Ben Whitelaw 0:00

So Dex, like other guest hosts of Ctrl-Alt-Speech, I'm gonna ask you the same question as we ask them, a way to alleviate the gloom of the podcast that sometimes, descends when we're talking about matters of, uh, speech and, and the internet. But tell me one thing that you saw on the internet that you thought was fun

Dex Hunter-Torricke 0:15

maybe it's not that disconnected from the gloom, but I have been loving the memes that people have been creating about the Nvidia, scandal. the most extraordinary usage of AI, ironically, to create songs. I saw an amazing music video, set in the '80s with all the tech leaders, called "We Were Always Blue," um, for, for the pivot, and it had just the most fantastic quality of writing of the lyrics as well. Our values never changed, they just updated overnight."

Ben Whitelaw 0:47

Yeah. Nice, nice. You know, there is upsides to this technology, I guess, in, in those ways. very fun. I, I would say non-related to this week's news agenda, but I came across a, 10-year-old project that has been going, in the background for a long time of a guy who's been taking pictures of UK bins. If you go to govbins.co.uk, I think it is, uh, or govbins.uk, he has been taking, with, with the help of volunteers, pictures of bins from around the UK, and he has been asking people to submit their bins, to take photos of them, and he's only got 69 councils left. So he's fastly moving towards one of the best collections of UK bin photos on the web, which I think is a, is, you know, a great accolade.

Dex Hunter-Torricke 1:34

That's amazing. It sounds like a lot of rubbish.

Ben Whitelaw 1:37

I'm gonna send my photo in now. Hello, and welcome to Ctrl Alt Speech, the podcast where you make sense of the major debates shaping online speech, platform power, content moderation, and the future of the internet. It's the 12th of November, 2026, and this week we're talking about what the midterms mean for big tech, regulating the frontier, and for our extended Patreon episode, how serious should we be taking AI safety fears? Ctrl Alt Speech is the weekly podcast of me, Ben Whitelaw, founder and editor of Everything in Moderation, and Techdirt's Mike Masnick. As you can hear, Mike is away this week, and I'm very glad to be joined by a man who knows a lot about these big tech companies. He's seen it from the inside. Dex Hunter-Torricke, welcome to the podcast.

Dex Hunter-Torricke 2:24

Great to be here

Ben Whitelaw 2:25

just to give you a bit of background on, Dex and, his qualities for being here and particularly on this week of all weeks. he's the founder of the Center for Tomorrow, a London-based nonprofit working on AI governance and how democratic societies cope with what's coming next. very relevant for today's topic. Before that, he spent two decades inside some of the biggest tech companies in the world. He was a speechwriter to Mark Zuckerberg and head of exec communications at Facebook. He's worked at SpaceX, he's worked at the Oversight Board and at Google DeepMind as well. So there's really nothing he doesn't know about how some of these companies work. Dex, welcome to the podcast. I wanna talk a bit about the Center for Tomorrow because it's something you've been working on since earlier in the year. just give listeners a kind of overview of the reasoning for starting it and what you've seen in the months since you've been working on it.

Dex Hunter-Torricke 3:10

after 15 years in big tech, I had a very strong conviction, that things were not working out for societies when it came to the promise of technology. I've always been a huge believer that technology can be an extraordinary force for good. No one's ever accused me of being anti-tech in my life. But absolutely the industry, the role that it plays in the world, and how this technology is actually arriving in the world and the effects it has on ordinary people's lives, that is something now which I have very serious concerns about, as clearly do many people in the world. And, uh, at the end of September last year, I left, Google, and earlier this year I launched, the Centre for Tomorrow. And my conviction is that if we do not change course as societies urgently, then, we are gonna end up in a future that will not serve the interests of most people in our societies and for most countries in the world. And we have a tiny window in which we need to radically revisit the social, the economic, the geopolitical, and the climate assumptions upon which our current plans for the future are built. And, within the industry, you know, I spent, a long time, you know, in the room with Zuckerberg and Elon Musk and others. it's just become incredibly clear to me that leaders have abandoned many of the values, many of the, responsibilities they have towards society. So I decided to go and work on trying to advance that agenda outside the industry.

