May 5, 2024 ยท 2024 #15. Read the transcript grouped by speaker, inspect word-level timecodes, and optionally turn subtitles on for direct video playback
Edit labels for this show, save them in this browser, or download a JSON override for the production folder.
Transcript Playback
Where is the Money?
Human Transcript
Timed transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Speaker 2
Hello everybody, it is Sunday, May the 5th, 2024. We haven't done a That Was The Week show for a few weeks. First Keith was away, then I was away, but we are back with That Was The Week and my old friend Keith Teer. nothing much seems to have changed in the world of tech and venture capital though uh this week's uh that was the week newsletter is entitled with its characteristic terrible ai art where is the money and keith is joining us as always from palo alto which might be renamed the money rather than where is the money. Keith, the money is in Palo Alto. You may not have it, but it's somewhere there, isn't it?
There's a lot of money in Palo Alto for sure, but it isn't earned by the people who are the subject of this week's newsletter. This week's newsletter is all about creatives, video creatives on the one hand, hence the guy in the background there, and writers on the other hand, the guy at the desk.
Yeah, everyone looks miserable, but is there anything new here? Where is the money? I mean, we've been talking about this, Keith, for years, you and I, one way or the other. I've been writing and thinking about it. The money has never been with the creatives, has it? Especially in the digital economy.
Well, the funny thing is creative money flows are very like venture capital money flows. They have the power law. meaning that with a large number of actors, there's only a small number of winners, and the curve of winning is skewed to those few who win. And that's the subject of Doug Shapiro's essay about... It's about video, and Doug Shapiro is an expert on video. He's got a TV background. He's... Yeah, I know, the... Full of the money.
What is the takeaway from... Well, his takeaway, which I tend to agree with you, is not new. His takeaway is that, you know, we're all talking about streaming media growing massively and Netflix and... So on. YouTube, of course. YouTube, TikTok even. And he makes the point that the total pile of money spent by consumers on video hasn't changed for 20 years or more, that it's the same pile of money. It's just moving from one person. The winners are changing, but the total amount we're spending is not changing. Therefore, video creation is not a growth industry. That's his main point. He does talk about opportunities to steal some of the pie. In particular, he's focused on live sports as a place that...
Yeah, his numbers seem to suggest that sports is way ahead of family and kids. And then right at the bottom is women's, which is surprising, and arts and entertainment, which is less surprising.
Yeah. So he's demystifying. I mean, he makes a living from consulting to broadcast networks and streaming networks and so on. So, you know, there's somewhat of a self-interest on his part to disclose what he considers to be surprising facts. I don't think the facts are that surprising.
Right, so let's go back to your... And Keith, you've got to change. You've got to change your AI artist. I don't know if that's possible. We'll talk more about AI. But it gets worse, in my opinion. Where is the money? We've got a terribly unhappy... Two terribly unhappy people. One who looks like Jesus. The other who looks like...
a very unhappy writer, and then we've got a videographer. Is there any money in the people who carry large cameras around and put their content on YouTube?
Well, that's the case. I mean, you and me, doing this video. I think neither one of us is doing it for the direct money that will come from... Well, you promised me that,
Because Seinfeld already has a brand. I mean, we've talked about this. If you've got a brand, you can use these platforms. But if you don't, I mean, presumably, what about all these video producers on YouTube who are apparently making millions of dollars? Is that just a delusion according to the pieces?
No, the PCs recognize that that's happening, but they say that that money is coming out of pockets that were previously spending on other video. So, you know, you're closing down your $300 a month Comcast, and now you're paying, you know, probably a little bit less than that for all your streaming, including YouTube and TikTok. And there are, there are people making hundreds of millions of dollars actually on YouTube. MrBeast is the most well-known. But I think there's lots of others. If you go on YouTube and you search almost any area of interest, you're going to find experts making videos with an audience. Sometimes the audience can be like, this is kind of weird, but I watched this couple of building renovating a home in Italy. They've been going on, I don't know, more than a year easily. And every week there's a new episode which involves lots of building. And it's really not that interesting except I kind of... Yeah, it's hardly Scorsese or Tarantino. But they're getting a quarter of a million people watching it. So that's quite a big audience. That's bigger than CNN's average daytime audience.
