Aug 30, 2025 ยท 2025 #32. 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
Users, Publishers, and AI: Everybody Wins
Human Transcript
Timed transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Andrew Keen
Hello, everybody. It's Saturday, August the 30th. I probably won't put this up, actually, until Sunday the 31st. Keith will put it up on That Was The Week. Today, it's the last weekend of the summer, silly season, not a lot of news, in my view at least, although Keith may disagree, which is why he's put together, in my view at least, on first sight, a slightly absurd headline with lots of people smiling. Users, publishers, and AI, everybody wins. We've heard that one before, and usually when we see everybody wins, it's everybody losing, especially... publishers and users. But Keith, I'm sure you will disagree. Although you have to acknowledge that we've seen these sorts of headlines before in Web2, for example, about users and publishers all winning and they ended up losing. So why is it going to be different this time with AI?
Are you disputing that that is a possible outcome?
Words and timings
Areyoudisputingthatthatisapossibleoutcome?
Andrew Keen
Well... I would be amazed. I mean, I'm naturally a skeptic. You're naturally an optimist. But explain how. I mean, at the moment, you've always talked. We've had shows where it's hard to understand where the revenue is going to come from. It's hard to understand how publishers can survive against smart machines who hoover up all the intelligence on the internet and spit it back out to users. Maybe users can benefit, although it's harder and harder to distinguish between human information and machine-generated information. So where's the win here?
So I break it down in the editorial into three things. Firstly, just to set the ground rules, AI needs access to the world's information in order to train, improve, If it doesn't have that access, AI ceases to be as good as it's promised.
Let me just respond to that. You always hear that AI needs access and permission and blah, blah, but it doesn't really. It's already acquired all the information, legally or otherwise. But go on.
I don't think there's ever an end. to AI needing access to information. I mean, the world doesn't stop. History goes on. It has to be real time and up to date. And it won't be AI unless it's better than any human. And humans obviously can read today's news. So AI needs to be able to do that as well. The second thing is publishers need a sustainable way to monetize their work. And they do that by subscription, sponsorships, advertising, or various kinds of direct payments. And the third thing is users want reliable discovery and knowledge.
Well, I mean, I'm not sure most users want... I mean, given the echo chamber nature of our current culture, I'm not sure they want stuff that they think is reliable. Often it isn't, but go on.
Well, I think if you put your brain into make yourself responsible for architecting a future you would like to live in, these three things are part of that. Whether you think they'll happen or not, All three parties have needs.
And let's just remind ourselves, those three parties are users, publishers, and AI. Although AI, are we talking about open AI? Are we talking about Google? Or are we just talking about some technology called AI?
It's a technology called AI, and they all are players in it. So they all need this. So you have a situation similar to the internet where publishers needed traffic, Google built a search engine that could send traffic. For Google to get paid, it had to have links that were monetizable. And users needed to be able to use Google and find it effective. It's not that different. It's just that the traffic is moving from a search engine into AI. And so an already existing business model, we've already figured this out, is called cost per click, cost per action or cost per thousand visitors. That business model already exists in the world and on the Internet, but it doesn't yet exist in AI because the three parties haven't yet figured out how to work together. And this week that blew up into a massive debate between Gary Tan of Y Combinator, Martin Casado of Andreessen, around a bit of a kind of obtuse subject, which is whether AI agents should declare their identity when they're crawling websites.
Okay, so I want to come to that because that's the core piece of news of the week. But your reading of history, Keith, sounds to me rather deterministic and teleological. You assume that users, publishers, and technology manage to figure out a way of working together. It's almost a logical narrative of history. Is that really how things work? It didn't really work out that way in Web2. Web2, everything was determined by the tech companies, and both users and publishers had to put up with it, especially publishers. They never were incorporated into the narrative in terms of determining what this thing should look like.
Well, yeah, but they all won. The thing is, they all won.
Words and timings
Well,yeah,buttheyallwon.Thethingis,theyallwon.
Andrew Keen
Well, that's hard on, hard on, hard on.
Words and timings
Well,that'shardon,hardon,hardon.
Keith Teare
You're saying that web publishers won in web 2? Massively won. Look at the New York Times revenues today compared to 10 years ago. Massively won. Only bad publishers lost.
