Feb 15, 2025 ยท 2025 #6. 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
Vance AI Speech: A Breath of Fresh Air
Human Transcript
Timed transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Speaker 1
Hello, everybody. It is Saturday, February the 15th, 2025, a day after Valentine Day. It's been a day or a week dominated by a certain J.D. Vance. Yesterday, he made a very controversial speech in Munich. which apparently laid bare the collapse of the transatlantic alliance. He attacked Europe over free speech and migration. So he's not the most popular fellow in Europe. And a couple of days before that, he spoke in Paris at the AI Summit, a classic podcast. Parisian event, talking about summits. Macron, of course, also spoke there. According to the Wall Street Journal, Vance's counsel on AI to the rest of the world was good. The Journal, of course, being a conservative newspaper. According to the Washington Post, which is a progressive newspaper, He pushed the America First AI agenda. Others, like Fast Company, ask what to make of Vance's speech at the Paris AI conference. According to my friend Keith Teer, the author of the That Was The Week newsletter, the speech was a breath of fresh air. It's once again Keith's Marxist foundations coming out. I was going to call you Marx, Keith. That would have been a true Freudian area. What do you admire about Vance's speech? Why is it a breath of fresh air?
Well, it's in the European context, that breath of fresh air. I mean, I think from an American's perspective, he didn't really say anything new. We already think of AI in the way he expressed it. But in Europe, the dominant discussion around AI still is all around safety. That is to say, AI is dangerous. We have to control it. We need to regulate it. And as a result of that, most of the American developments in AI are not even launched in Europe because in order to be made available to citizens, it has to go through various regulatory layers and that slows everything down. So in the context, Vance stood up on the platform in front of all of the people doing that regulation and told them, he basically rubbed their noses in it, saying how self-destructive their approach was for European success. His opening lines were, I'm not here to talk about AI safety, I'm here to talk about AI opportunity. And in the days since, there's been quite a big reaction in Europe to the speech, mostly positive from normal people and adjusting policy at the regulatory level. So it's quite a profound moment. carried himself very well. I mean, he was articulate, thoughtful.
Yeah, you say his speech marks a crucial inflection point. I wonder, though, if France was so self-interested as a MAGA person, why would one even to encourage Europe to develop AI? I mean, why not just... let it be like social media or the internet, where American companies dominate. Is there anything in America's interest, MAGA's interest, the Trump-Vance-Musk alliance, that would benefit from strong European AI companies?
Well, from strong European AI openness, yes. I don't think Vance thinks for a minute, there are any European AI companies that will be able to compete in that open environment. And so most of his purpose is economic. He's basically saying open up so that our guys can sell stuff to you. And the money will flow back to the US as it has done with Amazon and Google and every other major tech innovation in recent years. So it's basically an economic speech masquerading as a policy speech.
I wonder if there's... an opportunity for Europe, given the clear divisions now that exist between the US and Europe. I wonder whether there's an opportunity for Europe to start looking more sympathetically at Chinese AI companies. Did Vance, in his speech, did he warn Europe about turning to the Chinese, the other partner potentially?
Yeah, there's two parts of his speech I didn't really incorporate in the editorial. The first was a subplot all around China, which he didn't name, but he called dictatorships. We don't want dictatorships leading in AI. And then there was another subplot which was all about free speech and openness and not censoring, which was aimed at the Europeans, of course, and the Chinese. And so I do think Vance and Trump will go very far to allow free speech to the distance.
I think the funny thing is, in order to be consistent, they're going to have to allow all free speech. And they will, because they know that. And so weirdly, the Republicans become the free speech party, which makes no sense historically. But you know, is happening. And so I thought there was a lot of interesting things in that speech that symbolize a very confident America. However, the reason America is doing this is because it's weak, which is a paradox. Politically weak or militarily weak or economically weak. not militarily, it's super strong, but economically, it's relatively declining against China. It's the next Europe, America's the next Europe, China's the next America. And in that context, American, I don't know if you'd call it confidence, but brashness, it's Sounds positive to our ears and to mine as well, because it's pro-optimism, pro-progress. But actually, it's coming from a place of weakness, which you see in the tariffs and the anti-Chinese stuff.
