The Internet's Impact On Productivity, Meta's AI Plan, RIP Big Tech Perks

0:00Capital One's tech team isn't just talking about multiagentic AI. They already deployed one. It's called Chat Concierge, and it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. The truth is AI security is identity security. An AI agent isn't just a piece of code. It's a first-class citizen in your digital ecosystem, and it needs to be treated like one.

0:40That's why Okta is taking the lead to secure these AI agents. The key to unlocking this new layer of protection? An identity security fabric. Organizations need a unified, comprehensive approach that protects every identity, human or machine, with consistent policies and oversight. Don't wait for a security incident to realize your AI agents are a massive blind spot. Learn how Okta's identity security fabric can help you secure the next generation of identities, including your AI agents. Visit Okta.com. That's O-K-T-A dot com.

1:18Welcome to Big Technology Podcast Friday edition, where we cover the week's news in our typical cool-headed and nuanced manner. Joining us, as always, is Ranjan Roy, back from vacation. We have plenty to talk about we're gonna talk about the internet's economic value We're gonna talk about AI moves inside meta and Google We're gonna talk about substack and its fight with Twitter and the state of the newsletter platform itself And then also where are all the perks going inside tech companies that and more coming up over the course of the next hour or so Ranjan welcome to the show. It's good to be back.

1:50Great to have you back. I missed you man. How was Paris? It was good. There was trash on the streets. There was a lot of trash on the streets. I did not see anything burning directly in front of me, but our travel was a little messed up some of the time based on the strike. So what you're seeing is a bit real. So now you're back. And one of the stories that you read on your break was this story that Paul Krugman wrote about the value of the internet and whether actually it helped whether it actually helped increase productivity so basically what Krugerman is saying is that everyone's talking about the internet has been great for people's productivity maybe not though he says nobody's arguing the internet has been useless surely it has contributed to economic growth the argument instead is that its benefits weren't exponentially large compared with those of earlier less glamorous technologies for example in 1920 only about one in five US households had a washing machine By 1970, almost everyone had access to one.

2:51Don't you think that made a big difference? Are you sure that it made less of a difference than widespread access to broadband? So kind of a contentious argument. What do you think about it? Yeah, this one sparked a good amount of debate with some friends because, and I do think it's interesting. on one side, it does feel a bit for, I'm sure many listeners know Paul Krugman still catches flack for his prediction, I think in the late 90s, that the internet would not amount to much. That seemed to have gone viral lately. So this New York Times opinion piece he wrote almost seemed as a response to this argument where he does intellectualize the idea that yes, in fact, the internet turned into an economic disappointment.

3:34But I do think it's a really interesting argument because it's the idea that productivity, like labor productivity is not up. Total factor productivity is not up. The internet did not actually generate massive productivity gains in aggregate across the economy. For very specific sectors, it did. But en masse, it actually did not create some massive increase in productivity the way the industrial revolution, the power loom, the steam engine, these inventions really changed the entire economy in a fundamental way. He's arguing that the internet, at least from the way we measure output, hasn't. And it's true because it made me think, just take a business like TikTok.

4:19The actual measurable economic output that's generated are the advertisements created and the revenue generated from that. everyone sitting there and spending time scrolling through videos, that's not actually measured as economic activity or output. The jobs created, the labor created, those things all matter. But overall, if this is one of the biggest companies in the world, it's actually not creating a ton of GDP for any of these countries. Right. So let me throw out a theory here that I'm sure is I'm gonna caveat it's not fully formed I think that our measure of productivity at least in the US is wrong and we've effectively solved productivity with all those inventions that we had in the 1920s that we're talking about the fact that we're able to produce things at low cost that's a soft problem in some ways in the US and you know we definitely have issues with poverty but basic needs have largely been met and you know now all of our economy is not really working towards how do we produce more goods more goods it's largely a service economy it's an internet economy which requires a completely different form of measurement when you want to think about growth what do you think about that no no i think that's fair i think even when you think about companies and the way they measure their actual i mean revenue is still in terms of goods sold or services sold but in terms of actually understanding assets like you You can take a building or a printer and ascribe some kind of tangible value to that, but an actual algorithm you can't.

