Transcript Viewer
Dec 31, 2023 ยท 2023 #44. Read the transcript grouped by speaker, inspect word-level timecodes, and optionally turn subtitles on for direct video playback
Speaker Labels
Edit labels for this show, save them in this browser, or download a JSON override for the production folder.
Human Transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Speaker 3
I've got sunshine on a cloudy day When it's cold outside I've got the month of May Everybody say I guess you'd say Hello, everybody. It is Friday, December the 29th, 2023. If it's a Friday, it must be That Was The Week.
Speaker 2
Our weekly view of tech. One would expect predictions from my friend Keith Teare. Everyone else in the tech business is giving predictions for 2024. All sorts of predictions, seven predictions, 24 predictions, blah, blah, blah. And then of course, looking back at 2023, some sort of summary, 2024 predictions, all the leading writers and thinkers in tech are predicting 2024, except for Keith Teare, his last newsletter was from the middle of December, baby, please don't go. We did the show about big media and Google and what's Absent from Keith's newsletter is a prediction for 2024. Keith, are you just chilling out over Christmas or are you against the idea of predictions, especially when it comes to tech?
Speaker 1
Well, you know, people who do predictions are kind of making themselves the centre of attention. You would never do that, Keith, right? Well, certainly not by my predictive capabilities. I think I would I, you know, there may be a lucky break where I score high, but it wouldn't be due to any forward looking insight. I think forward looking insight is for fools. And I'd rather, I'd rather, you know, kind of sit here and watch others do it, and be really happy that I won't be wrong 12 months from now, because I didn't say anything.
Speaker 2
It's interesting the way you put it. Forward-looking insight is for fools, but it's not, you're suggesting that it's not really insight. What is it, just gambling and you usually get it wrong?
Speaker 1
It's guesswork. I mean, look, none of us would have predicted open AI 12 months ago. I mean, it kind of blew us all away. I think most of us would have predicted that Web 3 would still be a thing.
Speaker 2
I wouldn't. I've been thinking even though it was a thing in 2022, let alone 2023. Maybe that's true, yeah.
Speaker 1
So we definitely, we wouldn't have predicted most of what drove 2023, I don't think.
Speaker 2
So I take your point about forward-looking insight or wrong-sight is for fools, but what about 2023? It's happened. How would you How would you generalize about 2023? What's happened? I mean, you've mentioned OpenAI. Of course, it dominated the narrative on lots of different fronts. Is 20 world tech historians, when they look back at 2023, will the title of that chapter be OpenAI?
Speaker 1
I think it might be the thawing of China-America relations. I think Biden's meeting with Xi and the various meetings leading up to that signify a pretty big change in the tone of America and China relations and I think there's an overarching canvas for tech to grow in that's super important.
Speaker 2
And that of course assumes that Biden wins the next election because if he doesn't and presumably Trump will run or whoever's running for the Republicans will run on a strongly anti-Chinese economic platform, all that will be undone?
Speaker 1
Hard to predict with Trump because he blows hot and cold on China. He does that, I'm best friends with Xi, we like each other, but I'm going to play a hard game kind of duality. So hard to tell. But I think America's self-awareness that its future is connected to China's, not disconnected, that's a big deal.
Speaker 2
So I take your point. It's a very interesting observation. But what exactly happened? I mean, Biden made one or two warmer remarks. Has anything profoundly changed? There's still a chip war, isn't there?
Speaker 1
There is. And China's, as we all would anticipate, China's building its own chip capability. A new chip was announced this week that's going into some of the Chinese electric vehicles. Xiaomi announced an electric vehicle yesterday that is set to be able to do 450 miles on a single charge. There's new battery technology coming out of China, which is sodium based, not lithium based. The price of lithium has gone from $80 a ton at the start of the year to only $20 a ton. You know, so there's a lot of fairly interesting development.
Speaker 2
Why should that represent a Thor in American-Chinese relations. The Chinese don't need the Americans.
Speaker 1
I think the thought isn't so much, let's decide to like each other. It's more of a American recognition that it's future is entangled with China's, and that if it disengages, it loses any ability to control the cadence of that relationship. So I think it's a defensive Thor, not kind of a let's decide to love each other Thor.
