack, Slack conversation that leaked through this or were in the court filings, where they talk about, they're like, okay, external websites sound scary. So the execs are gonna love it. Basically like saying you're going to an external website. And I've you have been in that position. You're like, oh, wait, maybe something's insecure right here. And then it even says, wait, to make

• 41:05Speaker 01

  your version even worse, you can add the developer name rather than the app name, just to make it look overall a little shadier that if you're gonna go do this, to which the other developer responded, oh, keep going. Like, they were sitting there like, how do we make this? Seem like this kind of like shady spammy transaction when in reality it was not. And it just made me start to question like Apple's entire brand and what has gotten them to where they are today is the idea they're the ones for privacy and security and the user interest when all these other tech companies are sucking up your data and leading you through dark patterns.

• 41:47Speaker 01

  And suddenly you're seeing, this was November 2021, this was being discussed, like the dark pattern thing, they're in there. And to me, that can completely change the perception of the entire company.

• 42:01Alex Kantrowitz?

  Well, I used to think that, but Apple has been able to weather so many storms like this and nothing seems to stick. And I think it's just because people love iPhones so much that like this is not going to resonate with the individual user. So I think that that is not going to be a big deal. Also, I just love that the term external website sounds scary. Like, excuse me, what, what is an external website? Like, what the hell am I on right now? It's still a website. It's just a website.

• 42:32Speaker 01

  Yeah, but to the average user, I get it. Like, anything that makes it seem, yeah, anything that makes it seem like you are no longer in the safe walled garden of Apple's payment ecosystem and your credit card is gonna get, it could get stolen out there, who knows? But to me, there's another big thing and the pragmatic engineer, Gergely Oroz, is that how you say his name?

• 42:57Alex Kantrowitz?

  Gergay.

• 42:58Speaker 01

  Gergay, Gergay. So he had a great, he had a great thread and he talks about how the actual payment ecosystem is like a perfect representation of monopoly because it has not improved. The whole idea is this 30% fee is supposed to allow you access to this robust, safe, secure ecosystem. It's good for consumers, it's good for app developers. But he pointed out all these things that I hadn't really thought about, but are table stakes for any kind of subscription e-commerce. Giving a full or partial refund directly, like providing customer service messaging

• 43:41Speaker 01

  around any kind of, uh, any kind of transaction or especially refund, providing certain types of discounts rather than just a three or seven day or 30 day trial. Like, all of these things are just so normalized and part of like, running a good subscription type service, and Apple doesn't allow them. And then you start to think like, and then as he points out, it's because they haven't had to improve it at all. So again, the promise that it actually was the best place to be, and that's why you should pay 30%. In reality, this was like, the most cleanly, clean example of monopolizing behavior where it just was

• 44:23Speaker 01

  pure market weight and dominance. And in fact, that experience got worse and worse. And I'll say this, like, have you canceled a subscription on iTunes slash App Store, whatever, whichever one it's on now?

• 44:38Alex Kantrowitz?

  I think so.

• 44:40Speaker 01

  Like, it's 10 steps. You have to log into your iCloud account, which who goes into iCloud ever? You have to go file a dispute. You have to like, like, it's the most heavy, ugly, annoying process ever. And I'm so convinced that like, they make a lot of money off of people just not knowing how to do that, or not even realizing or not caring. And again, think about how many apps are like, seven day trial and then 69.99 for one year. Like, they've pushed the whole ecosystem towards the more scammy side of the way this stuff can work versus the rest of the internet.

• 45:19Alex Kantrowitz?

  I'm with you. I think this is the most egregious thing that Apple does with its power, or the most egregious thing I can think of right now. And I do love that the federal government is trying to Get them for it. And frankly, the reason why I think it is such a righteous crusade here is because the one that ends up getting hurt is the customer, always. Like, why am I paying 30% more for Netflix? It's the same service. I mean, I know Netflix doesn't even take payments on Apple's ad platform. So that's an extremely righteous reason to pursue action here. And just to bring it back to our discussion of where Apple's business goes, App Store fees are a

• 46:02Alex Kantrowitz?

  major part of its services revenue. Like, Apple's big revenue growth moment, again, is services. Wait, what services does Apple provide? Now, of course, there's Apple Music and Apple TV+ but a big part of the service, quote unquote, that Apple provides is extortion.

• 46:20Speaker 01

  30 billion, 30 billion. That's the estimated revenue from the, directly from the payment side of it. I mean, that's a lot of money. And what is anyone getting for that? That's the question.

• 46:33Alex Kantrowitz?