Well, we can spend the rest of the day talking about that. I mean, publishing has been complete. I mean, you're right on the New York Times. You're right on the Wall Street Journal, maybe the FT. But the vast majority of publishers lost because most of their businesses have disappeared in the last 20 years. Because they were bad at doing what they do. I don't want to guess.
All of local journalism has pretty much been decimated. Because it's boring and nobody wants to read it. Anyway, go on. I mean, we could spend the rest of time arguing.
So Google won because they built this search engine. The publishers who get the traffic beg Google. In fact, they're so worried about AI because Google might shrink. Why? Because their traffic might shrink. So they totally love Google.
And then users love Google. I mean, no one says Google's a bad search engine. It's deteriorated for sure, but we're all glad it exists. So you at least have to accept the premise that the scale of reading and the amount of money flowing through the internet is bigger than ever.
Yeah, but I would strongly push back on that assumption. I mean, the fact that I've written whole books on this, and many people have written books. The amount of revenue taken by Google and Facebook is astonishing. And whilst they've clearly been winners, for better or worse, I'm not blaming them. It's just the reality of things. I would strongly push back on the idea that publishers have won. I think publishing as an industry has lost dramatically. There are a few successes, but mostly they're either out of business or struggling for a business model.
Well, I would, I don't know if you'd call them bad or good. Everything's changed. So it's increasingly, for better or worse, it's increasingly hard for local publishers, in particular local newspapers, to survive in the internet age because of particularly the way in which the advertising industry changes everything. I mean, Craig Newmark and Craigslist started the revolution, and he felt so bad, he's poured much of the the wealth that he made in Craigslist back into publishing. But anyway, go on. So it's debatable. We could spend hours talking about this. You can't just state, well, everybody won, and anyone who goes out of business is bad, so they deserve to be out of business.
I mean, that's an absurd argument. Well, it's no more absurd than suggesting that hand-sewed cotton goods were punished for not adopting spinning jennies.
Yeah, but we all got cheap clothes, but the hand workers who made... the cotton goods that generated, I don't know, the Luddite revolution, they didn't win. They all went out of business.
Well, but then your argument is just teleological. Given your fetishization of progress, everyone always wins with progress because anyone who loses deserves to lose.
So you're going to disagree with this, but this is now the latest incarnation of progress. And the question is, Is AI going to replace search as the dominant place we all hang out? My guess is yes. Is Google search going to lose traffic because of that? My guess is yes. Is the business model that Google developed that publishers benefit from, not all, going to die or will it transition to AI? And my argument here is that it's completely feasible and doable that it transitions to AI.
Yeah, and I wouldn't disagree on any of the things you've said. So we're in the next chapter of this dramatic transformation of the publishing business from Web2 to AI, from self-publishing to machine learning publishing. So... Users, publishers, and AI, everybody wins. And you and I have talked about this extensively, Keith. It seems like the debate this week, the Gary Tan, Y Combinator, Andreessen Horowitz spat, is about how publishers generate revenue in this new age. Is that fair?
Okay, so you added another couple of players. We've got Cloudflare, which is a network-level management system, and BrowserBase, which is what you call a headless browser. So these are the two tech companies involved in this week's news. We've got Gary Tan, the CEO of Y Combinator, Silicon Valley incubator of early stage startups. And then we got Andreessen Horowitz, which is the most powerful VC in the Valley. So what's the story here?
So the sequence of events is number one, Cloudflare and BrowserBase announced a partnership for supporting a technology which will require an AI agent or a web bot, if you will, to carry with it a certificate indicating who its owner is. And that will allow a publisher to decide whether to let them in or not. So this is basically an authentication certificate that every AI agent would carry on engaging with web content. And a headless browser is basically a browser that an agent uses to engage with web content.
Which is a way of stealing. And they become accountable. So when you do a search on OpenAI for this show and they know exactly what's happening, it means that they must have got your or my permission, whoever owns this content, maybe... maybe substack to actually.
So far at least, so far legally, the courts have found in favor of AI that reading is not copying or stealing. So I believe that that's the case as well. So I think it's reading. But the publisher is being given a tool by Cloudflare and BrowserBase to control who can read. It's not about stealing. Even reading could be controlled.
So here's the question then. Given the vast fortunes involved, would it be, I mean, if Cloudflare and BrowserBase come up with this authentication system, would it be wise or not for publishers to use them to protect their work? Might it make them irrelevant?