Ben Whitelaw 4:40

And what does that, mean in practice? How are you doing that, and what kind of signals are you seeing as you go around? You know, you mentioned before we started the podcast that you've been in 10 countries in 10 weeks, which is a grueling travel schedule by anyone's standards. Like, what are you hearing in those, rooms and in those kind of meetings with different stakeholders?

Dex Hunter-Torricke 4:56

Every country in the world is struggling with a common set of challenges about how to manage the arrival of this technologically shaped future. And, everywhere I've been in the last 10 weeks, uh, and during this year, from China to India to Australia, the US, leaders are looking for answers. And so the center is focused on developing those solutions and working on helping all of those leaders, governments, industries, communities to figure out what a real agenda for managing AI and its effects looks like. And in a way, the way I think about it at least, those 15 years I spent advancing the story as well as the, regulatory strategies, the political strategies of these leaders, I now employ all of those exact same skills, networks, experience for a different set of people who are trying to fight for a future, of AI that's more humane and just and sustainable. And, the reception has been unlike anything I've ever experienced, in my life. truly now I think a critical mass of people and leaders has arrived who genuinely want to solve the future, and they realize that that solution is not gonna come within five or six big companies in Silicon Valley

Ben Whitelaw 6:06

Yeah. And, uh, I think you're gonna talk a bit about some of, you know, those conversations you've been having, particularly this week with some leaders in AI, being interviewed yourself on stage as well, I, I hope so. Hopefully those kind of-- that'll come out in the course of today's stories. But you talk a bit about that timeframe being very short, you know, this being the most important decade in history, I think it, you said at one point. is that a kind of cause for panic? do you think we have enough kind of reasons and agency, to kind of mobilize people to do things that help us inch our way towards that better future?

Dex Hunter-Torricke 6:36

The folks who believe in the better future and want to achieve it, they have to realize we're losing. We have a very small window in which to drive this profound transformation of our societies, of our institutions, of our policies. This would be hard under any circumstances, and now we're being asked to do something vastly more ambitious for it to be thoroughly global in nature. I don't believe any society, including the two superpowers, gets out of this century in good shape unless the whole world gets to come too. And we're being asked to do it as well, really, really fast with a set of incumbent leaders in many systems who are not the people you would be expecting to drive the most profound transformation of human history. There has been a historic hollowing out of capacity within states, within many of the, domains where we need to be shaping these kind of solutions. And so it's really, really hard, but in a way, we have no other choice. We have to fight for that future. We have to work as quickly as we can and, uh, see if we can turn this thing around

Ben Whitelaw 7:39

Yeah. l- let's dive into that then. we're gonna talk about... We haven't done a recording here on Ctrl-Alt-Speech since the US midterms. I somewhat thought that would be the kind of dominant storyline, this week, but we've-- there's about four other candidates um, that we could talk about. We're gonna, briefly talk about the US midterms, talk about the big story of, this week and, and today in particular, which is NVIDIA. and then we'll kind of move into the, I think the more kind of speech-related stories that I think are important to our listeners. the midterms was obviously a massive deal because AI, basically for the first time, was put in front of US voters, and they clearly rejected it. You know, the exit polls suggested that more than half of people, made their vote based upon AI and data centers and some of the kind of what's going on in their local communities. what was your take on that? You know, a lot has been said in the last week, but give me a kind of overview of where you sit now with a, a week's worth of hindsight.

