That Was The Week 2023 #14

Speaker

I've got sunshine on a cloudy day When it's cold outside, I've got the month of

Speaker

April the 21st, 2023. It's that time of the week. It's that was the week. But it could be later. It could be Halloween. We are talking this week, death and destruction. Everything has gone according to Keith, mags, DVDs, even BuzzFeed, Keith. So is death the story of this week when it comes to tech? It really isn't. But it was the most interesting collection of headlines I put together to make a front page. There's lots of other news apart from death. But it is worth saying, who knew Netflix still would deliver you a DVD? But they they do. And you can subscribe to a DVD only version of Netflix. They're closing it down in September. No more DVDs. And then two of the most well known computer magazines in America, Maximum PC and MacLife are the last two standing both closing down. And last but not least, BuzzFeed News, which is that kind of irritating clickbait headline style of news that you see everywhere you go, is also closing down, which must be testament to the decline in the advertising market for clickbait. So as you say, though, in the editorial, none of this is particularly surprising. But you say that these closures and in addition, you talk about Elon Musk's first starship offer an opportunity for us to acknowledge and mark. So you're telling us to be historical, contemplative, Keith, not to overreact, not to over-traumatize. I think it's always important to be historical. I mean, it's dialectics as opposed to hysterical. Exactly. And yeah, I think we should be historical. I mean, I mean, I don't know about you, but in 1994, when I did my first internet company, I think we already predicted the end of magazines. And it's now almost 2024. So, you know, 30 years later, and it's kind of happening. It's slow changes when you look when you look at it. And then when you go to the airport, the stores are still full of magazines. So someone's buying them. Somebody's buying them. Yeah. Yeah. I mean, The Economist is a good example. I always pick up an economist and about you when I'm flying. Yeah, they somehow have managed to survive but not computer magazines. What strikes me about if we're if we're using this as an opportunity to pause and think is how Netflix have successfully negotiated each new chapter in technology. Most companies just survive one, maybe two chapters. But Netflix seems to be perpetually reinventing itself. How has it done that? Well, I think a bit like Amazon, it has a core focus on home entertainment. What do people want to do at home and convenience. And they've managed to add quality to that with their own. I mean, it is remarkable when you think about it that Netflix make good movies, and good TV series, as does Apple. And so HBO is now not in a league of one, making good quality TV or Showtime, maybe two. And yeah, they basically have that core value of convenient home entertainment of high quality. And it works. And because we all love to, we all would love to settle down around eight o'clock at night and be entertained. I've always I have to admit, I've always been amazed with all the hype around BuzzFeed. Was there ever a reason to feel optimistic about it? Wasn't it always even at its best moment, a losing proposition? I think it was. I mean, you and I aren't gonna disagree about that. I would go even further. I mean, I was a scam, do you think in a way? Is it what a scam? It is a scam, but it's a clever scam. I mean, it's really the attention economy manifest as a really bad app that's really good at writing clickable headlines. And they're stealing your attention, and then disappointing you. It's a bit like those ones you click on, you know, the, they say the 25 best places to holiday, and you click on it, and you literally get ads on 25 pages as they force you to click through one at a time. And you do that once, maybe twice, but anytime after that, anyone who continues he does that. I don't know what you say. I mean, the thing with BuzzFeed news is, if ever there was a product that would be instantly destroyed by generative AI, it's that because this stuff is churned out. I don't know, historically, who buy by journalists making $5 or $10 or $15 an hour, I'm sure that generative AI will already create better quality content than BuzzFeed news. Yeah, I bet you I could copy and paste this week's. That was the week into open AI and ask it to write attention grabbing headlines for each story, and it would do a good job.