  Yeah, now, of course, I don't mean extortion, like in the legal sense, but sometimes it feels like there's, you know, you sort of, you know, be terrible if something happened to the nice business you had over there with that company. All right, let's talk a little bit more about what's going on with big tech, because this does, again, have a read through to the economy. Very interesting stuff with Microsoft, which crushed earnings expectations. Microsoft Chief Financial Officer Amy Hood said the band signals across the company's business have remained consistent so far this month. So everything's still going quite well in April. And Azure grew 35% year over year. I think that's OpenAI. What's your read, Ron John?

• 47:15Speaker 01

  I mean, I think the the general theme from this week was AI is still powering the economy. Like, I mean, maybe that's, that's why you got to be an AI-first company. But like, I definitely think, and again, in the whole, how much is of that money is Microsoft's own investment money coming back to Microsoft? I mean, this stuff gets a little difficult to fully understand. But but overall, I mean, I mean, when Azure is growing at 35% year on year, that's, that's pretty spectacular. And again, I think it's a sign that Every company is moving towards that paradigm and so you better better get on.

• 48:00Alex Kantrowitz?

  That's right. And Amazon, AWS grew 17%, missing expectations. I don't know. I think that next quarter is really going to be the interesting quarter to watch for Amazon. Did you know that 60% of the company's e-commerce sales came from third-party merchants? Okay, come from third-party merchants that get many of their products from China and then sell on Amazon's website? I mean, I tried to get this, speak to Jamil Gani about this, but he wouldn't give a number also, the head of Prime, but this is coming out in places like The Journal. The company is way over indexed on China, whatever way you look at it.

• 48:35Speaker 01

  Yeah, I think Amazon potentially more than Apple, or at least equally is in trouble. And actually, this is a good sign, the like, do you want supply chain Tim? Might not make great product with Apple Intelligence, but he can shift supply chains faster than anybody else. But I mean, Amazon made this strategic shift a number of years ago to push for more Chinese suppliers and allowing Chinese suppliers directly onto the platform rather than Chinese suppliers selling to American businesses that then resell on the platform. So,

• 49:11Speaker 01

  I mean, it's been a long time coming and obviously they kind of had to do it in many ways to compete against Shein and Timu, but they're they're gonna, this is gonna be a big issue and we're gonna see this stuff, yeah, play out a lot more directly in the the months to come. But hey, maybe you don't need $30, Alex. Maybe two are good enough.

• 49:31Alex Kantrowitz?

  Isn't that what Trump said?

• 49:32Speaker 01

  Yeah, that's what Trump said.

• 49:33Alex Kantrowitz?

  What was the New York Post headline? Oh, yeah. The New York Post headline, Trump admits tariffs will raise some prices cause shortages. Maybe children will have $2 instead of $30. And there's a picture of Barbie and Trump and the headline is skimp on the Barbie.

• 49:49Speaker 01

  Oh, God.

• 49:51Alex Kantrowitz?

  Classic.

• 49:52Speaker 01

  Do you think AI is going to replace the New York Post headline, front page headline?

• 49:57Alex Kantrowitz?

  Unfortunately, probably.

• 49:59Speaker 01

  You know what?

• 50:00Alex Kantrowitz?

  That's classic. It's very human.

• 50:01Speaker 01

  Yeah, that's a good one, though. No, you know what? AI can't come up with that one. That's too good. That's original.

• 50:08Alex Kantrowitz?

  Yep, skimp on the Barbie.

• 50:09Speaker 01

  Textual. It's never been done before.

• 50:13Speaker 01

  New York Post is probably the single most defensive, uh, front page headline writer is the most defensible job against AI.

• 50:21Alex Kantrowitz?

  Don't tell that to Marc Andreessen, who said this week that it was VC.

• 50:24Speaker 01

  Early stage venture.

• 50:25Alex Kantrowitz?

  Early stage venture. Hmm. Confirmation bias, maybe. All right. Very quickly, Facebook, 42 billion in sales in the first quarter. Uh, I think that is kind of insane. Zuckerberg says they're well positioned to weather macroeconomic uncertainty. Capex spending that they're going to go from AI is going to go, the original range was 65, uh, 60 to 65 billion. Now it's going to be 64 to 72 billion, largely on artificial intelligence investments. Meta, I've been through many cycles with Meta and it's hard to keep them down for long. They find a way.