There you have it. If you are one of these unhappy video... video makers in Keith's AIR go to Italy and put a camera on and start renovating your home. And then the news is no better, Keith, according to the pieces for publishing. It seems like, and correct me if I'm wrong, Keith, you always will, of course, the more new technology there is, the less anything changes in the creative industries. Is that fair?
Yeah. The one you're pointing to is Freddie de Boer, which sounds like a South African name, but it may be Dutch. Saying that he feels like a loser. A bit like you, Andrew. He writes books. I'm a loser, you say? No. You write books. And so does Freddie. And Freddie has been bemoaning the fact that there's no real... promise of money for people who write books, that it's a lonely endeavor that results in nothing. He sounds very depressed, actually, when I read this. But the reason I put it in is he also speaks about how Substack is not a solution. And most people, I think, who write think Substack may be a solution. So that's a new fact or a new point of view, I would say. And he's basically saying that in publishing, the publishers who do the work to promote a book really only put money into the top 1% of books. The rest is they pretend... It's a winner-take-all.
So it goes back to your video from Italy. The only people making money are weird books about... I don't know, food. I mean, real books. And he says that... He quotes someone in your piece that... If Hemingway and McCarthy were starting out today, they wouldn't get published. I guess in the same way as if Scorsese and Hitchcock were starting now, they probably wouldn't be able to realize careers. Is that fair?
I can't quite get my head around what would happen to those people, but it's certainly the... know the bar for quality is very high now and there's such a lot of good quality um that that getting into that game is hard but there's also the bar is really low for talented people to build an audience and the best talented people do build very large audiences in all kinds of niches so so that is also true so i think it would be wrong to pick one or the other i think you'd have to say to Freddie, you know, that really it's on him to find an audience for the stuff he loves to do.
Actually, I think your book called City Amateur stands the test of history quite well. It makes the point that the availability of tools to everyone means that people without talent will use the tools uh as well as people with talent and the average person won't be able to distinguish between the two i think that last bit probably is wrong i think people can distinguish but the first bit is definitely right how many cycles we're going
to talk about ai in a second now we got pieces you don't have them in this week's newsletter you've talked about them in the past and you will in the future articles about how ai now will create a renaissance of video or music and audio and and textual In fact, you even have a piece about a watershed for music videos being produced by AI. Is the likelihood that AI will only compound a winner-take-all economy? Well, actually,
that is buried in the editorial, but there is a video by Marc Andreessen and Ben Horowitz which says we are at the cusp of the time in history when you'll find billionaire companies that only have one employee, the founder. They quote Bitcoin as an example of that that is historical. One person wrote a protocol, minted some tokens which they still own, and the growth of Bitcoin has made them many, many times over a billionaire. And they make the point that AI is going to create the possibility of a lot more single person billionaires. And so that does kind of correlate to this question. Where's the money? They are answering that the money.
Well, they would say that. I mean, Horowitz and Andreessen, they've been saying it for years and they've been making personally more and more money. I mean, now they're part of the winner take all venture. community so do you believe them i mean is it just self-serving i mean i assume they kind of believe or at least they want to believe it well how many times keith do we need to keep on hearing the same old stuff and it never happens
They're not creatives. They're just platforms. So there's always going to be platforms. TikTok is another good example, but it's not necessarily benefiting the sad videographers or writers in your AI art this week.
Yeah, I agree it's tangential, but I do think computer software engineers and product thinkers are creatives. They're just creating software and tools, not videos, audio or writing. I think all of the above will benefit from AI. Sam Altman was interviewed this week. I actually didn't put it in the newsletter, but it's on YouTube. He's always interviewed.
He made the point that he was at Stanford speaking to a class, and he made the point that ChatGPT-4 is the worst AI that any of them will ever have to use, that it's already internally at OpenAI what everyone thinks will become chat GPT-5, but whatever they end up calling it, is many, many, many times better than chat GPT-4. And that means that you can use that for whatever your goal is.