Speaker

The the depressing thing, though, and this comes back to what you just said, is that the death of BuzzFeed news doesn't mean the end of BuzzFeed style garbage news, does it? No. I think, you know, YouTube is a really great example of this. Increasingly on YouTube, there are robotic voiceovers to repurposed be real material about topics that people care about. I discover it because there's a lot of Manchester United ones. And you start playing it because the headline is interesting. And you get this literally this computer robotic voice using a Chinese version of English usually, or a Korean version of English in the sentence structures. And, you know, it's enough to get my attention. And I get fooled quite often, clicking on those as opposed to better quality sources. So I think the race for attention is a race for money. And when money is available, there will be people who will try to race for it. And they'll appeal to our, you know, that book about the seven deadly sins, and how that correlates to marketing, lust, and I can't remember what they all are, but you probably can. The fact you thought of the seven deadly sins and only could think of lust means you have a one track mind, Keith. Exactly. Well, I am I am male, that kind of suggests a bad stuff, doesn't it really these days? So it wouldn't be that was the week if we didn't have something about Substack. And you have a piece, an interesting piece on Substack, can Substack save the social network? One reason this is relevant in the context of the decline of the death of

Speaker

this, these these bad websites, is that Substack by definition, can't be created by AI, I don't think yet, certainly not for the next few years. So it speaks to a high end in content. Yeah, I think I think AI would create a really bad version of Substack, there would be and then it wouldn't be Substack. That's the whole point. I mean, it has to be good to be Substack because otherwise no one's going to pay. Yeah, even $5 a month to access it. Exactly. And so I think Substack represents, when you talked about the cult of the amateur, Substack represents the countertrend to that. And these days, it's the cult of the, you know, the cult of AI producing content, which is a bad use of AI, in my view. You know, and Substack represents the countertrend. And it's a good countertrend. But is it for real? Or is it just for guys like you and I who have the time and the luxury to both produce high quality content and buy high quality content? Is it just an indulgence of the coastal elites, Keith? Well, it isn't going to be interesting to anyone that isn't, you know, thoughtful and a critical thinker, somewhat intellectual. So by definition, that doesn't describe the masses. There is kind of an elite that consumes that kind of content, which we're part of, better or for worse. And, you know, you and I are familiar with the British publishing scene. If you count the readers of The Sun, The Daily Star, and The Daily Mirror, and compare that to the number of people who read The Economist, for example, or The New York Times, you'd find that the internet is a reflection of the real world. Well, Ben, the internet is an outrage machine, a shame machine, and it was defined, or has been defined by Twitter. You have a piece about RIP, rest in peace, Twitter. You've been a Twitter optimist for a while, Keith. Are you beginning to feel that maybe it really is now, if not dying, certainly in the process of dying, that can take years on the line. I mean, it's like Yahoo. Yahoo's dead, but it still exists. Yeah, no, actually, I published this piece almost ironically. This is a piece by Vox that is, you know, Vox has a vicious campaign, a little bit like Casey Newton has against Elon Musk's Twitter. And they write with such veracity that they're really not journalists or even opinion makers anymore. They're more like the Fox News version of a Twitter conversation where they're campaigning, and they don't really care about balance at all. They only care about evidence that they're right. And this Vox piece is one of those. It's actually a disgusting piece, in my view. So I put it in there more. Well, it's hard to outrage Keith Tidd. Disgusting. What was disgusting about it? Well, look, you have a weird fetish for Musk, but he's an incredibly annoying man. I mean, he's smart, blah, blah, blah. I accept all that. But how could you possibly not get pleasure out of seeing him publicly shamed and losing all his money? Anyone would do that, wouldn't they? There is a sport in being against Elon Musk. I mean, he's so annoying. It's like watching a five year old have a tantrum and get away with it. Yeah, but you know, I don't think that's what's driving most of it. I think it's actually they're irritated with his tolerance for right wing views. And it's really part of the MSNBC versus Fox narrative that these people are jumping on the MSNBC side of that bandwagon painting Musk. This isn't just a new right wing Keith Tidd, because I can't stand Musk and I don't have a lot of patience with MSNBC either. It's just so annoying. And what he's done is... But it doesn't matter, does it, that is annoying? I mean, would that really justify... I'll give you an example. I mean, what was really annoying was the stuff on this public, what was it, publicly financed media making and trying to humiliate the BBC. NPR and PBS and the like. Yeah. But that was just playful in my view. I mean, you can be annoyed by that. It's playful if you watch Fox News. It's not playful for anyone else. Anyway, let's move on. So hopefully Twitter will be dead by the next time we meet next week. Although I doubt it. I doubt it. As one thing dies, you're a student of Marx. One thing is being born. Sam Altman is the anti-Elon Musk for the moment. We'll probably discover more about him as he becomes wealthier and more powerful and probably would eventually replace Musk as the wealthiest man in the world. He talked about LLMs, their size not matching as much moving forward. I'm not going to make any joke about the size of LLMs. But what does he mean? Why is this important? Why do you include it in the newsletter? Well, because the critics of OpenAI have focused in on the call for a pause in the growth in the size of LLMs and correlating that to how dangerous they might become due to the fact... What is an LLM? It stands for Large Language Model. And it basically means the training data that OpenAI-like environments learn from. And the assumption is that