Dex Hunter-Torricke 8:32

Big Tech has been pursuing a political strategy nationally for, years now. In the second Trump administration, they chose to completely sell out any shreds of their values and responsibilities and to become intertwined, essentially, with the Trump administration. And that was a means of pushing a deregulatory agenda and looking to speed run progress towards, AGI and all the things that the tech industry wants to do. at the local level, when it comes to building out the trillions of dollars of AI infrastructure and data centers, essentially the strategy was ignore all other considerations, including the effects of that infrastructure on people and communities, on people's wallets, and the effects on energy prices, on water resources for communities, on the actual blight on landscapes, the fact that you have incredibly, huge, noisy, ugly infrastructure being built in places where people don't really like that stuff. And all of those chickens came home to roost last week. voters emphatically rejected that agenda. You know, there were 41 ballot measures, you know, across the US which are reined in data center construction. You've had governors coming in, like the new governor of Texas, who's opposed to that entire agenda and calling for a pause on, some of the deals that were made, with Elon Musk. and then, of course, at the national level, with Democrats sweeping Congress, you can now expect a very serious agenda focused on regulation, oversight, accountability of the tech industry. And, all of that just reflects the natural, inevitable consequence of a deeply arrogant, misguided political and communication strategy from these companies. And in doing that, it has been the, most unbelievable, shot across the bow for the whole industry. the conversations I've been having over the last week with governments, regulators, decision-makers around the world, um, it has given enormous impetus now to people who want to rein in that, unrestrained power of the tech industry and find a different path. In a way, it showed that for all of the vast resources of the industry, the nearly $500 million that Elon Musk committed to backing Republican candidates, none of that is gonna work if the technology and the companies behind it are so unpopular.

Ben Whitelaw 10:49

Yeah. You would've thought Elon would've done a few focus groups with the, with all that money, wouldn't you? he would've, would've found out pretty soon

Dex Hunter-Torricke 10:55

he thinks he's the main character of the 21st century, so he doesn't need any focus groups. But you know what? People told him, uh, "We don't like what you're selling."

Ben Whitelaw 11:02

Yeah, exactly. I mean, talking of main characters, again, the big story of this week following a New York Times investigation, was the affairs of Jensen Huang, the CEO of Nvidia, who this morning, UK time, where, Dex and I both are, announced that he was stepping down as CEO of, of a company that he's led since the early '90s and founded. And this is, as big as a story you get in the tech space, frankly. a lot of the kind of chats that I've been in have been very interested in this from a different, a number of different examples and, and angles. what is your take on this and how should we be reading this story and, what kind of the implications of it are?

Dex Hunter-Torricke 11:39

In a way, I read it entirely as the same story as the midterms. the fact that it unfolded just a few days after the midterms, almost couldn't have picked more poetic timing. it is just, uh, the perfect example of the kind of arrogance, the leadership and cultural deficit at the top of the industry, and the huge negative consequences that can have for a lot of other people if the tech industry is allowed to operate in the, unconstrained way it has done for the last several years. here you had, arguably the most consequential CEO in the world, at this moment in history. somebody who really, probably has a claim to be one of the main characters of the 21st century, and, he engaged in a course of action so risky and ultimately self-destructive to himself, to the interests of Nvidia, to the entire, build-out of AI, that now, this has compounded all the, challenges that are coming for the industry from the midterms. And even before we got into the Jensen cycle, the immediate reaction to that blue wave was stock prices for the companies getting hammered, people trying to digest what it would mean to be facing this new regulatory agenda coming from a Democratic Congress. And now you've got a story that hits all, people's, favorite, you know, hot topic, items. You know, it's a sex scandal. It's an ethical conduct, crisis. It's one that involves national security and geopolitics. it's one that is about the organizational and cultural rot within a company dominated by its founder for more than three decades. so this is a reckoning moment for Nvidia, but it's also just one more example of why, tech needs to get its house in order.