Speaker 2
But that's your thinking. I mean, who's articulating that within the administration?
Speaker 1
Well, you have to read the signals. I mean, suddenly it went from no... And you're an expert as a CEO of Signal, right? Some signals. I mean, the fact that it went from no engagement to a lot of engagement, leading up to presidential level engagement, obviously tells you there's some thinking going on that denotes more engagement. That's all you can conclude. Now, my PhD was about the transition of empires from being empires to not being empires, the economics of that. So I wouldn't claim to be an expert, but neither am I a layman. And it is the case that the economics dictate that America is going to have to cohabit the Earth with its future replacement. And whether that future replacement is China, India, Africa, which isn't coming onto the stage as a leading player of economic growth during the next 20-30 years. No one knows, but there will be a replacement. It's not going to be an American-only world anymore.
Speaker 2
So let's continue to look back at 2023. It was a year, of course, of Sankt Bankman Fried and his public humiliation. You were always arguing that we need to wait until he got his trial to determine whether or not he was guilty. Will anyone remember Sam Bankman Freed in 50 years? Is he a symbol of the corruption of Silicon Valley, the corruption of his generation, this obsession, fetishization of digital currency?
Speaker 1
No, I think he's an outlier and he'll be forgotten. I think he'll be like... Except by his parents, of course. A little bit like John McAfee. Do you remember John McAfee?
Speaker 2
Andrew Keen, Keith Teare
Speaker 1
is probably the main story. As we speak today, Bitcoin is above $42,000. And the Graycroft, what is it? Gray something, Graycroft, I think, Bitcoin and Ethereum stocks that trade on the NASDAQ each up close to 400% in 2023.
Speaker 1
So, you know, they were down by a similar percentage two years ago. So the volatility of cryptocurrency and the lack of underlying product is probably, continues to be the main dialogue. Solana, which is the cryptocurrency that the All In Podcast guys invested in, is up to over $100 now. from a low of below 20, even below 10, I think, at one point. But there's no real products yet built on it. So I think the promise of crypto to be the infrastructure for a new financial world, I think that's still alive, but no one's yet done anything to show it.
Speaker 2
Well, that might, Keith, have been what you would have said about AI, this time last year. But AI, everyone's always talked about the promise. It's always we're on the brink. It's the old joke I always make about the Brazilian economy, always on the horizon, never here. Yeah. Can I mean, there's been so much said and written and spoken about AI. Is 2023 the year that AI became real? Or do you think that perhaps that might even in itself be a mirage, might be misleading.
Speaker 1
No, I think what will happen, and just to make one prediction about 2024.
Speaker 2
Well, we're getting to predictions of 2024 after the break, so focus on 2023 for the moment.
Speaker 1
Well, I think what has happened in 2023 is artificial general intelligence has- AGI, otherwise. Yeah, AGI, has been outed as a false god. because something that is not artificial general intelligence, open AI, is good enough to support a very large number of human activities that previously would have had to be done manually and personally. So I think the debate has shifted from is this AGI to look how good this can be. and there are a few outliers like Gary Marcus saying but it isn't AGI and of course they're right but it actually doesn't matter anymore whether it's AGI, it's still good.
Speaker 2
Yeah and whether we'll ever get to AGI is a bigger issue. What about the personalities of 2023? I mean Sam Altman It's hard to argue that he hasn't been the dominant figure in tech in Silicon Valley, perhaps even more broadly in a cultural sense in 2023. How will we look back at 2023 in the context of Sam Altman's achievements and tribulations?
Speaker 1
I think Altman shares something in common with Elon Musk. There was a couple of stories this week where His leaving his previous gigs was described, I think it was on the information, the Y Combinator CEO ship. And the accusations against him are that he's too self-centered and doesn't collaborate enough. Which of course, if you say that sentence you immediately think of Musk. And I do think that people who achieve things tend to be like that. They're super driven by something in their head, and they don't let things stand in their way. He is definitely one of those. And the way he handled being fired, and turned the whole thing around within 48 hours, I don't think was masterful. I think it's just like the natural instinct of a, you know, an operator with a vision. and there are a lot of people like that who tend to be the high achievers. We live in a time when determination and self-focused vision is frowned upon as being nasty or unkind and the atmosphere we live in is this atmosphere where the more you are a victim The More People Like You. And Musk and Altman refuse to be victims.
