SVB Reexamined, GPT-4's Potential, TikTok's Potential Ban

Ranjan Roy of Margins is back for our weekly discussion of the week's tech news. We cover: 1) SVB's depositors' role in the collapse 2) The over-financialization of tech 3) The continuing risk to the rest of the banking system 4) The re-rerating of late-stage startup valuations 5) Stripe's down round 6) The release of GPT-4 7) GPT-4 in Microsoft Office 8) How AI might eliminate modern apps vs. be a feature 8) TikTok's potential sale or ban 9) How AI might help you date. -- Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice. For weekly updates on the show, sign up for the pod newsletter on LinkedIn: https://www.linkedin.com/newsletters/6901970121829801984/ 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?

  This episode is brought to you by Genesis. What does it actually take to put Agentic 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 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?

• 0:43Alex Kantrowitz?

  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/alex for free consultation. That's akash.network/alex for free consultation.

• 1:17Alex Kantrowitz?

  Welcome to the Big Technology Podcast Friday Edition, where we break down the week's news in our typical cool-headed and nuanced manner. Joining us as always is Ron John Roy. We have so much to talk about. We're going to spend the next little bit talking about the latest on Silicon Valley Bank. But not only that, what's going, what's going to happen with startups that are now going to have to rerate their valuations as maybe the entire tech ecosystem gets repriced? We're also going to bring you the latest on GPT-4, what it means, what you should be excited about, what's all hype. And then, of course, we're going to cover the ban of TikTok or the potential ban of TikTok. I shouldn't get ahead of ourselves and the latest at Meta. Ron John, welcome to the show. Hello. Great to see you. We had a pretty good time at South by, I would say.

• 2:02Speaker 01

  Yep, back in cold New York. Missing Austin right now.

• 2:06Alex Kantrowitz?

  Yeah, me too. And for those listening, Rajan was scoping out the room that we were scheduled to talk in and came back with reports of like a dozen people, 10 people. And then we showed up and thankfully, it was a packed house. And to anyone that came out, we had big technology listeners there. So to anyone who came out there with us, thank you very much. And I would say it was a pretty exciting session for us.

• 2:30Speaker 01

  It was a good time.

• 2:32Alex Kantrowitz?

  It was a good time. Then the news just kept on spilling over and spilling over. And Rajan, you have a new piece about the continued fallout of Silicon Valley Bank. So we should definitely talk about that. And by the way, we're doing this again on LinkedIn. So if you're watching on the live stream and you have questions, we're happy to take them. For those on the feed, we do these Friday shows, they're news shows that cover the week's news on Fridays and the main flagship interviews on Wednesday. We take your questions. We're typically going to do these on on Fridays at 11 a.m. Pacific and 2 p.m. Eastern time. Today, we're a little bit early. Rajan, I think, has to go go to Europe, or I'm going to make that up at least. So but Rajan, let's talk a little bit about

• 3:12Alex Kantrowitz?

  what's going on with the continued fallout of Silicon Valley Bank. You had a piece in margins that was very interesting, talking about basically busting a lot of the narratives, I think, that we had around this, or questioning some of the narratives.

• 3:24Speaker 01

  Yeah, well, the biggest point that I've been I was trying to make was as we were in Austin last weekend and, you know, the world was collapsing and VCs, some of the famous ones were crying on Twitter loudly. One, one point that everyone kept making is let's wipe out, you know, the shareholders, let's wipe out management, fire them all. But we have to save the depositors. The depositors are mom and pop stores in Ohio or whatever else. And I take, take issue with that. Because what I was trying to point out is, so Silicon Valley Bank, and this is becoming more known, The New York Times even had a piece today about

• 4:07Speaker 01

  all the benefits you would get as a depositor. And the kind of interconnected nature of the way all of this worked, it's amazing. It's Silicon Valley Bank. Would take deposits from venture capitalists and their startups. The venture capitalists would not force, but highly recommend their startups to bank there. Then Silicon Valley Bank would turn around. They would invest in those same VC funds. They would invest sometimes in the startups who are depositing with them. And they would give loans of all sorts. They would give personal loans. They would give mortgages to founders who couldn't get them anywhere else. They would give highly preferential rates on those mortgages.

• 4:49Speaker 01

  They, uh, the whole field of venture debt, which is something that exploded Silicon Valley Bank. It was the, it was a very large part of their overall loan portfolio. And venture debt is a really interesting part of where we are in the cycle. Because it allows startups to not have to go out and raise more equity with the down round. But instead take some kind of debt or bridge loan in order to maintain a high inflated valuation. So venture debt was a huge part of keeping everything afloat. And then even something that's pretty innocuous, but capital call lines of credit. When a VC makes an investment, instead of having to actually have the cash from the investor directly.

• 5:30Speaker 01

  They could draw down a line of credit from SVB, make the investment and then over weeks or months, get that money in from the investor. So the whole world of super rapid fire investing that Tiger Global kind of perfected. They were a huge part of that cog or, you know, like, so you take all these different elements into account. And they were such a central part of the COVID ZIRP tech boom. I said that, you know, the way Robin Hood was for meme stocks, they were the analogous party for the entire boom.

• 6:03Alex Kantrowitz?

  But don't you think that the boom would have found its way elsewhere? Like the zero interest rates were a fact on the ground, right? This was just the bank that facilitated it. And to push back on your point a little bit. I mean, shouldn't a bank trying to win over a set of customers do what it can to ensure those customers bank with it?