• 51:01Speaker 01

  Yeah, I think Meta, the, the digital advertising impact is almost going to be like a second order effect versus the manufacturing one, which is going to be a lot more immediate in this quarter. We'll start to see that downstream and we haven't, I mean, I know Snap issued some like headwinds guidance, um, but, but I think Meta is overall relatively pretty well, well positioned right now. And I also say that, that now that it's been getting warm in New York, I'll admit, I am loving my Meta Ray-Bans even more. Like it, wearing Meta Ray-Bans, walking around New York City, listening to music on it, asking

• 51:43Speaker 01

  it questions while you're just wandering around. It's a good product. It's a really good product.

• 51:50Alex Kantrowitz?

  Now, would you wear them in a podcast interview like Zuckerberg has been this, this week?

• 51:53Speaker 01

  Absolutely not. Absolutely not. I got them in December and I haven't used them that much, but they're cool, like on vacation and when I was skiing and stuff, but, but now wearing them outside when it's sunny out isn't weird. So then you can just go around and actually use them. I still will not wear them. I, I don't have the transition lenses though, so they're just the dark ones.

• 52:15Alex Kantrowitz?

  Got to get those.

• 52:16Speaker 01

  I'm definitely not, not wearing it inside, but I, I don't know if I'd wear it. Maybe if we're an AI first podcast, Alex, both of us.

• 52:24Alex Kantrowitz?

  Gotta wear those Ray-Bans.

• 52:25Speaker 01

  Gotta wear the Ray-Bans.

• 52:26Alex Kantrowitz?

  All right. Next time we're in person, let's do it.

• 52:29Alex Kantrowitz?

  Okay, let's take a quick break and come back and talk about Nvidia and Anthropic going at it. And then this great letter that we got from a professor from a philosophy from Wilford Lawyer University. We'll do that right after this. This episode is brought to you by AppPoint. Everyone's racing to roll out AI right now. Co-pilots, chatbots, agents doing real work. But here's the part nobody loves talking about. All that AI runs on your data, and most teams have no single way to see it, secure it, and prove it's under control. That's exactly what AppPoint does. For 25 years, they've been the trusted layer beneath the world's most demanding data, now extended

• 53:09Alex Kantrowitz?

  across your entire AI estate. Your data, your cloud, and the agents acting on your behalf. It's how more than 28,000 organizations deploy AI with confidence. So innovation scales without scaling risk. AppPoint, the unifying trust layer for AI. Head to apppoint.com to see how enterprises deploy AI with confidence. Learn more at avpt.co/bigtechnologypodcast. That's avpt.co/bigtechnologypodcast. Violent threats against executives are growing at an alarming rate. When a company experiences backlash, executives are the first people blamed. When their address and other personal

• 53:52Alex Kantrowitz?

  details are sitting online, this backlash can lead right to their front door. The team at Ironwell knows this better than anyone. They've protected some of the most targeted executives and individuals on the planet for almost two decades. As a white glove enterprise security service, they protect your people with continuous personal data removal, proactive prevention tools, and human-led emergency support. So when someone goes looking for your executives, Ironwell ensures they hit a dead end. So here's what to do. Go to ironwell.com/bigtechnology, fill in the quick form, and request your free risk assessment. Their team will show you just how exposed your executives are and how to lock it down before a threat reaches their front door. That's ironwell.com/bigtechnology. Learn how to stop online threats before they

• 54:35Alex Kantrowitz?

  become real-world attacks. The link's in the show notes.

• 54:42Alex Kantrowitz?

  And we're back here on Big Technology Podcast Friday Edition, coming into the home stretch. Very quickly, Ranjan, I know we have this in our headline of the story of the episode, but let's just talk about it. It's kind of interesting. Nvidia and Anthropic are going head to head. I don't know if you caught this, but Anthropic really doesn't like the fact that chips are getting into the hands of developers in China. And they said that the chips are being smuggled in to China in baby bumps or alongside live lobsters. And they would like an increase of funding for export enforcement. Export

• 55:24Alex Kantrowitz?

  controls are only effective with proper enforcement, Anthropic says. Enhanced resources for the Bureau of Industry and Security would significantly improve control effectiveness. Now Nvidia, which sells a lot of chips into Singapore, and I think we can say with a good deal of certainty, many of those end up in China, is not happy about this at all. And on Thursday said, American firms should focus on innovation and rise to the challenge, rather than tell tall tales that large, heavy, and sensitive electronics are somehow smuggled in baby bumps or alongside live lobsters. Very interesting fight. I mean, it is interesting that we're going to start to see American firms become more

• 56:06Alex Kantrowitz?

  and more concerned about what's going on in China. And Nvidia, which I think depends to some degree on selling them, is going to fight back. And we're seeing that play out. And among two companies who really don't want to be publicly fighting with each other. What do you think?