That's a very good question. I'll come back to it in a second. Let's just now state that the second thing that happened is that Gary Tan from Y Combinator then called Cloudflare and BrowserBase an axis of evil, wherever we heard that before. And basically in defense of AI said that this pact was evil because it would constrain AI's ability to read.
Calling an axis of evil. Did he have two Xs in it? Only one. I mean, anyone who's called Gary Tan and spells their name with two R's, I think, is automatically suspect. And calling it, what did he call it? An axis of evil. I mean, this is this sort of provincial, Silicon Valley provincialism. I mean, it's such an absurd thing to say.
And then the third thing that happened is Martin Casado from Andreessen Horowitz came in and said he thinks that agents carrying identity is a perfectly reasonable piece of infrastructure that seems like common sense. And that... So he's agreeing with what Cloudflare and BrowserBase are doing. Yeah, and he doesn't think it's threatening to AI.
So this axis of evil comment from Tan, is it... sort of extreme? Has he been reading too much Ayn Rand? Is it just sort of childish libertarianism that no one should be able to control anything or anyone?
Look, I think if one wants to get into the fight, there's a very strong legal base for Gary Tan being right, even if you disagree with him. And lots of people disagree with him, by the way, but lots of other people agree with him. It's a very, very divided set of opinions.
So he made the, I mean, coming back to tan, Keith, it's a moral argument around an economic one, this axis of evil. He didn't say it was against the market. He didn't say it was against someone else's interest. He was saying an axis of evil. He's using this vague moralistic language. Is that because he didn't really have much of an argument?
His argument is that we all benefit from innovation and this will provide friction to innovation and that the So it's rather like you. It says everyone wins from progress, even people who lose. So my point of view gets developed after that. And that comes back to your question. Just remind me what your question is again that I said I didn't.
So my question is, would it be in the interest of a publisher to actually throw their, so to speak, their hat into the browser-based Cloudflare camp by having their data somehow tagged so that no AI can essentially, you don't like this word, but I'll still use it,
steal it? What I think is that publishers need traffic. That's the bottom line. And if AI is going to be where all the traffic goes, publishers ultimately need to make a business relationship to the AI companies.
I would disagree. I don't think publishers need traffic. I think publishers need revenue. And you and I have a show, for example, where neither of us have made much of an effort because we're both lazy. But we could probably sell sponsorship on this to some tech company. And I don't think they care that much about the traffic. They care about the... few thousand Silicon Valley influencers who enjoy our spats, but traffic doesn't really matter. We're not claiming to be... We're not on Instagram or... Let me rephrase what you just said.
I think you're arguing against yourself. You're right. Publishers want revenue. And that takes... There's only three real ways they get revenue. They either get it from advertising or from sponsorships or from subscriptions. And all of those are driven by traffic. So there's a correlation between traffic and revenue, clearly.
Well, no, but let me just come in here because this is, There's a huge difference between advertising and sponsorship. There's a fundamental difference when it comes to revenue. That's the problem with the advertising model, I think, historically, is it's really required massive traffic to generate any viable revenue, especially the way in which Google advertising has commodified advertising. Whereas sponsorship, the traffic doesn't really matter. It's your influence. I mean, obviously someone's got to watch it or listen to it, but it's not the key metric.
I am. It's one area where I am experienced as a publisher, and I know it's not naive at all. I know that I've been doing this podcast for years, and I've been through every version.
Well, I do know that I'm not successful at advertising, which is incredibly hard. Sponsorship, I'm less unsuccessful. And if, as I said, I could probably sell sponsorship. I have done in the past.
I'll agree with you this far. The dollars per thousand subscribers from sponsorship is way higher than the dollars per thousand views for advertising. And yes, sponsorship is a viable business model for people like you and me. I'll agree to that completely. But I also think it isn't true to say the sponsor doesn't care about traffic. The sponsor does care that their money is spent to help influence your readers or listeners
Yeah, I mean, that goes with that. I mean, that's obvious. That goes with it. But it's not the first question they probably were asked. They're much more curious. If we were selling sponsorship on this show, they're much more curious as to whether Gary Tan or someone from Andreessen Horowitz is watching or listening than whether we've got 10,000 viewers on Substack.
That's clearly true. There's a correlation between the quality of your subscribers and the amount you can charge a sponsor for getting access to them. But I think we're broadly going to end up agreeing that traffic is not irrelevant. It is a factor.