• 6:20Speaker 01

  Yeah, no, no, I made the point. It's a great business as a bank. Like, it's probably the best, uh, best review you could ever give for a bank is that they enabled wealth creation, you know, at an unimaginable scale for a huge concentration of its depositors. But that's also why depositors are not innocent in this. The depositors, I mean, at a few levels. One, they received incredible levels of wealth creation over the last few years or the last decade, thanks to all of these different products from SVB. But even more important, it was those deposits that killed SVB. It was the massive increase in deposits over the last two years. They over, they tripled their deposits, whereas the rest of the banking industry only grew deposits

• 7:03Speaker 01

  by 37%. So it was the very deposits that forced them to move down the risk curve and buy longer dated treasuries. And then it was those same depositors that sparked the bank run to pull out all the money and actually caused the collapse of Silicon Valley Bank. So again, depositors played an incredibly important role. So the idea that simply, you know, wiping out the fat cat shareholders and firing the management actually, you know, prevents any kind of moral hazard in this, I think is wrong.

• 7:35Alex Kantrowitz?

  Okay, but you're singing a very different tune than last week, where you said one of the fundamental responsibilities of a bank is the risk management stuff. So yes, the depositors put a lot of money in this bank, no doubt about it. However, you still have Silicon Valley Bank failing in its one obligation. So yes, I definitely understand that, okay, the depositors put the money in there and that was part of the thing that fueled the problem. And you have this, you know, great line saying that like the same depositors who got disproportionately rich thanks to Silicon Valley Bank and whose wealth was too much for Silicon Valley Bank to handle, ended up bringing down the bank. So harming the bank that enabled what they were doing. I think you've highlighted it here before. But again, like, it's the bank, right?

• 8:19Alex Kantrowitz?

  Didn't we say that was the bank's issue?

• 8:22Speaker 01

  Oh, my, my view on this has completely changed over the last week since we recorded last Friday. And, you know, sometimes views can change and that's a good thing. I think it was, I had heard like murmurs about these preferential mortgages from friends in the startup community, but I didn't quite realize the extent of just how integral Silicon Valley Bank was in enabling all of this. And I think that's where, again, you know, you saw all the kind of hyperbolic tweeting and stuff, but at first you're still thinking, okay, well, you know, are people going to make payroll, whatever else? But again, so much wealth was created. And I think the other point, by Sunday, the more I'd read, the more I'd spoken with friends

• 9:06Speaker 01

  from the banking community, it was very clear, the FDIC had signaled it themselves, you know, and simple math, that Silicon Valley Bank, majority of their, you know, assets were in very liquid things like US Treasuries and, you know, like liquid mortgage-backed securities. All that could be sold off in a matter of hours and everyone would have gotten back 50, 60% of their money by Monday, at worst. And then over time, maybe you take a 5% to 10% haircut. There's been a lot, even banks are issuing analysis, a lot of, again, like very reasonable analyses, why at worst you take a 5% to 10% haircut? So why was there such an outcry?

• 9:48Speaker 01

  And why were we threatened with the, you know, like collapse of the entire banking system when everyone was gonna make payroll? Everyone, I know the idea that they wouldn't was spurred on by, again, these same depositor investors.

• 10:02Alex Kantrowitz?

  There is a great clip in your post of Jason Calacanis, who was one of the biggest alarm sounders in this entire moment, talking about how SVB sent like eight people to his house to help do a mortgage. And, you know, it's kind of interesting to talk about this close relationship between these people who are getting something at SAP that they, they weren't going to get at any other bank and now cried to save it. And the reason why we're talking about this, right? We're not trying to cast blame. What we're doing is saying, we're analyzing this because we're asking if the bailout or the rescue, depending on the term that you like to use of this bank, is a moral hazard, where...

• 10:43Alex Kantrowitz?

  If we rescue the bank in this situation, are there going to be other, you know, depositors in the future that will not be careful about their their deposits? But, you know, Ron, and I, you know, I do have to disagree with you on this. Like, I don't think this is a huge moral hazard because, again, you know, this is, we can't, we cannot ask depositors to check into like the the balance sheets of every bank that they that they bank with, no matter how many benefits they get, that just seems arduous to put on the business owner. And then, like, you know, okay, so the bank might not have failed necessarily, but we were, you know, once the panic gets going, that's something that can spread across the

• 11:28Alex Kantrowitz?

  entire US banking system. And it has become a political issue. Janet Yellen was out in front of the Senate, I think, today, you know, getting grilled about whether she bail out other regional banks, which is an important thing. But the fact that they did bail out this bank means that they're not going to be faced necessarily with the prospect of bailing out the smaller regional banks across the country. Whereas if they didn't, there's a pretty good chance that that's what we would be talking about now is the failure of those regional banks.

• 11:54Speaker 01

  But i- is it? Is there a pretty good chance? I mean, this is where, yes, smaller regional banks don't have the same balance sheet, don't have the same risk profiles as Silicon Valley Bank. You know, again, their deposits, the overall banking system grew 37% over the last three years, where Silicon Valley Bank tripled. They didn't have all of these additional products that they're offering, you know, inflating overall wealth creation around this. It's just, it's, it's a completely different business. And again, if on Monday, I was talking to a friend who's who works as a VC, Sunday afternoon about this, and he was like, oh, no, if the FDIC really comes out on Monday and says

• 12:37Speaker 01

  everyone gets 50% of their money back on Monday, can make payroll, this all goes away. Like, we didn't even get there. It, everything ended up being complete panic, and this is the most biggest systemic problem imaginable, before we even had any inkling of what the resolution could be. And I think that the last point is, we're not out of the woods, and this didn't solve things. I mean, look at First Republic today. I think after getting the $30 billion rescue plan from all the big banks, what are they now? Um, I think it was down, oh yeah, it's down 26% again today. So, so the, the fear in no way is gone, but all that happened was this very, very concentrated, specific group of depositors

• 13:21Speaker 01

  was bailed out. And I will use that word.