Yeah, but that's not the point. So coming to OpenAI, you had a couple of interesting pieces, actually, which appear to contradict themselves. On the one hand, and these are both from the same newspaper, The Guardian, and one of the world's last newspapers, eight US newspapers are suing OpenAI and Microsoft for copyright infringement. And then you've also got a piece from... from The Guardian about OpenAI using FT journalism to train artificial intelligence systems. Where is AI getting its content? Is this clear? I mean, is it stealing it or is it doing deals with newspapers like the FT?
I don't think it's stealing anything because most websites block you from being able to do that using something on a robots file. So I don't think it's stealing anything, but it is using publicly available information to learn from. I think there's two things happening at the same time. I love it when there's contrasting newspaper articles like that, especially from the same publication, because my normal thesis is there's more than one thing happening simultaneously.
Yeah, but only one of them represents a valid future the other one doesn't so i think the law case represents a backward looking past by the way those eight newspapers are all owned by the same owner so it the headline's a bit misleading well it's good for um it's good for the
So this is like the Springer, the Axel Springer deal. Yeah. But there are other US newspapers. The biggest case is the New York Times, which is a very large case. So this is not marginal, Keith, the legal stuff.
Uh, I, I, I, I, I walked to the Sanford law school many times, um, but they didn't let me through the doors, but my legal, my legal scholarship tells me that, um, OpenAI losing a case against a newspaper because it read they're publicly facing stuff to learn from, it is not likely that OpenAI will lose that case.
So where is the money in this emerging AI economy? You have, on the one hand, the platforms like OpenAI. On the other hand, you still have traditional creators of content like the New York Times and the FT. Who's going to clean up here? Who's going to make money?
Well, I think this week's, is it the video of the week or did I put it somewhere else? But the video about IBM buying HashiCorp for 6.4 billion. Yeah, it's the video of the week. There's a fairly decent example of where the money goes. HashiCorp was created by two 22-year-olds who thought that as the cloud and data become more important, it's unlikely that big companies will only want to host their stuff in, let's say, Amazon. They're going to want to use all the different clouds. So HashiCorp built this abstraction layer that sits on top of all the clouds and lets you manipulate where your stuff runs. And that's all they do.
That's the plumbing, Keith. Again, it's not for the creatives. Let's go back to your... Yeah, but... It doesn't benefit the guy holding the camera or the writer.
yeah yeah but you're not asking uh whether the creators make money you're asking where does the money go and i'm telling you the money will go to infrastructure a lot
Plumbers, electricians, and in the case of the internet, it's deep infrastructure for data, for computing. That's just getting bigger and bigger and bigger. Data centres are getting bigger. All of that will happen. And then on the other hand, consumer-facing, the key to money is audience, always. Always. Well,
it isn't, because you've got people with billions of, or certainly millions of viewers on YouTube not making any money. Or for that matter, people with large mailing lists on Substack who are struggling, as Freddie DeVore reminds us.
You're not right about the millions on YouTube not making money. I mean, if you've got millions of subscribers on YouTube, you do make a decent amount of money. But there is, you know, we talk about this power law or winner takes all. I think winner takes all, isn't it normal for all creative endeavors that winner takes all? I mean, there aren't very many creative endeavors where the spoils are spread evenly, no matter how good or bad you are.
You mentioned Andreessen and Horowitz. The other person you connect to this week in the newsletter is Nick Bostrom, who became very well known. He's a professor at Oxford, became very well known for a pessimistic book about AI that Elon Musk, Bill Gates all embrace. But now he's making the reverse argument. Is he suggesting that AI is going to save us, not kill us?
Well, he's asking the question, what if he's wrong? And AI actually is the solution to a lot of things. and you know uh uh given he's obviously making money on both sides of the argument this guy but um i do think this second question is the more person of the two questions because clearly there's no insight in understanding that ai is limited i still read gary marcus a lot but i'm less and less you know marcus has a new book
out uh i should get bostrom on my keynote show my um And now I'm going to talk to Marcus about his new book. It's an Oxford University or Princeton University Press book. It's out in September. So there'll be alternative views to books. So what's Bostrom's argument?