And I wonder, coming to the Munich speech where Vance was pretty clear, Trump's always been clear that if there is an opportunity for Ukraine, Ukrainians have to offer American, so to speak, access to its raw materials, minerals, etc.
Yeah, exactly. By the way, one of our commentators, David John William Bailey in LinkedIn is saying, you need to explain this a bit to us, David. He says he absolutely is. He's also attempting a trillion dollar mob style shakedown. Anyone defending this is either deluded or or only reads hard right propaganda.
That's what he needs to explain. Put a second comment there, David, and give us a bit more context, because it's not obvious what you're saying. But if you're accusing us of being hard right or being deluded, obviously, we're not going to agree with that. But we're totally happy to take it on board and explain why not.
Well, but Keith, you always claim to be a progressive, you always claim to be a man of the left, you have a background in left-wing communist activism. Now you're on board with Vance. You were on board the week before with Musk. You're ambivalent about Trump. What does this say to you? What does this suggest about you personally? Or is the reality of politics these days that the supposed conservatives like Vance are actually progressive in their own way and the supposed progressives on the Democratic Party are actually conservatives?
Well, as you know, I don't like those labels anymore because I think they're trying to fit a modern narrative into an old set of boxes. I think, broadly speaking, Vance is an economic progressive. He wants the economy to grow. He wants GDP to grow.
Yeah, but not very many people can do it. So I think they really are serious that they believe innovation in tech and GDP are correlated. And I believe, Keith Teer believes, that GDP and social good are correlated. And so if you really want to be a progressive that wants people to have a good life, you have to support economic growth. And I think Vance does, and I think that's what his narrative is about. And he's basically telling Europe that they're going to get the opposite, which has been true, by the way, now for a decade. European GDP per capita is as low as $35,000 a year. American is $85,000 a year.
Yeah, it's astonishing shift. And this is going to be remembered, I think, as an important week in America, in the American-European relationship. You said that the aftermath of the advanced speech has been remarkable and telling. The EU dropped its AI liability directive. The UK rebranded its AI safety institute. Open AI removed diversity commitments. So our speech is now having an impact and particularly this Paris speech advanced when it comes to AI policy, both particularly in Europe, but also in the US as well.
Yeah, I wouldn't. give too much credit just to the speech. I think the speech is symptomatic of a lot of a lot of zeitgeist change. And everyone's getting in line with a new zeitgeist, which is tech is good. AI is good. censorship is bad.
Well, I don't know if that's I'm not sure I would call that the new zeitgeist, Keith. I mean, you're talking in Palo Alto, where there's always been a zeitgeist. I think if anything on in universities, book publishing, the reverse is true.
Yeah. So I'm an avid MSNBC watcher. I watch Morning Joe every morning with Mika Brzezinski and Joe Scarborough. And so I'm kind of imbued with the liberal narrative compared to what's going on and what's happened is a very rapid change from the days after the election when the liberal narrative was we need to look at ourselves has now become a narrative that the judges have to save us from the administration the administration is not democratic even though it was elected. And we got to rely on judges because there's no one else to rely on.
Yeah, I mean, we don't want to about politics doesn't mean the zeitgeist is shifted. It just means that the zeitgeist is shifted folk. I mean, the people on one side have shifted focus, but they still are not sympathetic to Trump, Vance, Musk.
One piece that you pick out, which I thought was interesting, is from somebody called Elizabeth Yin. It's nice to have a female author, Keith. Too many of our authors are male. Maybe I'm being too woke here, but the AI takeover, according to Yin, no one's jobs are safe. This isn't exactly news, is it?
No, she's really summarizing what we've been talking about on That Was The Week for quite a while. But I thought it was a good summary. And she gives some kind of prioritization. There's a section that talks about regulated professions, human centric jobs, creative and entrepreneurial jobs, energy and infrastructure and distribution. And she then breaks down where she thinks the main impact of AI is going to be. She kind of leaves it. where you kind of want more from her because she doesn't thoroughly go through all of these. But she's a VC. She does early stage investing. She's very good. And I thought it was.