• 13:25Alex Kantrowitz?

  Okay. So Twitter was, was definitely not been, been helpful for this, uh, in terms of the, the panic. And I'll say that more, that much. But one of the interesting things that I've also seen is that I, that I agree with in your post is that it's been wild to watch some of the investors say that they're speaking up for small businesses and they're speaking up for the little person. We're like, basically, you know, two tweets away, they're cheering Meta's layoff of 20,000 people. Now either you're, you're, it's, it's a little bit different. Of course, it's a business decision, you know, versus uh, a bank falling, you know, falling apart and then not being able to pay people. But if there was this much concern for the little person, why is it such a disjointed response to the idea that people could lose their job or not get paid?

• 14:09Speaker 01

  Yeah, this is the biggest part. This is the tell in this entire saga. When the same investors who have been cheering Mark Zuckerberg's year of efficiency and layoffs of 10,000 people and Elon, you know, like slashing Twitter's workforce for the last few months, to turn around and say they care about the everyday person who's gotta make payroll and feed their kids or whatever else. I, I just don't buy it. And again, it's amazing because some of these people have turned right around and went back to, if you saw Mark Zuckerberg's memo about announcing another $10,000, uh, 10,000 people laying, uh, being laid off.

• 14:50Speaker 01

  They're celebrating it again because as far as they're concerned, everything's okay is from their perspective.

• 14:56Alex Kantrowitz?

  So Ron, John, I'm, I'm hearing you and I'm also, now I'm starting to think, wait a second. Would you have done the, the bailout, the rescue, whatever you want to call it? I, I am in the camp that it doesn't really matter what we call it. But are you in the camp that we should not have done this rescue to the depositors?

• 15:12Speaker 01

  I think the FDIC should have, if they were very ready to return 50 to 60% of everyone's money on Monday, I think that actually would have been a very reasonable course of action. And again, we're seeing it. This has not solved the problem. This is not in any way solved with First Republic down, everyone's still panicking. It didn't solve the problem. It just helped a very, very specific set of people.

• 15:40Alex Kantrowitz?

  But think about, okay, so the argument that it did stop the panic is think about how much worse it could be today. But I guess these are all hypotheticals. But talking about how this could also spread to a larger, the larger banking system, there was this chart, actually, that David Sacks tweeted out, I think Balaji also tweeted out, and is worth talking about, how right now, we're in this moment where what happened to SVB is it failed because it had all these unrealized losses on its balance sheet, holding these long-term, long-duration securities, and people figured it out. But over the past few years, we have hundreds of billions of dollars, maybe trillions of dollars of unrealized losses being held across the entire banking

• 16:21Alex Kantrowitz?

  system. And for years, it was not so bad. But once you hit 2022 or so, when these interest rates go up, you start to see this across the banking system. And the argument is that this is a huge systemic failure. And from Sacks, what he says, I'll put it up on screen here for those on the live stream, the problem in our banking system, the problems in our banking system aren't over, they're just getting started. And now, at least, we have some time to deal with them. So why don't we seize upon this, Rajan? I mean, what is the long-term threat here in terms of more banks being hit by these long-duration securities issues, the same way SVB was?

• 17:02Speaker 01

  Yeah, so, again, I think that statement, the problems in our banking system aren't over, they're just getting started, to make that after being made whole on the Silicon Valley Bank deposits, is a perfect indication that that was just self-serving. You know, like, if all of these problems were truly out there and being recognized as something that needed to be fixed, the calls would have been for, let's regulate the hell out of everyone today, let's have like an emergency congressional, you know, intervention to actually nationalize banks, or whatever it is. And the other thing to remember, that graph that's being shared around, The other banks, their unrealized losses aren't greater than their overall equity

• 17:47Speaker 01

  value, the way it was for Silicon Valley Bank. And that's what made them much more vulnerable in this versus a JP Morgan or a Bank of America or whoever else. And I think that's the, it's misleading in that. Is it a problem? Sure. And is it something that, you know, needs to be thought through? Yes. But also, the banks, this is, and this goes back to the overall regulatory issue where banks over $250 billion in deposits had been already going through interest rate risk stress tests, whereas Silicon Valley Bank had lobbied and been removed from that exempt, had been exempted from that because they weren't over 250 billion. So at least there's some work being done, or there has

• 18:28Speaker 01

  been around larger banks on this. But again, if any bank loses, you know, over, what is it, a half of their deposits or a third of their deposits in 24 hours? Yes, everyone's going to be in trouble.

• 18:40Alex Kantrowitz?

  Yep. We have a nice little bit of pushback in the comments that I think I'm going to read from Ryan Smith here, talking about our comparison of the celebration of layoffs and the panic about startups not being able to make payroll. And he says, you guys are comparing well-established companies with startups. The well-established players will be fine. This is just part of doing business like layoffs. The venture cap-based companies might not even get out of the gate. I think that's a good point when we talk about, again, a good bit of nuance when we talk about how these companies are, are the reacted, the sort of polarly different reactions to Facebook layoffs versus startup layoffs.

• 19:19Speaker 01

  Yeah, no, I think that's a great point. Because, again, separating out, you know, venture-backed companies, especially late-stage ones from the big tech firms, it is a very important nuance. But I think, I guess, from my side, I was just making the point that the way they approached the idea that we have to worry about people getting laid off, that was the inconsistency.

• 19:42Alex Kantrowitz?