Speaker

they get worse in terms of danger to humanity as they get bigger, because they get cleverer and more devious. And Altman is making the point that there's only so much data that you can learn and it's pretty much peaking already. And so most of the innovation isn't going to be growing the size of the training data. It's going to be understanding how to align the LLMs with human needs. So all the innovation, and this is a perpetual feature of the web, is going to be on the edge. And a man who knows about the edge and has written about it extensively is Tim O'Reilly, one of the few grown-ups in Silicon Valley who is anything but hysterical. You have a piece in the newsletter this week on O'Reilly's piece about not being able to regulate or shouldn't regulate, which we shouldn't be regulating stuff that we don't understand. What is Tim saying? He's really calling for industry consensus audits of what AI environments are able to do and not able to do, as a kind of a ground zero base level understanding, apples to apples across all the different AI environments, so that you can actually judge for yourself if there are things that one should be concerned about. And if there are, then based on that knowledge, you can have an intelligent conversation about whether they need regulating or whatever. So he's asking for thoughtful analysis and processes that feed thoughtful analysis, which in contrast to, say, Gary Marcus, who isn't the by far. Gary makes some logically good points as well. But I think Tim is even much more balanced. The difference between Gary Marcus and Tim O'Reilly is that Marcus is a subject expert, but not a grown-up. Tim O'Reilly is a grown-up. He's not a subject expert. Yeah. Well, Marcus is a subject expert in a field of AI that is different to LLM's. And he's a self-confessed believer in a different path to artificial general intelligence, and sees LLM's as the wrong path, and therefore is compelled to become a critic. And to be honest, he'll throw everything, including the kitchen sink at LLM's, because he doesn't believe it's the right path. And he's pretty unconstrained in what he throws. And if he's right, then you could argue that the end justifies the means. And it's fine. Yeah. I mean, I think the thing on this generative AI is it's so, I think Tim's right, because it's so unclear what this is going to mean. I mean, we've been through so many hype cycles in Silicon Valley, Keith, you and I, that all this promises, like the Web3 stuff, decentralized companies, blah, blah, blah. None of it actually became real. Now, I think AI is real. I think most people accept that. But it's not clear what kinds of products. I mean, you didn't link with the Google announcement that they're including generative AI eventually in their new search engine. But that's a big deal, isn't it? Yeah. Have you used chat GPT much yet yourself? I've played around with it, but not really. I've done it when I'm quoting myself, just for amusement. So, I used it this week. You know, my team uses Python to write code to query the data on SignalRank, and I never learned Python. So, I asked it to write me a Python program. I told it about the database that we run in terms of its structure and the names of each of the fields, and asked it to write me a Python program that would calculate the bottom score of the top 5% of Series B rounds in the last 90 days. And it output this Python, and it kind of was outputting to screen. So, then I said to it, well, can you make it output to a spreadsheet? And it told me what to do, and I copy and pasted it, and it output to a spreadsheet, but not an Excel spreadsheet. So, then I said, well, could it be an Excel spreadsheet? And it gave me a slightly changed piece of code, and now I can just run this program, and I get an Excel spreadsheet with my results in. And, you know, it's amazing. It took about 15 minutes for the whole thing. So, I'm just convinced that as an aid to human output, it's just a fabulous technology. And if you think it's good for something else, like suggesting points of view about politics, then that's just a mistaken use of it. Let's go back to the real world. Meanwhile, when everyone's obsessing over chat GPT and generative AI, Apple continues to colonize the rest of the business world. They now launched a really interesting assault on big banking. They launched their Apple Card savings with 4.15 interest rate. You link here. How big a deal is this for you, Keith, in terms of Apple's assault on retail banking or on credit card companies? Well, I think the headline news is the 4.15%. Apple's whole system is run by Goldman Sachs, and Goldman Sachs customers don't even get 4.15% for their bank accounts, so they have a Goldman Sachs. So, Apple's pulled off something quite remarkable here in alliance with Goldman Sachs. And at 4.15%, I mean, banks are not even close to competing with that kind of interest rate. They do limit you to $250,000 that can earn that interest, but that's well enough for most people. And it seems to me that money and banking are