5:53And that does not sit on a firm's balance sheet. So I totally agree that the way we measure everything around digital goods and assets, it doesn't really make sense. And I think that's why I thought it was a good piece, even though I felt he was kind of just trying to cover his ass a little bit from his 1990s predictions. But it is. It's a reminder that on one side, I agree the way we measure has not caught up to the way the actual economy works. But on the other, I think it is interesting that, you know, in aggregate life expectancies going down, inequality increasing. Overall, did economic output and quality of life, how much has it really improved in the U.S.

6:35since the advent of the Internet in the last 25 years or whatever, 20, 30 years? versus the way at least the Silicon Valley VC community or others try to claim that this has completely transformed everyone's life and the way we live for the better. Well, I think everybody would be happy to have an iPhone or a smartphone, and that's definitely made life better. And we'll get to that actually at the end, about the degree to which it has, because we have a fun story to tell. That'll be our fun story of the week. but then again like there is something that I kind of agree with Krugman about which is that once you get basic needs met like what are you doing you know what is our economy all about and sometimes it is a question because it feels like in some ways that it's like you know okay so there's money that's been made to you know on production and then the rest of it is kind of like a wealth accumulation game for many people for large sectors of the economy and in entertainment and a service game and you know as you mentioned inequality has increased as we're doing this and you know a lot of people out there are suffering even though they're able to you know eat and put a roof over their heads and it's a question yeah no i mean you even think about the revenue per employee at one of the big tech companies has definitely you know multiples like exponential multiples higher than a railroad company worker in the 1920s or something like that.

8:04But that's why I think in terms of actual technological advances that, you know, increased overall economic output and productivity for the society at large. You know, I think it is an interesting question with the power loom or the steam engine or any other transformative technology. And where does the Internet at least sit in there? I think is up for debate. And I think he even recognizes that, I mean, we'll get into AI, I'm sure, even though I've been gone a few weeks and there's probably a new GPT that's been released or whatever else has happened. You're up to seven now. Damn it. But yeah, like maybe could that be the truly transformative technology where as logging into websites wasn't in terms of actually creating genuine economic output?

8:55I think it's a good debate. And I think it's one that, again, Krugman can definitely spark the ire of certain groups. But I think he raises a really good point here. So we're going to move on to AI topics. But I feel like there's one more thing we need to hit on this. And you just brought it up, whether artificial intelligence will be able to increase the productivity that we have in the economy. Right. That's the big deal. And there are people talking about how an AI can take one engineer and make them a 10x engineer. I could take a 10x engineer and make them a 100x engineer, effectively increasing productivity because it can help them code.

9:32Then you think about all business analysts, right, like the business analysis, business intelligence inside companies. You dump all the numbers into the AI. It spits out basically the findings. So you can end up going to work on higher value tasks, improvements in medicine, improvements in other areas of our economy. Do you think that we're going to, and obviously we don't know yet, but do you think that AI might be able to show not only that kind of intangible productivity gain, but also the tangible productivity gain that we would see in the numbers that Krugman is looking at? Yeah, I think I actually this is where I am bullish on AI.

10:05And I think especially from kind of a measurable economic output standpoint, AI hits that exactly in terms of increasing productivity, increasing the end output on the in the business world in terms of different problems or services or goods or anything versus I mean, if you think about like, the ability to more quickly connect with people via email versus a phone call has been incredible at a personal level, the way being able to connect with people. able to really stay in touch with people. But has it been exponential in terms of business productivity and output? Maybe not in the way you're saying.

10:40I do think, especially if we've realized the hype, AI can definitely be the type of thing that really changes the way business is done for the better. I'm actually bullish on it. So speaking of AI, we have this new report about where Mark Zuckerberg is spending his time. And I guess this is one of the least surprising reports on Meta I've ever read because it was obvious it was heading there. But there's a report in CNBC that Meta's top executives, including CEO Mark Zuckerberg, Chief Product Officer Chris Cox, and Chief Technology Officer Andrew Bosworth are spending most of their time working on artificial intelligence, according to an interview Bosworth gave at Nikkei Asia on Wednesday.