But that said, if AI... which seems to be the case, is talking traffic from the open internet into conversations between humans and AI, LLM chatbots, then the business model of publishing, whatever the business model is, whether it's sponsorship, subscription, or advertising, or a combination, needs to find a way to surface inside AI, hopefully without turning AI into an advertising platform.
we would like this conversation, which is highly relevant to what's going on, we'd like anybody on OpenAI today talking about this topic to have... a link to our show surfaced. It would be good for the user. It would be good for AI. We might even agree to pay AI a couple of cents for doing that. And AI, if it sends us traffic, that listener might convert into a subscriber.
OK, so let's use this show as the example of the question I asked, which you acknowledged was a good question. Should we, let's say this Cloudflare browser base, what Tan called this axis of evil emerges, and we have the option are building a kind of walled garden around this content would we want to be in
their network the answer is yes if they agree to surface links if if they summarize what we're talking about without sending the reader to us then the answer is no so it comes down to will ai surface links yes or no and
We want them to see our content, but we also want to turn them into subscribers. I don't know about you. I've got about 100,000 people who read the newsletter. I'd like it to be 200,000.
But I have about another 100,000 followers on both LinkedIn and Twitter that see my content and on Facebook, about another 30,000. So I want that to be a funnel that converts into subscribers in my sub stack. And if AI can help me there, then I want AI.
Although actually, we should talk about Substack too, because there's an important player. I mean, our business model, if we really wanted to make an effort would be to turn people who watch and listen to this on Substack into sponsors of our show. I don't know how the AI piece plays into that.
Well, it comes down to which sponsors are typically businesses. It comes down to which businesses want to be associated with our content. Sadly, the answer may be not, given our content. But
Yeah, but I don't want to be paid like that. As it happens, my editorial this week is completely in favor of the Cloudflare Browserbase deal and is trying to say to public...
don't be scared of this, embrace it, because actually it's good for you as well. And Gary Tan's knee-jerk reaction is understandable, but probably too short-sighted.
So you're on the Andreessen, I'm surprised that you're going against the incubator spirit of Gary Tan and the libertarian spirit and saying that what Cloudflare and BrowserBase is doing is actually good. And is this how potentially Keith, users, publishers, and AI can all win in this new AI?
Exactly. There needs to be, I call it a trusted third party, which is a common phrase in the internet. A trusted third party that has links and the economic models associated with those links and allows AI to learn from them in order to generate traffic back to the publishers. It's an obvious thing. It already happens in Google. Let's just recreate that same universe, minus advertising, in AI.
Well, you're breaking... You talk about breaking Google's dominance, but maybe Google even... want to break Google's dominance. It's the classic innovators dilemma. So this trusted third party matters. Was there a third party, a trusted third party in Web2 or was that what was really missing? What can we learn from history in terms of these other moments where there was so much potential and opportunity and things didn't quite get real.
To be honest, Google played the role of of course a commercial but nonetheless a trusted third party. I tried to turn real names into that and failed. Google won. Google became a trusted third party that made a lot of money from being the trusted third party because it could charge rent on traffic and, as you pointed out, probably more than it deserves.
Yeah, well, I mean, but I'm in your camp when it comes to... I mean, if the market's there, then they deserve it. If it isn't, then they don't. We did trust Google in the old days. Now, when you use the traditional Google search, you get a lot of advertising, which makes it a little dodgy.
Yeah, well, it's evolved into a self-interested commercial traffic generator. So, yes, you're right. But I do think Google's going to shrink down to the market... deserved by its Gemini AI. And OpenAI and Anthropic are going to have the market deserved by their AIs. And so there will not be a dominance anymore. There'll be many players, way more than those three, who will all need a trusted third party as a source of monetizable traffic. And
Yeah, I agree, although the markets might not agree. I mean, Google stock is up dramatically in the last six months since this has become more self-evident. Maybe it's because they're reinventing itself. So you use this phrase, Keith, in your editorial, which made me automatically suspicious as a natural skeptic. You talk about something called a virtuous circle. Yeah. What exactly does that mean? I've heard this term before, often in Silicon Valley, where we've always talked about virtuous circles and what we've got often are vicious cycles. So why is this a virtuous circle?
And I will say it's a theoretical virtuous circle because for it to be virtuous, this trusted third party platform has to come into existence. It doesn't exist yet. But the virtuous circle is that Publishers allow AI to learn on the content, possibly by charging them as well. In some cases, the content and learning will be valuable enough that the AI's will pay. Secondly, that AIs, having learned, surface content from those publishers with attribution that can be clicked on. And thirdly, as users, as we use AI more and more, it exposes us to publishers' content instead of hiding it. That would be the virtuous circle.