  Yep. One of the things that I talked about looking at, I mean, we had this conversation on the podcast on Wednesday with Omalik and Chris Tollis about why this became politically tricky for people. And even President Biden hesitated before going through with it. This was why we waited all weekend to figure out what was going on. My, my point to them, and I think I could have made it more forcefully, and I wrote about it in Big Technology this week, was that there has been this over-financialization of the tech industry where tech can do a lot of good. And we love using the products, again, like we're talking on LinkedIn here through StreamYard, and we're going to put it out on Spotify and Apple and Overcast. Like, it's a miracle of technology that even allows us to have this conversation at scale,

• 20:27Alex Kantrowitz?

  where the tech industry sort of lost the public, or a chunk of the public, is this over-financialization that it did, where instead of just being able to press a button and get a delivery and tip a DoorDash driver, it actually included your tip for, DoorDash actually included your tip for, as part of its minimum payment to drivers. Until someone pointed it out in the press and it was no longer tenable for them. And even then, they tried to fight it, saying that this is what drivers or what dashers prefer. And I do wonder if, if this is a moment where the tech industry says, okay, maybe some of this over-financialization is, is a little bit counterproductive, because there is a cost to it on, on the political side. And that's what we're starting to see.

• 21:12Alex Kantrowitz?

  I'm curious what you think, and I also know you have some feedback about that show, so why don't you hit me with it?

• 21:17Speaker 01

  Well, no, I, I think, and if we dig into over-financialization, I think that's like the most important point of this, is, and you know, listening to the show, Om, Om was definitely a beneficiary of this system. And so to people who weren't, it sounds very different, where you're hearing all these, this special treatment, preferential treatment. You know, it's, it's great for that person. And I think like, the, the, in terms of over-financialization, if we get into, let's say, late-stage valuations, and I think this is a really important thing to watch right now, is how this all plays out. Think about what I'd said earlier, is like Tiger and other firms were able to invest at speed because they had these capital call lines of credit.

• 22:03Speaker 01

  And in, a founder who can borrow and take out a personal loan or a mortgage against a valuation of their company is going to have a huge incentive to go for that inflated valuation. Like, all of these things worked together to create All these massive valuations, and this was one of the biggest problems. But for a while, again, it all worked because then the company could IPO. They can pass it off to the retail investor and everyone else walks away happy who was in on that entire chain of events. And I think what's going to be really interesting to watch right now is what this does to late-stage private valuations. Because the, I mean, whatever money was already slowing down, it's going to become significantly slower.

• 22:46Speaker 01

  I do think that's a good thing because I think that is probably the most distorted portion of the market. Like you just had Tiger, the stuff just came out. I think they were down 33%. They marked their private valuation, their private books down. Whereas the comparables in the public market are down 50 to 60%. Similarly with SoftBank, all of these funds have been able to maintain the illusion of a certain valuation or a certain, you know, like worth of their portfolios because of these inflated valuations. So that's at the high level, but then when you go down, it's still crazy to remember that that founder of that company could then use that to borrow against it for their own personal gain.

• 23:28Speaker 01

  And I think like to see how that part of the market plays out is definitely going to be the most interesting thing to watch over the next few months, aside from a potential global banking crisis.

• 23:39Alex Kantrowitz?

  Yeah, fingers crossed that doesn't happen. Thank goodness we bailed out these companies. We won't have that. Okay, sorry. So let's talk about it. Did this, is this rerating or this revaluation of big tech VC funds or big VC funds and of startup valuations? Is that something that's a direct outcrop of SVB? Or is that just something that like now we've had these rate raises and all of a sudden all big funds are going to have to, you know, take a look at the valuations of startups, take a look at the valuation of their portfolio and then adjust?

• 24:12Speaker 01

  Yeah, I am so curious. One, SVB invested in a lot of these companies alongside the same VCs that were their depositors. But then, when their books are sold off, is there a forced repricing of these assets? Because if you're the buyer of an essentially bankrupt fund, or distressed fund, you're not going to pay at that high valuation. So I'm very intrigued to see how that plays out. But another thing that there was in this Wall Street Journal piece around Tiger Global, it's crazy to remember over the last few years, they would not only value companies at a very high level, then they would go out and raise money off

• 24:52Speaker 01

  of these. Like, they, there's a, their PIP16 fund had raised in 2020 and said it was generating a 20, 20, 22% internal rate of return. And then that's already been marked down to 9%. But they already used that to raise their funds after that. So like, again, that cycle of ever increasing money, everything was predicated on these high valuations. So I do think, again, that there will be pressure on this from many, many levels. And I think that's going to be where we see some very interesting shakeouts.

• 25:24Alex Kantrowitz?

  And we're already starting to see at least one sort of shakeout, which is Stripe announced a Series I fundraise this week.

• 25:31Speaker 01

  Series I?

• 25:32Alex Kantrowitz?

  I've never heard of Series I. I mean, we might just keep going down the entire alphabet. But, okay, so you read the high level numbers, that the fundraise is more than 6.5 billion at a $50 billion valuation. And you just say, oh my God, that's a ton of money. And how do they raise that much in this environment? And that valuation is obscene for a private company. But then you look at the valuation that they raised at last, which is 95 billion. And, you know, it's almost now, the company is almost now valued at half of what it was. Is this sort of the sign of now we're starting to see the sanity come into the late-stage private market? Is this already beginning?

• 26:15Speaker 01

  Absolutely not. Because...

• 26:18Speaker 01

  I would say the most amazing part of the Stripe fundraiser is that a large portion of it went to actually paying the tax bill for employees restricted stock units that for them to issue new equity, they would have to, it has been so long that they would have to pay off this huge tax bill for employees. And the amazing part is, again, these massive funds are actually giving money to allow for that. It was $2.3 billion of the 6.5 billion were to cover withholding tax for employees. Like, it's just to prevent going into the public markets and having to open your books. They know it's so bad that they're unwilling

• 27:03Speaker 01

  to do that, that they're willing to literally just keep this going and going and going into Series I, rather than just recognizing where the business is and letting the larger public market say this is what it should be valued at.