11:23Zuckerberg, it seems like five minutes ago, was spending all of his time on Metaverse. Now he's spending all of his time on AI. What do you think about that shift? What do you think it says about the direction of that company? Yeah, we've been saying this for a while now. I remember I was actually on CNBC in December talking about this exact topic about whether meta was going to move away from the metaverse. And it was the day that Boz, the Facebook CTO, sorry, the meta CTO had written a post saying they were committing to the metaverse. that, you know, like basically trying to justify the whatever$13.2 billion that they're losing per year to say, no, we're actually going to go even bigger into it.

12:03This is the future. It's funny because in that interview, they really try to wiggle out of it by saying that the work on AI that they do now can then, of course, be used in the long term for generating metaverse worlds and generative AI can be used for, you know, the metaverse, whatever it's going to end up being. But yeah, I think, of course, it's the year of efficiency over there. They got to go all in on AI. Do you think, are we going to see Zuckerberg wearing an Oculus helmet in 2023? or is he going to show up in Horizon's world anytime publicly in 2023 again? Or is it just AI now? Yes, I think he's still going to.

12:53I think I'm growing more and more of a believer in this metaverse idea, although I do think it's going to take some more time for it to go to fruition. So I don't see them giving up necessarily quickly. However, I think that the AI is going to be baked. It's not just going to be baked into metaverse. It's going to be baked into the Facebook product when it comes to content creation, right? Giving people the opportunity to create images, for instance, is like almost the perfect way for Facebook to get posting back up and people engaged again. So I think that's coming. Using AI to enhance photos and videos.

13:26I think that's coming. And there's been a report that I read that I really believe is going to be true. that they are going to use AI for advertising where an advertiser submits one piece of creative for instance one ad and Then you hit a button and generative AI for meta makes a thousand different possibilities of that ad Then you hit you know play and it starts running that ad across the network right now You're running it to billions of people out of the thousands piece thousands of pieces of creative You find the one or two that resonate the most and advertisers then get a minute get the opportunity to double down on that It's what the most sophisticated Facebook ad shops have been doing up to this point.

14:05And it's what they are going to enable through their tools. And it's just another data point that shows that they have a chance of circumnavigating some of these Apple changes. I'm going to run out and buy Metastalk right now from that because you just sold me on it. I mean, I think the two, you're right, the two exact points, the increasing the amount of posting that the average user does, which is, I mean, tremendously gone down and giving them easier ways to create great content. You're right. This actually does open the door on that. And yeah, on the advertising side, I've thought about that a lot.

14:43and if they can actually bake those type of tools in because i mean it's actually a perfect scenario for them because if they both allow you to generate those thousand different ad variations and it's their same algorithm that's actually targeting those ads and seeing what's working and it could even reallocate spend to the ones that are working as you said that's exactly how sophisticated ad shops work right now it's done manually at least on the content creation side The algorithms already know how to do that to reallocate spending to winners. So if it's all automated, that can definitely work.

15:17But my one piece of pushback here would be on the content creation side, because they've gone down this TikTok route of only the best content wins and basically because it's not follower based anymore and now it's much more driven by just pure engagement, it actually is even more of a disincentive for anyone to post unless you're like a professional content creator. So if you have a nice tool and you can create your cool generative AI post, still no one's going to see it because it's going to be pretty average because it's generative and everyone is able to do it. On one side, it helps them solve the problem, but because at a higher level, they've totally changed the way the platform works, I still think it's not going to do it for them.

16:10Yeah. So I would definitely be more skeptical on the creation side than the ad side, but the ad side, I mean, it builds a world-class advertising platform and Meta is one of the few companies that can do it because of the size of their audience. So that's something I think is, is worth looking out for and would not be shocked if we, if we see some big announcements coming up on that front in the next couple months okay speaking of AI baking it into the main products finally it looks like alphabet Google is going to add chat to the Google search experience so this is from the Wall Street Journal a conversation with Google CEO Sundar Pichai it says Google plans to add conversational artificial intelligence features to its flagship search engine and Pichai is saying this as it deals with pressure from chatbots such as chat gpt and wider business issues what do you think about the fact that google is going to start adding this um functionality into search itself because that's a big step right for bing who cares right add the chatbot by the way that bing chatbot is really good and getting better i'm just going to say it here but for google it's a very different situation okay okay so i just got back from two weeks of traveling still i use trip advisor i use like what search through reviews i i literally it felt like it was still 2010 traveling around like in the way i was actually trying to find stuff to do trying to make an agenda and yes like i played a lot with the bing chatbot and i got back some very reasonable agendas and suggestions and stuff but i didn't actually stick with it i wasn't actually going to take the suggestion to actually, you know, put my whole family in a car and go somewhere because Bing told me I still kind of, and you know, maybe that's a behavioral thing at the moment.