So in other words, when you go to Anthropic or ChatGPT, when you go to Claude or ChatGPT, you wouldn't only see the output. You'd see how the sausage was made. You'd get a view of where all this data comes from. Which would be interesting in itself, but it would also lead you back to the source.
Exactly. It's a little bit like when you do a PhD, they ask you to go to find the original material. You can read books about topics, but when you're writing a PhD, you have to go to the original source material. well ai will will probably want to do the same it'll want the original source material and that means the web will not die it may be buried inside ai conversations but by providing attribution and links it's accessible and lives on and and so the business models of publishers get to continue
Although I don't know how that would work. I mean, most data is so complicated. It has so many manifold origins in so many different ways that it literally, it would be the kiss of death for AI if every time you went into it, it was like reading someone's PhD dissertation. Who wants that? People just want fast information, quick, easy, cheap, ideally free.
I think buried inside that objection, there's a technical issue. And that technical issue is all about relevance, which is an everlasting issue in the internet. If the user is looking to buy a specific thing, it's easy because there's SKUs which is the unique code for a product and there's many merchants selling it and the AI can probably know them and can tell you what the price is and whether it's in stock and stuff like that. So some information like things you're gonna buy, like newsletters for example, they all have unique names. uh some of them are easy other things like topic specific general stuff is harder because there may be multiple sources and you've got to have some kind of a ranking algorithm like google has that suggests which is authoritative yeah i mean i think
this will be the this virtuous cycle i mean it sounds interesting but it sounds like the kind of thing that jimmy wales might fantasize about you and i might fantasize about it but most people just don't care most people aren't
interested and i just well that's why you need a builder to build it and i i mean to be self-obsessed a little bit here. It's been my whole career to build things like this. I've always built trusted third parties with real names or Egeo. And this is an area of passion for me, but I certainly know I can go and build this.
now there is some self-interest here let's let's rephrase keith's headline users publishers and ai everyone wins if we trust keith is that fair if i was to build it that would be true but i i think a trusted he is gonna he is the the knight on the white horse he's coming along by the way everybody No users, publishers. Well, not users, publishers and AI. Everyone's going to be saved by Keith because everyone will win if we put the Internet in his hands.
It's an interesting segue to Doc Sills, who I'm sure you know, last week responded to our newsletter and wrote a piece about AI and hotels where he he's been writing for quite a long time about something he calls personal AI. And he made the point that he thought at first that our newsletter last week was reinforcing his claims. And then he read it. Our newsletter? I mean, it's your newsletter. We do a video. He read it and then he realized, no, he's talking about a new centralized infrastructure. And that is true. I am. And just as Google was able to get trusted third party with a centralized search engine, I do think this era will lead to some kind of centralized infrastructure serving all the different parties.
Yeah, I would push back, and this may be a conversation for a later time because we're moving on here. But I would push back in two ways. Firstly, there's nothing inevitable about publishers. And secondly, I'm not sure anyone cares about trust anymore. I think it's an overused word. And I don't think anything needs to be verified. For example, everyone believes the US president at the moment, or half of America believes everything that comes out of his mouth, half don't. And they do that intuitively, whatever he's talking about. So I'm not sure there's any connection these days between trust and information. We're entering a truly new age, maybe a post-postmodern age where The idea of objectivity and truth and facts of being transformed into something else. I'm not welcoming that. I'm not celebrating it, but I think it's a reality.
As a builder, I think of it that this is a once-in-a-generation chance. to get rid of the advertising-plagued internet that we've built since the 1990s and replace it with a knowledge-based internet that surfaces links to publishers where relevant. And money still flows, but advertising kind of goes away in those...
Yeah, I'm not sure that the people at WPP or Publicis or the millions of people employed by the advertising industry would be pleased to hear this. But we will come back to that, Keith. We need to move on. In terms of your faith in AI, my interview of the week is with someone who has no faith in it, particularly OpenAI and Sam Altman. Gary Marcus was on the show. I headlined it. Why, even... Sam Altman wants to be Gary Marcus from Son of Sam to Son of Gary in a single ChatGPT release. Gary, and I'm sure you're going to disagree with him on this, Keith. Gary was very dismissive of ChatGPT5. He sees it as a disaster. Is he wrong?