And the one thing she says, which I don't think anyone's going to disagree with, fewer workers, more AI. I was at an event this week in San Francisco. where there was a panel with some VCs and entrepreneurs on exactly the same question she's asking, where are the cuts going to come first or what sectors are going to be most dramatically affected in the short term? And people sort of, people weren't entirely clear, but the one area that comes up is healthcare seems the most, that's the lowest hanging fruit at the moment.
Yeah, there's a funding event this week from a company that applies AI to biology, specifically cancer, programming anti-cancer cells. So you're going to see AI in everything, and that will lead to an acceleration of invention for sure, because the individual, who is still really important, by the way, there's another article about that this week, the individual now has an army of talent in AI able to help them make progress. It just speeds everything up.
Yeah. So what other AI news you, in the summary, there's a couple or three pieces on Anthropic. I use Anthropic. I like it. They're growth soars to 34.5 billion in 2027 revenue. That's, of course, speculative. And they announced their next major AI model could arrive within weeks. Are anthropic competitive with OpenAI, Keith?
Yes. And they're better than OpenAI at some things. They're already better than OpenAI at coding. and I use them for coding, that if you put it in context, those three anthropic pieces sit alongside the Google piece and the OpenAI pieces, and what it tells you is, The last couple of weeks have seen a major acceleration of product roadmaps and plans, mainly in response to the DeepSeek news, I think. Yeah,
it's interesting that, I don't know if DeepSeek was a one-week wonder, but there are no headlines, at least from you, on DeepSeek. It seems to have stimulated change, as you suggest, rather than change it in its own way. And then your Google piece is interesting. that they're rolling out a new memory feature for Gemini AI, allowing recall of past conversations, which is increasingly getting to the point where these AIs, if not human or sentient, they certainly are able to remember things and have conversations with it.
Yeah, and that becomes much easier once you go from LLMs to LLMs with agents. An agent is... a piece of software that speaks to an LLM to complete a task. And so you can, you could have in software, a memory agent or a recall agent whose only job is to say, has this question been asked recently and what did I know last time? And bring it into the context for whatever the current question is. And I think we're gonna see more and more of this. I've spent most of my week building a multi-agent system for my company, SignalRank. I have a question taker agent that you ask a question of. it then farms out to a database agent or a chart drawing agent or, you know, a reasoning agent. They all have different jobs and they come back and they give their answers to the original agent and then it gives the answer to the user. So this collaborative process, agents kind of concept is becoming very real now. And memory is one of those. I think perplexity is the most advanced.
Yeah, we were talking about perplexity before we went live. You convinced me I use anthropic, but you said for me it's probably wiser to use perplexity where I still have all the access to anthropic, but it adds a layer and some more intelligence. As I said, I was at an event this week where one of the The venture people from OpenAI was there who talked about Sam Altman's projection that in the not-too-distant future, there'll be billion-dollar individual startups. Are you suggesting, Keith, that that's not that far on the horizon, given the power of AI, that individuals can do the entire startup without needing the help of anybody else?
It depends on the startups. If the startup is mainly software, that's probably true. But if it needs, you know, account management and billing and all the other things...
I think right now it's the hard bit. Like reconciling invoices to receipts. LLM's not very good at that. So I think it's coming with... Two things, reasoning agents and then agents that can use tools to follow, you know, do actions, if you will. So it's coming and it's probably coming this year and it will accelerate. So, yes, it will get there. I think the headline of a single founder of a billion dollar company is just a headline. I mean, but it's directionally correct.
I mean, it does reiterate Elizabeth Yin's point that no jobs are safe. I mean, in finance, in HR, in coding, in content. I mean, I'm using it more and more to summarize these conversations. I don't need a large editorial force. So clearly dramatic change. And in fact, your startup of the week, Keith, the robotics startup figure is in talks for a new funding, almost $40 billion valuation hardware company. Does this speak of the reality of this new AI revolution? It's not just theory, it's practice now.
Yeah, well, so figure, it's gone from 2 billion to 39 billion in less than a year. And why? Because one of the major car companies has signed an agreement with it to have these robots on production lines in its factories. And the startup of the week, by the way, is Aptronic, which is a different humanoid robotics company, also raising a lot of money, but slightly earlier in its journey than Figure. And I think it's... Oh, I made a mistake there.
I have to add. I thought it was Figure, so that's my error. And AI certainly wouldn't have made that error. I'm going to add an Aptronic product. image to this.