17:58But I think this is where when Google says it will include AI in chat, these announcements, especially from Google's perspective, are getting a bit tiresome for me, because like, I want to see it. I want to see it in action. Already, they, you know, they released this incredible demo. Did you see the uh google docs workspace generative demo where it's like yes it's like people are on a meeting it it instantly and magically extracts the insights from those meetings creates a doc and then creates a deck and goes basically from like meeting to deck and presentation and action plan without you having to do anything good luck like i mean honestly like when i see when when that happens in my day-to-day life i will be the happiest person in the world but until i see it happen remember duplex and restaurant reservations were supposed to be automated and that was from 2018 we were promised so so i don't know i think microsoft's brilliance in this and open ai even more so is being product release first and announcement second and i think that's where google is still kind of with the old playbook of fancy demo and announcement first and then waiting on the product release i think it's going to be really problematic in this battle right now oh yeah and it's uh it's interesting to watch sundar make this little press to our right he went on hard fork he talked to the journal he doesn't talk he does not talk to the media this is extremely rare for him to give two interviews in the span of a week not around a major developer conference satya andela did say that he wants to see sundar or wants to see google dance excuse me and this is them dancing and it's just a sign that like there is something going on inside that company that is um they're scared i think they're playing from a place of fear and we'll be it'll be interesting to see what that pushes them to do on the product perspective because yeah you're just going to see more features no i and i and i agree google should be a player in this a big player if not the player again they they are years ahead on this or they should be as you know on par with an open ai they should be on par like and where you know like you had jan lacuna on the show talking about how facebook does have the models like google should definitely have them they have the products they have the user base i mean i'm a google workspace user and i think it's a much better overall product suite than Microsoft Office.

20:27But they need to start getting this stuff out. They need to start really pushing this and showing that they have the products to match up the technology. In both interviews, there have just been clear indications that Google got caught by surprise here. I mean, just this quote from the journal story where Sundar says, we were iterating to ship something and maybe timelines changed given the moment in the industry it says everything you need to know maybe timelines changed what are you not sure if your timeline changed or not you're the ceo of a company this is your biggest challenge right now and you're going to say maybe timelines changed when they asked him in hard fork about whether he was caught by surprise by open ai's bot instead of just saying no we've been developing this technology we know what it can do he said they built a good product it's just an admission that they got beat which is which is a very interesting very astonishing well but but also i think one thing i really want to push on um in general is this kind of narrative that google was caught behind because they're trying to be safer and like that it's almost like the excuse that companies are using right now is an almost i mean i hate to say like trying to attribute it to wokeness slash like uh worries about being you know better on the diversity front or being safer in terms of like how the ai works i don't think that's the case i think yes like uh microsoft but even like microsoft remember tay yes that was a disaster but they still were aggressive on this i don't think it's wrong to be aggressive in pushing out new products that are using older models that are already tried and tested like yes gpt4 slash whatever other gpts i missed in the last two weeks um i think like releasing products on stuff like that it's worth you know making sure things are at least working properly but but i think it's a bit of a cop-out the way they've been trying to say oh well you know we're trying to be as safe as possible because otherwise everyone's going to get mad at us.

22:32I think that's definitely been a way they've been trying to escape the blame for just being behind on this. Yeah, absolutely. And from every indication that I had, Lambda, the internal Google chatbot was better than chat GPT in the original stage. But I don't know. We'll see. Maybe OpenAI and Microsoft have better technology using GPT-4 than whatever Google's working on internally. Hopefully we'll see soon. yep so there's another interesting thing that's that started to emerge in my discussion earlier this week with joe marchese and mike mcnano and also in the article that i just wrote in big technology that came out today on friday where it seemed clear that chat gpt and bing chat and bard were all demos trying to get others to build on top of the technology that exists underneath them right so effectively if you like using chat GPT why don't you build your own bot or something that you can use for our for your company on our tech that to me seemed like the selling point now very interesting thing has happened these demos work there's obviously a lot of interest in the API's but alongside it they've become destinations themselves even surprising the companies putting this stuff out now what they're doing is kind of opening them up we We talked about plugins like a couple weeks ago or earlier this week where you can now develop your own application within something like a chat GPT and maybe eventually within Bing.