Speaker 2
What about Musk? I think Musk has his own following of fellow victims. One thing I'm sympathetic on Altman is a lot of the criticism on his previous job is he misled people. But you know this as well as anyone, Keith, when you're a CEO, there's a very thin line between driving the future, inventing the future, misleading people. You have to tell a story and that story when it comes to startups is always euphemistically is exaggerated and isn't that fair?
Speaker 1
Yeah my experience of that is being accused of changing the story as if a story has to stay the same and if you change it you know you're doing something criminal The whole nature of startups is the story has to adjust to your most recent learnings. And stories do change. They don't tend to change wholesale. It's usually nuanced change, let's call it. I would say that's happened at SignalRank in the last year. Our story is definitely, you know, word for word, not the same as it was a year ago. But it's essentially the same story. And so I think anyone who is in a leadership position has a responsibility to the truth. And the truth doesn't always align to the story you were originally told. Does that mean you were misled?
Speaker 2
I mean, as a founder, like an Altman, or even a Musk, you're creating your truth. So yeah, there's a lot of movement. It's like those People on the street with their card games, it's sometimes hard to watch their hands or know what to look at. I agree with you on Altman. I'm not sure. I mean, you're a big fan of Elon Musk. I have to admit that I'm not convinced that 2023, and I don't know what you think, 2023 was a very good year for him. I think he's shamed himself at X. And I mean, maybe he's achieved stuff on other fronts. What kind of year do you think 2023 has been for Elon Musk?
Speaker 1
Well, he's done a few things. Buying Twitter certainly put him into a new situation vis-a-vis everyone else. He's the owner of probably one of the most important media properties in the world, arguably the nervous system of news. And that, of course, means he made lots of enemies. The second thing he did is he stated an opinion that he wanted Twitter to be available to people he personally strongly disagrees with on the right, even the far right, a little bit like Substack did last week. And that made even more enemies because we live in times when the left want to de-platform anyone that disagrees with them. And Musk is saying, no, we will platform everything.
Speaker 2
So you're arguing that Musk is a hero of the left?
Speaker 1
Musk is a hero of free speech for sure which means he's hated by the left because the left by the right the right like him more because they're the ones being censored so the left arm being censored is only the right who is censored for the most part yeah i mean i can't think of anyone on the left that's been censored can you
Speaker 2
I can think of many, just think what's happening on the Palestine issue, Gaza, I mean half these people who open their mouths on American campuses are being shut down. That's another issue, it's not a tech issue. I have to admit I strongly disagree with Elon Musk, but maybe we'll do a whole show on Elon. Finally, when it comes to 2023, one thing that I think you must be thrilled with is it was a very bad year for your friend at the FTC, Lina Khan. According to the NY intelligence, she was expected to break up big business. but apparently she's broken the agency. You talked about 2023 being the year that America and China made peace, economic peace. Is 2023 the year where Kahn or Kahnism at the FTC failed?
Speaker 1
You know, I think this is not a win from my point of view, just because it's so obvious that she is not pursuing an agenda that could work. Her agenda is defined by an attempt to expand antitrust law into things that it would be super hard to define as antitrust. So she's trying to control big tech when big tech isn't breaking any rules. And the courts still in America are, you know, We're speaking with Keith Teare, the CEO of SignalRank, the author of The Invaluable That Was The Week newsletter on Substack. Keith,
Speaker 2
You mentioned Lena Kahn and AI and copyright. One of the late stories, late breaking stories, if you're doing a regular show, I'm sure you would discuss it, is the news this week that the New York Times is suing OpenAI and Microsoft. A couple of questions on that. I mean, this could be perhaps the big story of 2024. But how does this connect with your observations about Kahn? Is it separate?
Speaker 1
It's as desperate by the New York Times as Khan is when it comes to antitrust law. It's an act of desperation. It almost certainly can't win.
Speaker 2
Well, why is it desperate? The New York Times is a profitable business, has a huge subscription base. What's desperate about it?