• 27:18Alex Kantrowitz?

  What's the alternative? I mean, if they did say, if they did go to that, I mean, so they're already taking this haircut from a 95 billion valuation to 50 billion. What do you think, I mean, the market, if they send it to the market and said, okay, value us at this moment?

• 27:35Speaker 01

  Yeah, no, I mean, I think a firm, a firm's down what?

• 27:38Alex Kantrowitz?

  Can you even say we're a $10 billion company?

• 27:39Speaker 01

  85% add yen's down. Like, obviously, that's the thing. Public market comparables are all down significantly more. When we see, well, it was like a 45% haircut for Stripe, it's still not near where the public market comparables are. So that, so, okay, then you get into this whole, you know, the argument can be private markets are better because they're longer term focused, whereas public markets, you know, are everyday liquid quarterly earnings, all these other things. But that all sounded, I think, good in theory when you had one, when it's a company just at an earlier stage that's uncertain about its revenue and profitability paths. But then also, Series I, like, this is, how long has

• 28:23Speaker 01

  this gone on? And, and this is where, again, going back to SVB, once some of those products and services and relationships are no longer there, I can't see how this stuff keeps happening.

• 28:38Alex Kantrowitz?

  We're here on Big Technology Podcast Friday Edition. Ron John Roy is with us as always. A few bit of of housekeeping notes. First of all, for those of you who rated the show on Spotify or Apple Podcasts and let folks know that you're here and listening, those five star reviews really went a long way. We have two great interviews coming up next week and I'll reveal or the next couple of weeks and I'll reveal one of them right now. Kevin Systrom, founder of Instagram, and now he's the founder of this new news startup Artifact News is coming on next Wednesday for a discussion of what goes basically, you know, what goes on when we're moving to media world where we're all being recommended content by AI. I think you're going to love that discussion. So

• 29:19Alex Kantrowitz?

  thanks to those who rated and review. And if you haven't yet, if you can rate us five stars on Spotify and Apple or whichever one of those you listen to, that would really help the podcast. So thank you for that. Also, Ron John is going to be out on vacation the next two weeks. We're still going to be breaking down the news here. I know that Aaron Griffith of the New York Times will be joining me for one week and we'll see who's going to be here the next week. So we hope that you join us for both of those. On the other side of this break, we're going to talk about the release of GPT-4 and then maybe the potential for a TikTok ban here in the United States. We'll be back right after this. This episode is brought to you by App Point. 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.

• 30:03Alex Kantrowitz?

  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 App Point does. For 25 years, they've been the trusted layer beneath the world's most demanding data, now extended 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. App Point, 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.

• 30:54Alex Kantrowitz?

  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 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 wide-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 ironwall.com/bigtechnology, fill in the quick form, and request your free risk assessment. Their team will show you just how exposed your executives are

• 31:39Alex Kantrowitz?

  and how to lock it down before a threat reaches their front door. That's ironwall.com/bigtechnology. Learn how to stop online threats before they become real-world attacks. The link's in the show notes.

• 31:55Alex Kantrowitz?

  And we're back here on Big Technology Podcast, talking about what we talked about, Silicon Valley Bank. I think we've covered that exhaustively, and I'm glad we did. It's a massive story. But now we're going to talk a little bit about the latest in AI. Joined as always by Ron John Roy. You can get margins, and I'm going to get it right this time, at readmargins.com. Latest article is up there, so recommend you check it out. Okay, big news on the AI front is that GPT-4 is here. It's a step up from GPT-3.5. The way that OpenAI describes it, they say it's a large multimodal model that accepts image and text inputs and emits text outputs. And that while less capable than humans in many real-world scenarios,

• 32:38Alex Kantrowitz?

  it exhibits human-level performance on various professional and academic benchmarks. Basically smokes law students on the bar exam. It can crush your score probably in any test you took back in the day. And it also does some of this really amazing stuff, like take an image that someone might have drawn and can turn it into a basic website, which to me, watching that happen was like, again, like so many times recently is watching AI perform magic. Rajan, I'm curious like what you think of the development here. Is this incremental? Or is this really the step change that a lot of people are talking about?

• 33:12Speaker 01

  I think it's a step change and we've been talking about generative AI a lot over the last few weeks. It's something I've done a lot of work in over the last year and a half. And this is big. And I think that multimodal nature, the thing to remember about this, I had primarily been working with text, text to text, like input, output. There's a lot that's been done, DALL-E, Midjourney around taking text and turning it into image. Again, I want a ridiculous photo of an otter using a laptop on an airplane. I think I saw something that Midjourney has been improving over the years as a base case. But now what's going to happen is you can actually take video and image for the input

• 33:55Speaker 01

  and then turn it into text, you can, which is where you saw the idea that you could take like a napkin sketch. It'll understand it and then it'll actually turn it into code and then code a website for you. So I think this gets really interesting for real world applications. Everything in marketing, this was a big, like making a social caption using GPT-3 actually wasn't that interesting. Because again, it's, this stuff is so visually based that it really didn't make sense unless you had to like manually describe the entire photo. Now you can actually take again, video text input, then generate text, video or images. So I think this is big.

• 34:40Alex Kantrowitz?

  And they also said that part of it, I think in the New York Times, they said this is that GPT-4 has already been baked into some versions of Bing. And I mean, when you think about how good that chatbot has been, it's like, oh, okay. If that's the thing that I've been talking to, then yeah, this, this does seem like a real development. One of the places we're going to start to see this is in, kind of hilarious, but in Microsoft Office and a version of it, I suppose, in Google Drive, where you're going to be able to be in Excel.