24:00And it does make me wonder whether this is going to scare developers away from working with these companies or whether eventually, you know, they won't have any choice. and the AI can just subsume their experience. Because if you're like, for instance, a open table or chatting inside and chat GPT with the user or an Instacart, do you eventually lose your functionality where the AI starts to take over and ends up placing that order at the restaurant for people? So I'm curious what you think about it. I mean, the headline that I had was, as AI gets better, could it swallow the company's building with it?

24:36What's your take? No, I think this is a really interesting question. I know in that podcast, you all used the example of Kayak, that Kayak builds a plugin today, but Kayak is kind of the middleman layer between the customer and the airline. So why couldn't Delta and the few other airlines all just go directly into ChatGPT and feed their data? I think it's a really interesting space. Like, I actually do think in the airline ticketing space, it is interesting because there's only a few players that really need to connect their data to OpenAI or whatever other platform. And they could create the service and literally displace Kayak on day, I mean, in a very, very short timeframe.

25:21But I do think this also gets back to the big debate around, like, where is the value? Is it in the original data set and those relationships? Like OpenTable has relationships with every single restaurant or, I mean, you know, some large percentage of them. Is that where the value is or is in the value, the actual ability to integrate, aggregate and, you know, distribute that data in the form of reservations? And I think there's an argument to be made that it's actually those relationships with the restaurants. OpenTable has above others. Resi has with certain restaurants. So the idea that ChatGPT, right now ChatGPT looks cool because it's able to take the data that OpenAI has already aggregated and serve it to you in a more quick, concise format.

26:11I do think it's, I still think there's a lot of value in those original data sets, especially if companies are more protective around them and are able to use them in smarter ways. Yeah, so obviously this is not going to happen overnight. There's relationships like you mentioned there are barriers to getting the data But there's a lot of for instance enterprise technologies like task management for instance That once the AI AI gets smart enough and if people are chatting with it It can be like hey by the way you have this to do to you know today or good morning Here are your tasks have you completed this or even you know Here is some help with this or even as the facts on the ground have changed in this project I am updating your presentation that's going to be ready for tomorrow.

26:55I mean, this is, of course, a number of years out. And, you know, maybe it's some fan fiction about AI. But given what we've seen up to this point, I don't think it's too fanciful to think that this is possible. Yeah, you had brought this up. And still, I think, tweet of the year is still going to Benedict Evans for his tweet about that, you know, maybe it's not an assistant to a spreadsheet. maybe the point of AI and generative AI is that we don't need spreadsheets at all. And it's going to change the actual interface and the way we interact with data in businesses. And I think that to me is the most kind of fun, exciting part of this is, is, is it going to be chat the way we try to like book travel?

27:37And does that displace Kayak? Or is the entire way we even approach booking travel or like coming up with agendas? Is that going to change? And whether or not it's a chat interface i think that's so that's the fun stuff and i can totally see why uh especially in the vc world while why thinking about this in terms of like completely new industries and massive companies completely disintermediating the existing you know dominant players it's got to be a fun time maybe that's why everything is taken over with such uh intensity so ranjan you've been away for a couple weeks probably had some time to think deeply about this quick temperature check we're in the middle of a hype cycle around ai there's no doubt about that where what is your feeling in terms of this technologies like especially generative ai what is your gut feeling in terms of its potential to disrupt the state of business today yeah this this is the tough part is having used this a good amount and actually worked with this i think it has tremendous potential um My worry is, is the hype cycle is like, I mean, and this is what I've written about voice computing and Alexa is that can the hype cycle actually prevent a technology from ever realizing its potential?

28:54And I know, I mean, it almost sounds ridiculous in some ways, but imagine like right now I am told that I'm going to be like, be able to book my travel in a completely new and innovative way. and then I go try to do it and everything looks the same. And even six months from now, a year from now, is stuff really different? But it's every bit of disappointment someone feels based on that the hype that was presented to them is going to chip away at the longer term potential because people will potentially stop working on solutions. Honestly, I mean, I still don't know. Could crypto have been integrated into our daily lives by now?