Bostrom's argument is that there's enough to suggest with what we've already seen. that there will be very few tasks that AI will not be able to lend its hand at, including increasingly in the physical world using robotics as a... Yeah,
Well, I think the trend is for AI to be freely available to every citizen of the planet. I mean, you can get pretty good functionality without paying for it.
He will give it away for free when OpenAI 5 comes out, I suspect, but... You're good at making jokes, but you're interrupting the flow of my answer with your good jokes.
But basically... It is very likely, according to Bostrom, that AI will be able to tackle many, many, if not most tasks, both intellectual and physical. Not that humans will be made redundant, because any human effort that engages with that content is going to make it better, most likely. So the human element doesn't go away. It just becomes supplemented or enhanced. And when that happens, well, what happens to the human race? What happens to the people that used to be needed to do some of the things?
Yeah, so you can be optimistic about technology, as I am, but pessimistic about its impact on humanity, as Nick is. I'm optimistic on both. Or you can be just pessimistic about everything, like Gary Marcus. So I think there's really three buckets of belief. Everything sucks, and it's going to make the human race worse. Everything is fantastic, and it's going to make the human race worse. or everything is fantastic and it's going to make the human race better.
And that's the Keith Teer, Marc Andreessen, Ben Horowitz argument. We're still in the early stages of this. One interesting piece that you connected with is that Anthropic now has a Claude chatbot app for iOS. I've been impressed with Claude. What do you think and why is this significant?
I use it a lot. It's significant just because it's competitive to open AI. And it means that, you know, there's more than a single big player. In fact, there's quite a few big players. And I think that tends to reinforce the point I just made, which is it's going to tend towards being free. You know, big companies like Microsoft can give things away for free, so can Apple, that are not core to their revenue streams. And OpenAI can give a lot away for free to consumers as Claude can because most of the money is going to be made in the enterprise. So I think I put it in there just because good for Anthropic and Claude is very good. I use it for SQL programming. They do limit your I pay. I pay $20 a month for both. And it does stop you after a certain number of questions and forces you to wait two hours to ask more. So they're kind of throttling the use a little bit, even when you pay them.
I mean, Google is starting charging for Gemini, which goes against their original business model. Can Anthropic... which is very much of a third or fourth player in this race, can they really survive? Or will they be acquired or an acqui-hire, as Microsoft did for the Suleiman company, the Reid Hoffman Suleiman company? Can Anthropic really compete against OpenAI and Gemini?
I think it can. I think it's got enough legs standing to keep standing. It's whether it can support the cost of doing so. It does have partnerships with Amazon and Microsoft, by the way.
They should be able to. I mean, the fact that there's more than one means antitrust shouldn't really arise. As long as OpenAI is standalone, independent, big, and probably the leader, I think the number twos and number threes can do a lot of things.
It's like the Premier League, Keith. It's just Man City, Arsenal and Liverpool. Because all we're seeing now in this supposedly revolutionary open AI economy is Google versus Microsoft versus Amazon versus Apple. History isn't changing, is it?
Well, you have to give credit to open AI and Anthropic for playing those big guys off against each other and managing to survive. The fact that they're still around and haven't been acquired... probably is testament to how smart they are, but also how scared the big guys are.
Is Microsoft playing the room, though? I mean, what's the relationship? There was a lot of news last year, of course, about OpenAI and Microsoft. What is the relationship exactly between OpenAI and Microsoft? Is OpenAI increasingly just a division of Microsoft, or are they genuinely independent?
So this is the cloud platform that Microsoft's... Yeah.
Words and timings
SothisisthecloudplatformthatMicrosoft's...Yeah.
Speaker 1
So they're a supplier, and because they're a supplier and OpenAI can't pay them, they're also an owner, because they're getting paid through ownership. So that's a trade-off. It's a deal I did with Microsoft back in the late 90s.
Well, it's gonna be interesting. Certainly not the last we talk about this. Meanwhile, we haven't had many stories of layoffs, but picking up one new story you have this week, Peloton to lay off 400 employees, their CEO is departing. Does that speak of the end of COVID economics, really?