• 56:21Speaker 01

  I think I'm convinced Trump is going to address this one. I'm making the call right here in the next three to four days because it's just too perfect. If you hear smuggled in prosthetic baby bumps or alongside live lobsters.

• 56:37Speaker 01

  I mean, that's just too enticing to make a comment on, but I don't know. I mean, we're going to start to see more of these very nuanced divisions like that you wouldn't quite expect. And I agree, like at that scale, and I'm sure Anthropic is a highly dependent customer on Nvidia as well. Like it's pretty surprising that they're doing this, but I don't know. And Nvidia has to play both sides, basically. So I think we're going to see more of this coming out, especially as whatever happens on the trade side happens, we'll definitely see more.

• 57:13Alex Kantrowitz?

  Not to be a guy that plays both sides of the fence here, but I definitely am on team Anthropic that like, we know whatever export controls are in place aren't working. So if you're serious about the export controls, you may want to try to enforce them. And then I'm also on Nvidia's side here that if you wanted to smuggle a GPU across borders, it's going to be very difficult to make it look like it's just a normal pregnancy. And a GPU certainly does not look like a lobster. That's just my opinion. I would, I could pick out a GPU server next to a lobster.

• 57:50Speaker 01

  I don't know. For, for my chipmule behavior, I like to go with the lobsters over the baby bumps, you know, it's just lobsters. Actually, do we export lobster to China? I think I remember reading a while ago, that is a big export to China. So might as well.

• 58:10Alex Kantrowitz?

  Makes more sense. This is why Red Lobster went out of business is because they started serving GPUs instead of the actual thing.

• 58:17Speaker 01

  Actually, the next one is, you know, like another major export to China from America is the like leftover pig and cow parts because like things that are chicken feet as well, things that are not eaten in the US. So yeah, keep an eye on those chicken feet to the customs officials. I have a feeling that's where they're going next.

• 58:39Alex Kantrowitz?

  I'm just imagining a, a man sitting down for dinner at a Red Lobster and tucking his little napkin into his shirt collar and being so excited, drinking a little beer maybe. And they bring the plate out and they say, sir, here's your lobster. And it's an H100.

• 59:01Speaker 01

  It's an H100.

• 59:02Alex Kantrowitz?

  God damn it.

• 59:05Speaker 01

  Send it back.

• 59:05Alex Kantrowitz?

  That's not what I ordered.

• 59:06Speaker 01

  Send it back.

• 59:08Alex Kantrowitz?

  This is not well done.

• 59:11Speaker 01

  I like that we went down that road rather than the prosthetic baby bump because that raises a whole host of ethical issues.

• 59:17Alex Kantrowitz?

  So I don't even know how you would do that. Okay.

• 59:21Speaker 01

  No, no, let's not.

• 59:22Alex Kantrowitz?

  Let's talk about this great letter that we got from Professor Brian Williston from the Department of Philosophy at Wilford Lawyer University. So we spoke about AI consciousness at the beginning of last show, and we always love hearing from listeners, especially when there's thoughtful feedback that will help us advance the discussion and think a little bit more. So what Professor Williston said was that we talked about how there's consciousness, which is what you feel, and intelligence, which is knowledge.

• 59:53Alex Kantrowitz?

  He said, here's the problem. In assessing whether or not any entity is consciousness, is conscious, all we have to go on ever are displays of intelligence. The only bit of consciousness in the universe that we can access directly is our own, and we take it for granted that other humans are indeed conscious. Why? Because I make an inference from their displays of intelligence, which I can access to what I presume is going on inside of them. This is exactly the same boat we are in with respect to LLMs. Intelligence in all cases, humans or machines, is all we have to go on. Well, that really threw me for a loop, I would say.

• 1:00:34Alex Kantrowitz?

  I definitely helped expand my mind here on this question because, wait a second, if the only thing that we can determine whether something is conscious or not is its intelligence, whether we basically look at them and say, does it, is it, does it make, is when, is there something going on inside this being that is like actually being versus just existing like a rock? Then who are we to say that machines aren't conscious or can't be conscious? Because I think my implication was if you can't feel, you can't be conscious. But how do we know? Now, I'm still on the side of this AI isn't conscious, but I do kind of think that it's

• 1:01:16Alex Kantrowitz?

  a very interesting prod at the ideas we shared last week and want to turn it over to you to get your response.