• 35:07Speaker 01

  Google Workspace.

• 35:08Alex Kantrowitz?

  Okay, Google Workspace.

• 35:09Speaker 01

  Yeah.

• 35:10Alex Kantrowitz?

  Sorry.

• 35:10Speaker 01

  You know how they name things, yeah.

• 35:12Alex Kantrowitz?

  I know. I'm going to get an angry email from Google PR after this, but that's okay. We can handle that. But It looks pretty amazing where you can like basically drop in a bunch of numbers and then start, you know, conversing with the Excel spreadsheet, asking about different trends and, and where things are going from that, you know, in your business. What do you think that that's gimmick? Or I mean, it sounds to me like that's gonna be pretty useful, actually.

• 35:35Speaker 01

  All right, two points on this. One, God bless Microsoft and Google for being smart about this because, think about like Meta tried to sell us the Metaverse by being in like meetings, you know, like with a with no legs and the whole virtual meeting thing. Versus Google, their demo video is like, okay, you don't have to do slides anymore. You don't have to do spreadsheets anymore. So I think from a pure marketing perspective, both of them knocked it out of the park. I think like make like really hitting pain points for people. They did an incredible job. But the second part is, and trying to be balanced and nuanced here, I'm very excited about GPT-4.

• 36:18Speaker 01

  I was a little like Google, if you watch, they have this video where it takes an email thread from work, turn like extracts meaning summarizes it for you, turns it into a presentation all with like one line or something like that.

• 36:34Alex Kantrowitz?

  It's pretty incredible.

• 36:35Speaker 01

  I don't think we're anywhere near that. I until I see it in real life, until like I actually use it. I still feel like literally this morning I had to go on Google Drive and I had to look up how to do a date constraint search. And there's like a whole set of parameter. It was a pain in the ass like it's the current product is so far from what they have in the demo. That it makes me a little worried or suspect that we're getting back into overhyped territory. And it actually reminded me, do you remember the Google Duplex demo in 2018?

• 37:07Alex Kantrowitz?

  Yes.

• 37:08Speaker 01

  Where they had Sundar was on stage and they had a it was like someone was book an AI was booking a haircut appointment with a real person. And theoretically, this was a real phone call that was made and it was understanding the other person and conversing with them.

• 37:25Alex Kantrowitz?

  And hesitating and going um to mimic human speech.

• 37:28Speaker 01

  Yeah.

• 37:28Alex Kantrowitz?

  Can I get an appointment at a even though it didn't have to wasn't really thinking.

• 37:33Speaker 01

  Oh, Google trying to be human. Um, but but again, this is where I my frustration with this whole space is.

• 37:42Speaker 01

  Everything being promised, like there's such incredible things. As you said, anyone who's used the new Bing Chatbot, even ChatGPT, this stuff is amazing. Having worked on it in real business applications, when you actually are like digging into it, it's incredible. But over promising on this stuff, my worry is, again, we get into another hype cycle where when the first iteration of someone, because Satya even in his memo, I think he said, "Sometimes it'll be right and other times it will be usefully wrong."

• 38:14Alex Kantrowitz?

  Yes.

• 38:15Speaker 01

  That's a technologist speaking, because that's someone who's thinking, "Okay, I will, you know, like that usefully wrong will then help us train it further and further." To an average user, that first usefully wrong ends it for them. They're no longer interested. If it turns into usefully wrong, where your boss gets pissed at you, that's definitely not usefully wrong. So I think we're still in that, I mean, not still, it's certainly accelerated over the last week or two or a few months, where we're being promised the world and I don't know if we're there yet.

• 38:50Alex Kantrowitz?

  Usefully wrong does seem like some amazing marketing speak, but I can see it. I can see, like, I have more faith, and this has sort of been where we sit on the two sides of things, I have more faith in people to be able to navigate and see the faults of AI than I think you do.

• 39:06Speaker 01

  I think you do.

• 39:08Alex Kantrowitz?

  I think that they could, it could be usefully wrong. But there's another thing that's interesting here. We're like, there's all this action, all this interest in what's going to happen. And when generative AI gets into Microsoft Office or Google Workspace or Workplace, you know, don't hold me to this, Google. But Benedict Evans, who's again a friend of the show, and thank you Benedict for dropping a link to the show in your newsletter this week. I'm sure there are some people listening who came through that, so thank you for being here. And he made a great point talking about how, and I'm just going to read his tweet, "Microsoft and Google adding generative AI into Office apps is a classic pattern of incumbents making the new thing a feature." But the new thing generally also enables completely new ways to solve a problem.

• 39:52Alex Kantrowitz?

  Easier spreadsheets is less important than why is that a spreadsheet? So his point is basically like you're instead of having to ask questions to Microsoft Excel, and I'm just gonna extrapolate from what he's saying. Maybe one day you're just gonna upload all your data to some AI bot and then have a conversation that way. So why make it a feature on the product versus make it the product? And maybe that's where they're vulnerable.

• 40:17Speaker 01

  Yeah, why I love Twitter, Elon Musk aside, is because sometimes you get those flashes of brilliance that make everything make sense. And that, I felt that tweet was so perfect because it is. It's having a co-pilot alongside a spreadsheet assumes spreadsheets make sense. And I'm actually the first to believe that spreadsheets aren't going to make sense where we're going, and that's a good thing. That, you know, like having it embedded into Google Docs and the way like writing blog posts in that traditional way or whatever else, like the entire user interface is going to change and that's a great thing.