being separated. And if you have more than $250,000, anyway, you're not going to put it on your Apple Card. There are lots of ways of earning 4.51%. Not so many ways. I mean, you can, you know, that's before tax. So, obviously, you're going to pay tax on the interest rate. So, it's probably closer to 3% once you've paid your tax. But there aren't that, I mean, certainly with banks, you can't do that. You can go and buy municipal bonds and treasury notes and things like that that will get you into that ballpark. And you can even get tax-free municipal bonds, although they're more at risk. Apple's isn't at risk. It's like a risk-free 4.15%. Yeah, it's interesting. And I mean, are Apple doing anything in the AI space, or are they just concentrating on real-world business problems? You'd have to say they're not doing anything with AI as it faces to consumers. Like, Siri really isn't AI. But they are doing a lot with AI when it comes to software, like their camera on the phone is pretty much 100% AI. So, AI does play a role in a lot of their products, but not as a packaged chat interface to consumers. That is going to have to change. I mean, one of the stories this week, I don't know if you're going to call it out, Thomas Tungus says every startup has to become an AI startup. Well, every company probably has to become an AI company. Yeah, and every company, I mean, all these big tech companies were AI companies before generative AI. I mean, Amazon, Apple. So, we've got two startup of the week, Keith, and it is a company called SeatGeek. What's interesting about them this week? They're interesting because they filed for an IPO. And the suggestion is that if they succeed, it may be the sign of a turning point in public markets because it's super hard to do an IPO right now. So, they filed for an IPO. There's a bunch of others lined up. Are they in any way original? Are they any different from any of the other secondary market, to use a euphemism? I think they have partnerships with venues and performers or teams so that they are not just the secondary market. They're often the primary market as well. And they're very powerful. I've used them a couple of times. Have you ever used them? No, I don't buy on the secondary market. I absolutely avoid it. I mean, these days, tickets sell out so fast that you end up... I just bought tickets for Manchester United versus Borussia Dortmund in Las Vegas that's happening at the end of... With anyone else, you could be the only person there. I doubt it. It's going to be a full stadium. But the way they did it... A friendly in Las Vegas, a summer game, is that going to fill out? It will, yeah. Last time I saw them in the US, they filled the bowl, the stadium in... Berkeley Bowl? No, the Los Angeles one. I forgot what it's called now. But it's 100,000 people and it was full for a game against LA Galaxy. For you, Geek... Seek Geek is not interesting in itself. It just talks about the climate of the public markets. Is it realistic to hope that the markets, the IPO market for tech companies might come back by the end of the year? I would say no, based on what I know today. But they're clearly... Somebody's telling them yes. So let's see. And finally, the tweet of the week for this week, the last... No, it's the second to last Friday in April. Neil Thanandar, what's he been saying, Geek? So this is a bit inside baseball for venture capitalists. But he's making the point that it's super hard to make money from venture capital. And it's harder still if you specialize in writing big checks to a small number of companies. The easiest way to do it is to write small checks to a large number of companies. And this is an ongoing debate within venture capital. He's a strong advocate of small checks to a large number of companies. And because of that, he then talks about how do you allocate to good companies when you do that? And it's a long thread. I recommend clicking through and reading the whole thread. But the short version of why I put it there is it's kind of arguing for the existence of my company, SignalRank. I knew that was the reason. I wasn't going to say it, but at least you've acknowledged it. So say it, Geek. Why is this such an ad for SignalRank? Well, it's making the point that really good seed stage investors find all the good companies, but don't have enough money to keep going. And if one was to put money into those companies, one would do very well. That's basically his thesis. And in our case, we're targeting, you know... Even in bad times, in bad times, you just lose more money. Actually, in bad times, two trends happen simultaneously. The first is there are less funding rounds done across all classes. And the second is the good ones shrink as a proportion of the whole. But the good ones turn out better than the good ones in a more flush period. So it's a bit like searching for gold in a river full of silt. You know, there's less gold. The river isn't flowing as fast. But if you can find the gold, you're going to be one of the rare ones to find it. And that is true.