29:31Could Web3 have been a real thing? like sure i mean payments still suck in the us why hasn't that been fixed like was it the hype cycle around crypto that actually killed its potential its potential right so i just got a a comment on my newsletter today talking about like what's going to happen when people get fatigued with chatbots and i think that's a totally reasonable question especially given that we've had already a wave of chatbot hype that didn't turn out to be anything. And this time, though, I think is different. I don't think that we're going to see the fatigue because these things are so good.

30:07That's just my take on it. Are you using any kind of like chatbot apps or anything outside of kind of opening Bing? No, I'm a Bing boy. You're a Bing guy? Bing boy? Yeah. All right. I'm a bard kind of guy, I think. I actually haven't used bard. Then you can't be a bard guy. It's against the rules. Okay, let's take a break. We're here with Aranjan Roy. He is the author of Margins, which you can find at readmargins.com. It's not blocked on Twitter. We're going to talk about that and more in the second half. Capital One's tech team isn't just talking about multi-agentic AI. They already deployed one.

30:51It's called Chat Concierge. And it's simplifying car shopping. Using self-reflection and layered reasoning with live API checks, it doesn't just help buyers find a car they love. It helps schedule a test drive, get pre-approved for financing, and estimate trade in value. Advanced, intuitive, and deployed. That's how they stack. That's technology at Capital One. These days, it feels like every dollar should be working a little harder. But figuring out where to put your cash can be confusing. That's where Wealthfront comes in. Wealthfront is a tech-driven financial platform built to help you grow your savings into long-term wealth.

31:29Their high-yield cash account through program banks offers a 3.5 % APY on your uninvested cash as of November 7, 2025. And there are no monthly fees, no minimum or maximum balance to earn that rate, and you can even make free instant withdrawals to eligible accounts in just minutes, 24-7. So your money can always be within reach. Right now, Wealthfront is offering new clients an extra 0.65 % APY over the base rate for three months on up to a$150 ,000 balance. That's a total of 4.15 % variable APY when you open your first cash account. Go to wealthfront.com slash big tech to sign up today. This is a paid testimonial for Wealthfront.

32:07It may not reflect the experience of others and there's no guarantee of future performance or success. Wealthfront brokerage is not a bank. Rate is subject to change. Promo terms and conditions apply. For more information, see the episode description. And we're back here on Big Technology Podcast live on LinkedIn and YouTube, taking your questions if you're on the stream and also discussing the week's news. So another number that just came out today is the jobs report. The U.S. economy added 236 ,000 jobs. Sounds like a lot, but it also shows that we've had some cooling growth in jobs and the market is kind of, I don't know, wobbling a little bit.

32:41We're expecting weaker economic growth. We've been talking about a recession for like, what, seven years at this point. And, you know, we're still not in that territory. But it does make you wonder whether maybe it's coming now. What's your read on it, Ronjan? Yeah, I think the thing everyone should look at listening over this weekend and going into Monday is this report is this should have been a very bullish report in the sense that job growth is slowing, growth is still relatively stable, meaning that it's actually almost the perfect soft landing scenario for the Federal Reserve. The idea that, OK, we're not going to overheat because the job market stays overly hot.

33:20We're getting down to a more manageable point. But if you look at the activity on yields today and over the last few days, the two year has jumped above 4 percent again, which is where it was before suddenly a Silicon Valley bank happened and everyone thought, OK, the Fed might even cut if not stop hiking rates um the 10 years started moving like basically over the last week yields have completely been retracing towards the levels that they were before uh all the silicon valley bank worries started to resurface and i think that's saying that there's still worry that the fed is going to need to hike inflation is still a threat but also there's a lot more worries about slowing growth, which again, it's like the worst of both worlds.

34:08This should have been a bullish thing. And I think we'll see on Monday, the stock market, whether it actually responds negatively to this. Yeah. And we'll have plenty more to talk about next week. Jim McKelvey, who is the co-founder of Square. He also is the founder of Invisibly. And he's also on the St. Louis Fed. He's going to come join us next week. And we're going to talk deeply about this economic data. It's going to be super fun. So companies now are adjusting to the current environment. And this brings us to a segment, which I call Google, where's my stapler? Google has pulled back perks across the board.