Speaker 1
Well, because they're so smart, they can, to some extent, see the future. and The Future Doesn't Look Like The Past.
Speaker 1
because there's a paywall and their subscription business has been super healthy for the last five to ten years because of that whole underlying infrastructure. The future, somebody says to ChatGPT what's going on in Ukraine and we'll get back a curated answer drawing on all kinds of sources and we'll never click on a link and we'll never go to the New York Times and we'll never need to subscribe to it. So the entire underlying ecosystem that has led to their subscriber growth is gradually being eroded by some new behaviors. We can assume that those new behaviors will grow in 2024. Of course, search isn't going to go away, but the balance between search and curated answers from AI is going to shift in favor of the latter. And that threatens the New York Times. So the question they have in their boardroom is, how are we going to get paid? And their point of attack is to go to the largest beneficiary of this new era, OpenAI, and try to get it to pay them some money. And so this court case is really a negotiating tactic for a payment. Even if they win, which I don't think they will, But if they won, the payment wouldn't replace their subscription payment. So their best bet is to build a New York Times version of OpenAI that does a better job than OpenAI could of indexing their content and charges subscribers. I'm a subscriber to the New York Times. I don't know how much I pay monthly, but I think it's more than $20, which I pay to OpenAI. charges for access to the best AI version of the New York Times.
Speaker 2
So what you're saying is that New York Times should have done a deal like Axel Springer did with OpenAI rather than take them to court.
Speaker 1
Axel Springer's deal is more like being integrated for money. I'd say that's plan B. Plan A would be to evolve the New York Times into an AI where the underlying writers are feeding it content to its model to get trained on and make that content proprietary. And if it's good enough, people will use it. If it isn't, they'll use open AI. It has to be better. And one would assume that if the content is better, the AI will be better, will hold up. But is it better enough to justify paying for it separately? That's the big question. I think I'm a skeptic. I think open AI probably is good enough. And I think Learning on publicly available information will end up not being a crime.
Speaker 2
We will talk much more about this in the new year. I do want to get to your predictions for 2024, even if you don't want to make predictions. But we always end our show, as regular viewers and listeners know, with Start-up of the Week and X of the Week. And it seems like there's only really one answer to this, Keith. What is the start-up of the year in 2023?
Speaker 1
It has to be OpenAI, obviously. It can't be anything else because it now controls the chessboard. If you could think of tech as a massive chessboard. OpenAI controls the chessboard. It controls the pace and the cadence, the recent changes in the structure of the board and the board's ability to override the product Timeline, taking safety into account is the entire, you know, nexus of that. So I think it's OpenAI. You would wish it could be one of the nuclear fusion companies, because I think nuclear fusion, if it were to be real, would have a much bigger impact on humanity, even than OpenAI, although together, they'd be profound together as well. But I think it's open AI. I think nothing else has come close to delivering on a promise that will impact so many people in a good way.
Speaker 2
Yeah, you and I completely agree on that. According to OpenAI's website, they are creating safe AGI, artificial general intelligence that benefits all of humanity. We will see certainly, that will be a story that will run and run and run.
Speaker 1
One thing I'd say, Andrew, is there is a reasonable narrative that has to run alongside it, which Altman does engage with this narrative, which is who's going to benefit? They talk about benefiting all humanity. And there are certainly many narratives in which that ends up being true. But there's also the narrative that, you know, the shareholders of open AI become super rich and the rest of us become serfs to the technology. And so the social implications of open AI, I would like to see that talked about more than whether it's AGI and whether it's dangerous.
Speaker 2
One particularly rich owner of OpenAI is Vinod Khosla. I think he was the earliest investor, certainly earliest outside investor with Reid Hoffman. He had a fascinating quote, I think, and it touches on the New York Times case against OpenAI. And Khosla wrote back in October, and we featured this, to restrict AI from training on copyright material would have no precedent in how other forms of intelligence that came before AI trained. There are no authors of copyright material that did not learn from copyrighted works, be it in manuscripts, art, or music. You can't separate Gauguin's influence from Matisse, Velazquez from Dali, Picasso's from Pollock, Beyoncรฉ's from Taylor Swift, nor Charles Taylor's from Yuval Noah Harari. The list goes on. It's from a broader post that that he had in Medium. I'm not sure I agree. In fact, I don't agree. But I think it's an important X, isn't it, Keith, that he's laid out his argument, his story for the coming war.