Continue talking, I'm rather embarrassed. You've probably already got one. I think I think I remember you having one. That said, they both speak to the same truth, which is AI is going to manifest itself in the physical world in the form of humanoid robots sooner rather than later.
Yeah, and that was another of it. Tim Draper was one of the speakers at this event I went to in San Francisco. I know he's an investor in your firm. That was his big prediction. So we have an apptronic image now building robots for humans. So are they just a kind of an earlier version, Keith, in some ways of figure?
An earlier version, possibly more advanced in concept because they started later when there's, this software gets better by the week. So the later you start, the more advanced the software is that you can leverage. So we're not gonna see an end to this. There's gonna be a lot more of it. I think humanoid robots are really interesting because the physical world is built for humans.
you know, steps, ladders. Currently, but I'm not sure that would be the case, especially when it comes to, say, self-driving cars and roads. That's going to change as well, isn't it?
Well, you still have roads because they still are.
Words and timings
Well,youstillhaveroadsbecausetheystillare.
Speaker 1
Well, you still have to have roads, but I'm saying the roads themselves will become more and more suited to self-driving cars as opposed to human driving ones.
Well, I'm sure the Chinese will do that, not the Americans, not even in San Francisco. Meanwhile, there is still lots of tech news. Still, there's this open feud between Elon Musk at OpenAI. Sorry, that was another Freudian error. Sam Altman at OpenAI and Elon Musk. Musk this week had a bid to buy OpenAI for around $100 billion. Is this just sensational? meaningless stuff keith is this froth or is it meaningful in the long run the musk
altman yeah did uh fight well the specifics of this are super interesting because it's very clever of musk what what musk is offering to buy is not open ai he's offering to buy the not-for-profit part of open ai now altman is trying to put a value on that not for profit because he wants it to go away or at least be subsumed. And he's trying to do it at a very low valuation so that the stakeholders in the not-for-profit don't get much. So Musk put a super high price on the not-for-profit to force the board at OpenAI to put a proper value on it as it transitions or to stop transitioning it, one or the other.
I think if I was on the board of OpenAI, I'd be very worried. They rejected his offer yesterday, by the way, but that will not be the end of the world.
What is Musk doing? Is it just because he hates Altman and he's annoyed that he was one of the co-founders and he's no longer involved? Because if he does indeed do what he seems to want to do, which is weaken, even undermine OpenAI, I mean, the real winners are probably Anthropic and Google, aren't they, rather than Elon Musk?
Well, because it's Musk, there's more than one thing going on. He has economic interest in XAI for sure. He's also really pissed off with Altman because he considers that Altman basically stole the OpenAI idea from him which is not really true when you get into the facts, but he believes that. And not only that, but lied by making it not for profit and then turning it into a for profit when he promised he wouldn't. So Musk basically feels like he's got the moral high ground and that gives him the energy to fight. Altman is clearly tired of the whole thing. He's just trying to do what he's trying to do, you know, and having a light shone on him.
So this is the first time you have articulated some concern about OpenAI. You've always been quite bullish. Are you suggesting that your bullishness in the past is changing a little bit?
I don't think so, because I think this is a bit of a sideshow. The biggest news this week about OpenAI is the decision to abandon their O3 model, not abandon it, but incorporate it into a new GPT-5 later this year.
It would do reasoning, it would do operational stuff, actions, and it would do what LLMs do, including being capable of video and image production all in one, and probably will retain its position as the best across all of those different things. So I don't see that anything bad is going to happen to open AI. I do think Musk can be an irritant and it could force them into corporate decisions about valuation and merging their different components that aren't to their liking. That could happen.
Well, there'll be a lot more of that this year, Keith. Interview of the week you were kind enough to include in this week's newsletter is with Greg Beto. Most people won't be familiar with Greg Beto. He's a tech writer, journalist based in the North Bay, San Francisco. But he's also the co-author with Reid Hoffman, who everybody knows, of a new book called Super Agency. What could possibly go right with our AI future? And from a progressive point of view, it's optimistic about AI. So I guess Hoffman is one of the few progressives, Keith, who actually is optimistic about AI. Is that fair?