I think it does. Peloton was definitely a COVID winner. Its valuation multiplied many times over during COVID and then has declined pretty much at the same speed since. And it's hitting the fan, as you would say. I think the CEO was already a new CEO, by the way, who was brought in to fix things and clearly has failed. So I put it in because there are a lot of layoffs. Google did some layoffs this week as well. And they're not.
Yes, I think so. I think it is. Product, by the way, is very good for anyone who wants a bicycle with a screen that lets you log into classes and train at home. It's still a great product. It's just the market for that product has shrunk a lot.
well we got to that point startup of the week and it's very much in keeping i think with the themes beehive attracting 33 million dollars it's a platform for newsletters does this speak keith of the the increasing vitality of the digital plumbing industry it's not for creators but guys like freddie de burgh can use beehive yeah
Well, Beehive is best compared to Substack, which I use and you use. Beehive's claim to fame is it supports sponsorships and advertising, whereas Substack really is all about subscription. And this investment happens at a time when it's already got over $20 million in revenue. So clearly something is working. I should disclose, by the way, that SignalRank were one of the investors in that round. Through a partnership with one of their seed investors, we helped them do their follow-on. So... Yeah, I thought it's noteworthy mainly because there's more than one player in this space.
So if you're on Beehive, as you said, I'm on Substack, so are yours. This show's on. What do you get with Beehive that you don't get with Substack and vice versa?
So if Freddie DeBeer is watching, should he go to Beehive?
Words and timings
SoifFreddieDeBeeriswatching,shouldhegotoBeehive?
Speaker 1
Only if he can attract advertisers. I mean, there is something in it. I mean, if you think about this show, that was the week. Probably there's a whole bunch of tech companies that would be more than happy to sponsor this show for...
Do they really do it? These companies all claim to do that. In all seriousness, if we went to Beehive and said we've got a few thousand followers or audience, they're not going to do anything for us.
Well, they're aggregated in a little bit like some of the podcast advertising networks did for podcasts. They get 500 podcasts in their advertising network, and they sell the whole thing to buyers based on content, how appropriate the ad is to the content. But they're an aggregator. The advertising layer is an aggregation layer. So, yeah.
Not good for the cred. Coming back to your AIR, not very good for the videographer or the writer. What becomes of Beehive, you invested in it? Does it become a target for companies, platforms like Spotify, or can it really survive on its own?
The jury's out. Let's see. I think you would probably argue that it needs to scale, probably needs to get to about... you know, at least 10 times its current revenue to be independent. I wouldn't discount that that's possible. It is possible.
Yeah. Well, I'm not sure if your algorithm's right, Keith. We'll see about that. Finally, X of the week is coming back to AI and robotics is something about figure. from Brett. Who is Brett adcock? Is he the CEO figure? I think he might be. So anyway, that was a big feature on 60 minutes. Figure looks to be the robotics company of the future. Are they the leader in all this?
They certainly have come from nowhere very fast. They only started working on this a couple of years ago. So they've got here very fast. I think Tesla might be catching up. And I think that Boston Dynamics, which launched an electric robot that doesn't need to be plugged into the mains, is also in the race. I think there are going to be domestic robots soon, soon meaning two or three years.
different it is this is this is where you blend software and hardware into things that can be done in the real world in the case of figure it's car production lines they're being trained to do specific jobs on a car production line so when finally
keep going back to our disillusioned writer and videographer from your open air from your AI art. At what point do all these people give up and get replaced by robots from companies like figure? I mean, if they can lift boxes, presumably they can make videos and write novels.
I think the answer to that is never, because I think the human pursuit of creative output is deep in the soul. And even if there were robots doing perfectly good things, as happened this week with the music video for the new single, which was made by OpenAI Sora, 100% made by Sora, there will be robotic robots.
Yeah. That's right. It's become quite popular. So humans are never going to stop wanting to create. That's who we are. And some of us will be great at it, even when robots are good at it. Even if they're not paid? Well, mostly they're not paid, right? I mean, most good artists only get paid after their success.