Speaker 1
Look, it's self-evident to me that you cannot copyright human knowledge. And in fact, it's really hard to assign human knowledge to a single creator. I think there is no such thing.
Speaker 2
So are you arguing against copyright law in itself?
Speaker 1
You're saying anyone should be able to... Well, no, you can copyright a manuscript or a form of words. Or a song. So that it can't be plagiarized without attribution. I think you can do that. But what The New York Times is asking for is way beyond that. It's basically saying if we wrote about it, you can't train on it. And create derivative thoughts or even new knowledge based on ours. Well, that on the face of it is ridiculous.
Speaker 2
I take your point, but let's just focus on the X from Koestler. Is this the best argument? Is this the Silicon Valley argument? Andreessen would, of course, agree. I assume that the lawyers at OpenAI will also agree. Yeah.
Speaker 1
I think it's the height of hubris to believe that you are the sole source of a piece of knowledge and no one can train on your knowledge because you created it. You've got to be a bit more humble. It's very likely you didn't create anything.
Speaker 2
Well, let's have a little bit of hubris now rather than humility, Keith. You are our authority on 2024. I know that we can't see the future, although you argued that the New York Times are so smart, they can see the future. So some people see the future more clearly than others. What should we be? I know it's hard to make predictions and especially very concrete predictions, but what should our viewers and listeners be looking for in 2024?
Speaker 1
Well, I think the whole idea of creativity It will be renewed with some new meaning. What these new tools give us, and by tools I'm thinking not only of OpenAI, but the video making tools, the image making tools, the music making tools, what they give us is the ability for a single individual to create new things without needing to build teams. And the entire past hundreds of years, innovation required teams, large teams, sometimes expensive teams. What you now have is the ability for a single individual to use tools to do what previously teams did. And that should create a huge release of human creativity, like the likes of which even YouTube would feel dwarfed by. And so I think that the signposts, things like Substack and YouTube that come from the past, when married to open AI, you know, stability, anthropic, Dali, mid-journey, some of the new video creation tools, there's some amazing trailers made from screenplay scripts that emerged last week that you would think were made in Hollywood. That's going to unleash the individual on the stage of history. Now individuals always been on the stage of history, but usually they've had to build teams using often flawed human beings to work with them that dilute their vision. Inevitably, because they're not the same. That is going to be changing. Andrew Keen's Keen On can truly grow in all kinds of directions just with Andrew Keen using the tools of the day. And as you know from doing Keen On, that idea of needing a team is an ever-present, you know, source of friction because teams cost a lot of money and you've got to be sure that you're going to make money. And you've got to talk to other people. And you've got to talk to other people. So that goes away.
Speaker 2
So are we seeing then... We know that there are companies like Anthropic and OpenAI. We know there are companies like YouTube and Substack. Are you saying that what we will begin to see in 2024 and onwards is the combination of YouTube plus Anthropic will equal something that we haven't seen before?
Speaker 1
Yes, but it won't be a central destination, it will penetrate everything. It'll be in everything, that combination.
Speaker 2
So it's like Google in that sense, or what Facebook wanted?
Speaker 1
Yes, but even more it's like the internet, you know, we don't the internet isn't a company The the internet is all companies Leading to a technology we all experience in all kinds of different ways. It's closer to that It's going to be an all-pervasive underlying set of tools that are available everywhere and will manifest everywhere including when you go to the movie theater when you turn on your TV when you listen to a podcast and Those tools will be everywhere.
Speaker 2
Wow, that's exciting. I'm not sure we'll only see the earliest manifestations of that in 2024, won't we?
Speaker 1
Yeah, I think 2024 is the co-pilot stage of that. Imagine when you get to the point where the tools are clever enough to build things themselves and talk to each other.
Speaker 2
Well, it's the old question. This is the Gary Marcus question about AGI. What happens using Keen on when they don't need Keen? I mean, that's what you're suggesting.