• 1:01:24Speaker 01

  My first reaction on this was, this is why I love our listeners. Honestly, like, actually, genuinely intelligent, thoughtful, and like, long responses to what we say. I mean, that's, it's kind of like, reflects the type of conversation in the Discord channel as well. Like, this, I love getting this kind of email from listeners. But then, the other thing too is, I agree, it made me at least be like, who are we to say that machines are not conscious? Like, it's one thing to be like, I don't think they are, but to definitively say they are not,

• 1:02:06Speaker 01

  given kind of how he, the framework that he set up, and even how we were talking about last week. Now, I think maybe I am in the camp of, I will not take a position and I will recuse myself from, because again, our displays of intelligence are talking about lobsters and GPUs, so. I don't know what that says about our own consciousness.

• 1:02:28Alex Kantrowitz?

  Right. I don't know, I guess like, the other side, the other definition is self-awareness. And you could say that these AIs today are already self-aware.

• 1:02:37Speaker 01

  Yeah.

• 1:02:38Alex Kantrowitz?

  Maybe that's a stretch.

• 1:02:39Speaker 01

  No, no, but there's been a lot of behavior that show, actually no, but we didn't even talk about sycophantic ChatGPT. It did not solve itself, they had to go into the system prompt, so. So I think, I don't think we're self-aware yet.

• 1:02:54Alex Kantrowitz?

  Okay, but I think that we can certainly see a path to it getting there.

• 1:02:58Speaker 01

  Maybe, I think we'll.

• 1:03:00Alex Kantrowitz?

  Self-awareness is, I think is a lower qualification than consciousness.

• 1:03:04Speaker 01

  Is it before or after AGI?

• 1:03:08Alex Kantrowitz?

  Maybe it's right at AGI. I don't know.

• 1:03:11Speaker 01

  Again, if a Waymo can drive the streets of New York City.

• 1:03:15Alex Kantrowitz?

  We'll call it.

• 1:03:15Speaker 01

  I'll give you self-aware AGI, not conscious, but close.

• 1:03:19Alex Kantrowitz?

  Well, maybe we can end with the greatest sign of technological progress we've seen in millennia, if not the length of time that's longer than millennia, maybe eternity? I don't know. What's longer than millennia? Double millennia's?

• 1:03:34Speaker 01

  Double, 10, I don't know if there's a unit for 10,000 years, but we'll check with AI, because we're an AI-first podcast in a moment.

• 1:03:43Alex Kantrowitz?

  And it is a beautiful new gadget that you're gonna tell us all about, called the Mag 2.

• 1:03:49Speaker 01

  I am now, as we've been doing more video, I, with respect to the YouTube commenters who are like, why is Ron John looking down at the camera like that? And what is this angle? And I'm someone who enjoys some good framing in any kind of video. This company Plexycam has this new, it's a magnetic thing that attaches to a webcam. So you can put your webcam in the middle of your screen, right over the other person you're talking to or right next to it. So you can look, when you're looking at them, your eye contact for the video is actually looking into the camera. A lot of people, I'm guessing if you've done video, you know, you kind of try to force yourself to look up to

• 1:04:34Speaker 01

  the middle of your laptop. I'm a big fan. This is my first time using it. We'll see how the video comes out, but I'm a big fan of this. And I'm going to put this under the camp of AI first podcasting.

• 1:04:48Alex Kantrowitz?

  All right.

• 1:04:48Speaker 01

  I don't know why. It's, it's actually magnetic. It's one of the oldest technologies in the, in the world.

• 1:04:54Alex Kantrowitz?

  Magnetism, the original artificial intelligence.

• 1:04:57Speaker 01

  Magnetism? There was a big transformation where you had to go magnet first, otherwise you were.

• 1:05:02Alex Kantrowitz?

  It's going in the performance review. Yeah. Magnet deficient. Get out of here.

• 1:05:08Alex Kantrowitz?

  Speaking of which, let's do that ourselves. Ron John, great to see you as always. Thanks for coming on the show.

• 1:05:13Speaker 01

  All right. See you next week.

• 1:05:15Alex Kantrowitz?

  All right, everybody. Thank you so much for listening and watching. It's always great to have you here. Join us on Wednesday for a discussion about whether scaling is over with Gary Marcus and AI skeptic. And then Ron John and I will be back on Friday to break down the week's news. Until then, we'll see you next time on Big Technology Podcast.

• 1:05:34Speaker 02

  Now, more than ever, we need someone on our side. But Rob Whitman? Whitman's votes drove up costs, cut Medicaid and risked 350,000 Virginians healthcare to pay for tax breaks for himself and his donors. He's not on our side. Shannon Taylor will fight corruption, cut taxes and bring down costs. This November, vote for Shannon Taylor. She's for us. Paid for by HMP, the HouseMajorityPac.com, not authorized by any candidate or candidate's committee.