He's completely wrong while also being right. It's a funny thing with Gary. What is he right about? He's right that LLMs... and not the final solution for general intelligence. He's right about that. It's completely wrong to not credit progress in how good LLMs are. They're really, really good at what they do. And to be honest, GPT-5 is the best so far.
Yeah, I mean, I agree. I think what Gary does, it's a good trick. You kind of do it also, is... He just assumes that he builds everything off AGI. So if it's not AGI, it's bad. And he claims that Sam Altman has always talked about AGI. Maybe he has, but it doesn't seem to be the core thing. thesis. Sam Altman is not a teleological businessman. I think he thinks that chat GPT can be fine even if it doesn't reach AGI. So Gary creates this impossible promise of AGI and then says, well, if you can't get there,
then you failed and i'm not sure that's right yeah exactly he you know he he might say the same about a tesla car that it doesn't go 500 miles therefore scrap it well you know 400 miles is pretty good and and uh yeah i i find him frustrating because i think he's clever enough i think he's quite politicized because he i i imagine
Yeah, well, he has a huge following. I mean, you said you had 4,000 followers on Substack. He's got over 100,000. He's a big deal. People want to hear what he has to say. I mean, maybe that message always needs a messenger, and Gary's become a messenger.
You have to admit that you and he do have something in common, this sort of teleology of his is AGI, yours is progress, so everything is justified. His is the reverse of yours.
I never said it was owned by you. I'm saying you're an example of that. Well, is there any other... Let me rephrase the question and we always argue on this one. Is there any, in your view, given how you seem to justify all forms of progress and that even people who lose their jobs and their incomes and their livelihoods because of new technology, is there anyone who has ever lost in technological progress?
No, you're not answering my question. You've reversed it. Is anyone ever lost? Or is your argument that since there's this thing called progress that drives all of world history, we're all inevitably and unavoidably winners?
You know, it's like when you get divorced, Andrew, and your kids are sad. And then two years later, they're happy because they realize it was a bad marriage. And there's always... You've never been divorced, so you haven't even lived through that one. Well, I'm familiar with it. I was raised by a single mom who separated from my dad, so I've lived through it. And I can tell you what was sad initially became happy quite quickly. And so you've got to look at history, not in the frozen moment, but you've got to look at progress as a measure of gain over time. And I define it by human freedom. That is to say, separation from necessity.
Okay, enough of this. Let's move on very briefly. We've got a couple of minutes left. Just start up of the week. Google DeepMind launches Alpha Earth. So you've always been a bit of a skeptic on the Google front, but Google's doing something right. What's DeepMind up to?
Actually, I think Google's doing a lot right, given that it could possibly lose its nest egg in search, or at least shrink. it's done a great job of transforming itself into an AI company. And DeepMind, which it bought for nothing, by the way, from the UK, is the core of that. And this is a whole earth model capable of understanding the impact of variables on the whole. And so you can model climate as a function of Other variables, for example, like chopping down forests would be one.
Yeah, and of course, let's remind ourselves Google acquired DeepMind. Demis Hassabis, who now is defining AI strategy, it's reasonably likely will be the next CEO of Google. So it's a huge deal. Their acquisition of DeepMind might ultimately prove as smart, as prescient as their acquisition of YouTube.
Good. And then finally, your post of the week is, we haven't talked about Elon much recently, but SpaceX releases what somebody on YouTube called an amazing new Starship. What's the post of the week?
So this is acknowledging SpaceX's successful launch of its version three Starship, which successfully took off The booster rocket successfully came back to Earth. It released a payload and then it came back to Earth and landed in the exact spot it was meant to. And why is this huge? Because the two previous attempts failed and blew up and they've now done a successful one. And usually when that happens, they've figured out lots of the variables and now they can go to the next stage. which is um to supply this to nasa for the next nasa missions which by the way ultimately include going to mars so it's it's a pretty big deal and the technology the technology is is is you know always hiring maybe you should have changed the
title of this week um what would we say humans the universe and elon everybody wins right let's see but uh Well, we'll see. Next week, I'm in Australia, so we're not going to do a show. You might do your own show, but we'll be back in a couple of weeks, everyone, where there'll be a little bit more real news, maybe rather than just Keith and Andrew arguing over progress. But have a great couple of weeks, Keith, and we'll talk again in the middle of September.