Yeah. I mean, he really represents that part of the liberal spectrum that was in the New York Times article last week suggesting the Democrats should embrace technology and innovation. And the book is symptomatic of that. I didn't have a chance to listen to the interview, Andrew. Give us a flavor of what he said.
Well, it's a standard. It's like listening to you. He believes that it actually... I mean, he's not quite as... extreme as you, but he believes that this progress will ultimately benefit. He distinguishes himself. Beto, I thought, created some light for the space between him and Hoffman. I think he sees Hoffman as being slightly more optimistic than him. But it's about super agency. You and I have talked endlessly about agency, about humans being able to shape their lives. And of course, that's the big debate for the critics. It's the AI that will shape us. For the optimists, AI will enable us to shape the world. It's an age-old argument, and it's not going away. Another figure on the left, if that's still a term that means anything, is Albert Wenger, a good friend of both of us. He's your post of the week, and he comes back to the J.D. Vance speech. He says, praising this speech by Vance is mistaking jingoism and wishful thinking for true global leadership with a real vision of AI and humanity. I'm assuming you don't agree with Albert on this.
Actually, I do agree with him. I think I wanted to take a positive view of Vance's speech for its optimism. In the context of Europe, it was a great speech. Albert's right that the American framing is entirely jingoistic and AI isn't. AI is entirely global and humanistic. So there is a contradiction between a declining superpower being a champion of progress for its own nation versus what Albert would prefer, which is a leadership that is truly global in nature.
Yeah, it's interesting that the first comment on Albert's tweet or X, whatever you want to call it. There was someone called Beth E slash ACC who says this may be the most E slash ACC speech of all time. I didn't know what that meant. And it meant effective accelerationism. Are you familiar with this term, Keith?
It's interesting you say they're the same thing because I don't think anyone thinks that, but I think you might be right. But as long as you put them in the right order, I think if you get acceleration and growth and value, you're going to get a better life.
And even Albert acknowledges, like you, that there are aspects of the speech which, in your language, are a breath of fresh air. He said, the only good point was the clear pushback against regulatory capture. Is it going to be effective? I mean, is it clear that the days of Lena Kahn are over, Keith? Are we the end of the period of regulatory capture, whether it's in Europe or the US? As you say, one of the consequences of the speech was that the Europeans have taken a step back from regulation.
much older and perhaps less powerful figure especially in trump's america i mean what can warren do is she can talk a lot and get people annoyed but she can't actually do anything whereas lena khan actually controlled regulatory capture i mean she was the head of the fcc
exactly but i think i think warren will be i mean i'm i find her intensely irritating um it's amusing to me that musk is asking how her net worth went from two hundred thousand dollars to you know double digit millions and it's because she gets subsidized by pharma because she's pro vax um and so she's plugged in wow
that's that that's a controversial observation i mean you said anyone who's anyone who gets supported by Big Pharma is Provax. Are you anti-vax? Does that mean that anyone who's anti-vax is not going to get the money? I mean, most of us are Provax.
Well, that's not a very generous interpretation, although it does suggest that when you give Elon Musk the keys to the Treasury and find out exactly... and the IRS, then all these things are going to get revealed. And we should end with another interesting X from Albert, which I think gets to a lot of this. He said, if you're young and capable and care about democracy, you should work for Doge. What do you make of that? I tend to think he's right.
No, I'd use the Obama example. Obama wanted to get a really good health care plan. And as soon as he was in office, he made speeches saying, I won't be able to achieve what I want to achieve unless you, the people, are kind of on the streets, because Washington is averse to change. And it turned out that he had to make all kinds of compromises, and he ended up with what today we call Obamacare. But his experience was an experience of being blocked. And Trump basically has been through that himself. We're probably mostly thankful for that, based on his first administration. He now is wiser and older, and he's not prepared to let the bureaucracy stand in his way, and Musk is his weapon. And there is something positive about a better, cheaper state, and more democratic if the elected people can do what they said they were going to do.
Yeah, and bring the expenses down. If you're young and capable and care about democracy, you should work for Doge. Wise words from Albert Wenger. We will return to all these themes, Keith, in the future. Have a good week, and we will see everybody again next week. Thanks so much. Bye, everyone.