• 40:57Speaker 01

  So I think, and this is exciting, this is what makes this moment really, really interesting because Microsoft and Google are trying to integrate it into existing products, which obviously is a huge advantage, but it really opens the door for completely new ways of working and that's kind of cool.

• 41:17Alex Kantrowitz?

  Yeah, I'm sort of irrationally excited about the fact that like you will just be able to upload data, like for me, like uploading listen data. Hey, what days are better to post the show? What type of topics do well? You know, all those type of things or even, you know, with the newsletter. And instead of like having to like put it in a Google Sheet and examine, like to be able to converse with the data sounds unbelievable.

• 41:42Speaker 01

  Yeah, but this is one of those that Google Analytics, I think it was like the 3-pointer, it was like a major update a couple of years ago, was supposed to introduce natural language queries. Then no longer did you have to do reports, you had to, you know, what was our traffic over this month period and how's it compare? It doesn't work well, and at least for me, and maybe others have figured it out, but it's still just traditional reports and stuff. So when this happens, I will be excited, but until then, I'm waiting, I'm waiting.

• 42:17Alex Kantrowitz?

  The Biden White House is telling ByteDance effectively, you either divest your ownership in TikTok or you face a ban. And this has been part of a, a long discussion about TikTok and its place in the United States. There's certainly worries that TikTok could have an influence on the culture of the United States or potentially make the data unsafe, which I think everyone I speak with is, is less concerned about that. The real concern is sort of the cultural influence. TikTok has a very interesting ownership structure. Okay. 60% of ByteDance shares are owned by global investors. 20% are owned by employees and 20% are owned by the founders.

• 43:00Alex Kantrowitz?

  You know, when you put that all together, does divestment actually? I mean, I know it's a hot buzzword. I know we've been talking about it for like four years at this point. Does, does divestment actually do anything?

• 43:09Speaker 01

  Okay, I, the news this week, I've been making this my call. I think on this show a couple of weeks ago on CNBC, multiple times that TikTok will be banned in the U.S. this year, we're getting closer to it. I think divestment is impossible and I think there's a couple of reasons behind that. One, infrastructurally, it is so ingrained and tied to the entire ByteDance infrastructure that to properly divest it, I, I just don't see technically how that works in any easy rational way. Again, everything is still built on ByteDance infrastructure. Even at the, I actually, at South by, I met some people from Douyin,

• 43:51Speaker 01

  which is the Chinese equivalent of TikTok and their marketing like expansion to China. And their offices are in the TikTok office in LA. Like it's, everything is intertwined. There's been a million articles that have come out around how, you know, how intertwined the company still is with the parent company. So I think that's the biggest part that I don't see how it happens just at an infrastructural level. But then the second part, and going back to the late stage valuation question. ByteDance, there was actually two days ago, the Abu Dhabi, one of the large investors, invested at a $220 billion valuation.

• 44:30Speaker 01

  That's down from $330 billion, and that's down from Tiger had invested at a $400 billion plus valuation at the end of 2021. So ByteDance, like plenty of other tech companies, is in a very precarious situation itself financially, and TikTok is a significant chunk of its revenues. So if you are a potential buyer, imagine going into that deal. I would feel bad for the investment bankers trying to work on that and assuage any potential investors. Like, I mean, not only do you have the threat of American governmental intervention, just as a business, you are going to have to realize a massive haircut on whatever that private valuation has been.

• 45:15Speaker 01

  And I think ByteDance is going to be afraid to do that because it could bring down the entire company's valuation. Who knows what other second order effects there are there? So I think between the technology and the value and the actual execution around that, I just don't see who buys it and how it happens.

• 45:33Alex Kantrowitz?

  So you just think that it's just going to get banned outright instead of the sale?

• 45:35Speaker 01

  Yeah, I think on one side you have where divestment is just operationally hard, but then on the other, and this is what I've been saying at the beginning, from the beginning, I think the data location is scary, but not necessarily, you know, a deal killer in the sense that like, yes, that is something that theoretically maybe you can make work. The idea that the Chinese government is following your views of who you follow and whatever else, even though there were reports of it being used to track journalists who are covering TikTok. But to me, the bigger issue, and this is the cultural and political issue, our entire

• 46:18Speaker 01

  pop culture is defined by an app owned by, in part, by the Chinese government, or at least whose management still, for all intents and purposes, has to answer to them. And that's nuts. It's a black box algorithm. And this is the thing from the beginning for me, it's amazing that we've gotten this far, is that it's a black box algorithm defining what we talk about and think about as a, I mean, especially for a majority of 20 to 35 year olds, like. It's that's crazy to me. I don't see how that's tenable for any, any duration.

• 46:52Alex Kantrowitz?

  There has been this view that like, if you do ban TikTok as a political party, you're doomed. And I thought that that was, as I would sort of ascribe to that as well, that it will be a politically costly thing to do in the United States. Therefore, I didn't see it likely to happen. However, we do have copycats on YouTube and reels and they're getting better. And it might not be, I'm starting to be open to the possibility that it might not be the hit that I and many others have imagined.

• 47:18Speaker 01

  If my millennial chat groups are any indication, everyone's sharing reels and even YouTube shorts now I'm starting to see. Whereas that never came through before. Obviously not Gen Z myself, but, but I think, yeah, like everyone has a competitor, short form video has been commoditized. And if your patriotic calling as a 22 year old is you don't use TikTok, you use Snap Spotlight and YouTube Shorts and stuff. I mean, I don't know, will that, there really be a massive backlash to that? I, I can't see that happening.

• 47:57Alex Kantrowitz?