Speaker 1
No, they're still being modelled on Keen. There's a new tool that came out, I saw it yesterday, where you can train an AI with your voice just by speaking.
Speaker 1
and then you can take a music track and separate the music from the voice and then tell it to do the voice in your voice and you can do that with any song in the world and basically you know the song becomes you become the singer with your voice and you know that that gives you just a small glimmer of what you know what the future will look like totally separately, you can create music now on Google. You can basically create music from text. So once you create music from text, and you can add your voice, and it sings in tune, you know, somebody is going to make a hit from that. And you won't care that they use tools to do it, because it'll still be good.
Speaker 2
This isn't very surprising. I mean, the future of you and I have talked endlessly about the future of creativity. Is there anything on the horizon in 2024 that might surprise us?
Speaker 1
I think nuclear fission would be the biggest, sorry, fusion would be the biggest surprise. There's a lot of talk about it. There's a lot of talk that it is worthy of short term investment because it's almost real.
Speaker 1
I don't know enough. It's one of the few areas where I'm naturally a skeptic, which is not like me, but it's mainly through ignorance. I don't really know enough. So that would surprise me in a good way as well, if that was real, because I think that would transform energy production on all scales, from how a car gets powered all the way through to how the world lights up at night.
Speaker 2
And finally, Keith, 2024, of course, is going to be the year of the American election in November, which will, for better or worse, shape the year. Silicon Valley seems quite ambivalent or indifferent. No one seems here to care much about Biden. No one loves Biden very much. and Trump isn't taken particularly seriously. Can the election have a big impact on tech or will tech? It's had a great year in 2023 that the big tech companies are driving not just the US but the world economy. Is it conceivable that everything will go well and Trump will be re-elected or even worse there'll be some sort of civil war or unrest?
Speaker 1
Well, I think tech Tech is a kind of a tool that sits on top of ideas. So it can take the ideas of the time and it can manipulate stories and such. But I think the source of people's ideas that lead to election results is more deep in their experience of life. And America is going through a hard time. And so there's a lot of dissatisfaction, even though Biden Andrew Keen, Keith Teare titled That Was The Year
Speaker 1
then it's likely that the incumbent will have a hard time everywhere. And I think that's the source of it. I don't think tech causes that. I think tech is a tool used by each side to try to win arguments. But ultimately, you can't win an argument against people's human experience. That's what will drive everything. The only way of understanding Trump having any level of popularity is that. It couldn't be explained rationally or logically based on his ideas.
Speaker 2
And finally, Keith, two days to New Year. I'm feeling particularly generous. So let's end with your own particular startup, SignalRank. It's done well in 2023. Where would you like SignalRank to be this time next year? What do you want to achieve with SignalRank in 2024?
Speaker 1
Well, Single Ranks already succeeded against its primary goal, which is to be able to predict future valuable companies early in their life using a scoring system that correlates to future outcomes. We've already succeeded and that works. We've also succeeded in finding a group of people who need that, which is early-stage investors who want to keep investing in their best companies later on. And to do that, you need to know, well, which are the best companies. And that's seed investors. We have over 200 partnerships. So we've already succeeded there. The next stage is the hardest stage, which is we have to persuade investors in venture capital that they would do well to give us their money and it will perform better than if they give it to late stage funds. And, you know, so 2024 for SignalRank is all about raising capital. We just, in two days from now, we'll close the most, the latest round. We raise capital constantly, by the way. It's not like we raise a fund and then we stop. We're constantly raising, we sell shares. So we just closed the December raise. There'll be another one between January and March and every 90 days we do this like rolling raise. The average raise in 2023 was about $3 million per quarter. We want to shift that to closer to 10 in 2024 per quarter, and then in 2025, closer to 50. And so it's happening in the context of the worst funding environment I've known. So the fact that we've done it at all is a credit, I think, to the team. and let's see if we can step it up to a higher level. It's all about capital now because the rest works.
Speaker 2
You better go and raise some money, Keith. Absolutely.
Speaker 3
I've got sunshine on a cloudy day
Speaker 3
When it's cold outside, I've got the money for you I guess you'd say, what can make me feel this way? It's my girl