  Yeah, well, I do think it's possible, but we do, who knows? Okay, so let me just drop this comment in quickly. So Michael DeDasco mentions that GPT-4's word limit is 25,000 words versus 3,000 in GPT-3. And those are big training, big changes in the training and to work with new models. So it's obviously able to handle a lot more text. And we're going to see a lot more interesting use cases come out of this stuff. And that is the perfect segue into our final segment of the day. We always try to end with something a little bit fun. And here's a real question of the week is whether generative AI will solve dating. Because I saw this thing, it was a screenshot from Reddit, asking about Reddit AI thing, asking whether this is ethical.

• 48:42Alex Kantrowitz?

  This person made a dating app bot to get dates and it actually works. And the person said that. He created a bot designed to learn his preferences based on his previous matches, allowing it to understand what type of girl and engage in meaningful conversations that are tailored to his interests. And he says in the first month, the bot scheduled 13 dates for me, all of which were with people who matched his preferences and had similar interests. And he no longer has to waste time swiping aimlessly or struggling to come up with conversation starters.

• 49:18Alex Kantrowitz?

  Is this the future?

• 49:19Speaker 01

  I am glad I'm currently married. I am glad I am not in the dating world right now if this is what's going on. But, but in a way, let's take this back to Benedict Ebbins' point. Swiping through endlessly, maybe that does go away. Maybe we are finally at a point, and I think like dating apps, it promises for a while that using algorithms will find you the right match. Maybe they will. Maybe they, maybe they'll actually get us there. Or the, uh, the ultimate logical extension of this is, uh, in the end, we'll all fall in love with bots anyways. So maybe this is, this is just gonna get us there one step quicker.

• 50:00Alex Kantrowitz?

  We have a comment. Yep. Again, what if bots, um, you know, date, date each other and hit it off? Will they become a couple?

• 50:07Speaker 01

  Can your bot cheat on you or can your bot with another bot?

• 50:11Alex Kantrowitz?

  Yeah. And maybe those bots end up feeling the true, what, what real love is and, and we never do. Uh, humans sort of lose that capability.

• 50:19Speaker 01

  I'm not gonna, I'm not gonna deny it. I have a, I have a three and a six year old and I have, I was joking about this like a couple of years ago. It is not out of the realm of possibility that like 15, 20 years from now, I will have to be in the conversation that like, I love this robot. You don't understand me. I just love him. I, I don't think that's a complete impossibility.

• 50:43Alex Kantrowitz?

  Just wait. And I'm not saying this is gonna happen to you, but just wait until people have their spouses leave them to start dating bots.

• 50:51Speaker 01

  I think there was like one viral story where someone claimed that they left for a bot. But honestly, if we look, if we think about kind of like cultural representations of AI, her. The movie was the best because like it was so good for anyone who hasn't watched it. Joaquin Phoenix falls in love. It's like the voice of Scarlett Johansson. She's an AI bot. That's in this like always on earpiece of his because I remember what was so cool about it was the future didn't like it wasn't flying cars and everything else. It was just kind of like somewhat normal and look like, you know, a representative of what today's world looks like. It just had

• 51:30Speaker 01

  this entire layer over it. And I, I think that might be the most accurate representation of where we're going out of anything versus Terminator or some other crazy thing like that.

• 51:43Alex Kantrowitz?

  Well, it is interesting and I know you're good. This is, we should probably start our next conversation about this because the debate is metaverse or generative AI and now like funding has gone from, you know, metaverse to generative AI, but ultimately like, I think that's where we're just going to end up in a metaverse filled with generative AI.

• 52:00Speaker 01

  Will you date a VR avatar or a generative AI bot in an always on earpiece? That's, that's the debate, I think.

• 52:08Alex Kantrowitz?

  This is the, it's a matter of personal preference, Ranjan.

• 52:10Speaker 01

  All right.

• 52:11Alex Kantrowitz?

  I wouldn't dare ask what goes on in Europe.

• 52:12Speaker 01

  Whoever you choose, whoever you choose.

• 52:15Alex Kantrowitz?

  Thank you so much, everybody for listening. Thank you, everybody watching on the feed. Very lively feed as always here on LinkedIn. So thanks for all of the great questions. Again, we will be back on Wednesday with my interview with Kevin Systrom, wishing Ranjan a great trip next couple of weeks in Europe. I'm sorry, I'm disclosing your location.

• 52:33Speaker 01

  I'll see you all in a couple of weeks. It's some time off finally.

• 52:36Alex Kantrowitz?

  Yes. Excellent. And, and again, thanks to all of you for listening. Really appreciate it. It's been, it's been great being able to take you through some of the SVB news and some of the other news going on and, and we won't relent. So that will do it for us here this week. Thanks again for listening and we will see you next time on Big Technology Podcast.

• 53:08Speaker 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' health care 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, theHouseMajorityPac.com. Not authorized by any candidate or candidate's committee.

• 53:37Speaker 03

  In this life, you can't win.

• 53:39Speaker 04

  When the world gets you down.

• 53:40Speaker 03

  No!

• 53:41Speaker 04

  Escape to Pluto TV.

• 53:42Speaker 03

  This is objectively fantastic.

• 53:44Speaker 04

  Relax into the movies and shows you already know and love, all for free.

• 53:48Speaker 03

  Shwing!

• 53:49Speaker 04

  Even our news is uplifting.

• 53:50Alex Kantrowitz?

  You stay classy, San Diego.

• 53:52Speaker 04

  Watch School of Rock, Crazy Ex Girlfriend, Anchorman: The Legend of Ron Burgundy, The X-Files, Wayne's World, and so much more.

• 53:58Alex Kantrowitz?

  We're not worthy!

• 53:59Speaker 04

  The best things on TV really are free. You just have to know where to look. Escape to Pluto TV. Stream now, pay never.