Ben Whitelaw 13:15

Yeah. I think it's a really, interesting time, and that is the context on, on which this week's stories that we're gonna talk about sit. you know, the first one that we'll dive into, I think, is a reaction in many ways to, to a lot of that, anti big tech sentiment that has grown over many, many years. this fear around what the kind of major labs and, and major data center and semiconductor companies are doing. some of those companies sit, quite outside of, our usual remit here on Ctrl-Alt-Speech, but, it's important that to bring them into the conversation because they affect a, a lot of what we do as a society in response. And so let's start with the Paris call for AI safety then, which is a, a again, a notable moment. Got a lot of trumpeting on earlier on this week. Emmanuel Macron announced that he had convened a group of, governments and companies to essentially, get another one of these kind of voluntary agreements signed to do a number of things, for a common evaluation framework of AI models for independent evaluations, by third parties, which is something that a lot of people have called for for a long time, and something that has been very top of mind since the Hugging Face incident earlier this year, a shared kind of incident reporting process so that, you know, these kind of major incidents don't go unreported for days or weeks at a time. we saw a number of different countries on the initial, you know, agreement. A number of countries have been added since. OpenAI and Anthropic are on there with a number of other AI companies who frankly kind of matter less. what's your sense on, on the value of this kind of agreement, Dex? we have seen it feels like many of these before. I was having to kind of think about quite how many voluntary agreements we'd had. is this the latest in a long line of, of fairly meaningless agreements, would you say?

Dex Hunter-Torricke 15:00

I will confess, until you mentioned this to me, Ben, when we were chatting earlier, I had no idea this was even a thing. I was talking with Yoshua Bengio yesterday at an event, and we never even mentioned this once. It didn't even come up. this to me, uh, is worth a warm bucket of spit. this is a thing which, even Google and Meta didn't take it seriously enough to sign up to it. the stakes are so low if they don't actually go along with it. But, like, when you look at the details, right, it's entirely voluntary. there's no deadlines. There's no penalties. some of the stuff sounds fine, but like, okay, come on. This is, like, as basic as it gets. Like, this is stuff that, like, these companies would probably be doing anyway, unless they literally just wanted to look absolutely irresponsible. It's just very basic, you know, sort of, governance stuff for them. saying that they're going to report incidents, well, you would probably have to do that anyway for a lot of these companies, you know, as part of their financial reporting, as part of their obligations as public companies. so I read very little into this. it is just the latest in a whole bunch of these declarations which are voluntary. and of course, we had the, US equivalent of this when Dario Amodei called for the pacing of the frontier in September, and then we had, Donald Trump with a number of CEOs at the White House who then, claimed they were all committing to this very hastily pulled together set of bullet points. but like, does it actually change anything? What do we actually think will happen if every single thing in this set of voluntary commitments was actually even abided by and implemented? Like, what comes after that? Because you could do every single one of these things 100% flawlessly, I don't think most of them will be, and it doesn't actually solve any of the problems. Like, we still have AI advancing very quickly. We still have public sector institutions and governments utterly ill-equipped to manage the effects of it. we still have a very narrow, rigid agenda amongst businesses, governments, regulators, communities on how to manage the social, economic, political effects of AI. So for me, it just feels underwhelming in a way. It sounds like Macron just wanted to have a thing to announce

Ben Whitelaw 17:00

Yeah. I mean, he has, been doing that fairly often as he kind of ends his term, right? You know, there is this seeming attempt to kind of create a legacy for himself, which is borne out in other things we've talked about on the podcast, including a social media ban for under 16s, which was not, well received by French courts and, is probably not gonna go anywhere, at least until there is a, an EU-wide, version of it. But, you know, Macron is seemingly doing all his best to be, be one of those main characters, let's be honest, and, and has been like that for a while. I wanna talk about some of the countries that are on the list. there has been a lot of talk in AI safety and AI regulation circles about kind of middle powers coming together in an attempt to counter this, winner-takes-all narrative, and the role that middle powers like France, like the UK, like Germany can play. there are notable number of middle powers missing from the list. There is no Japan, there is no South Korea, there is no Australia, there's no Brazil or India. what's your take on the ability of these middle powers to come together to create leverage and, do something that breaks up this, global power dynamic that we're seeing where China and the US are, are going full tilt against each other?

Dex Hunter-Torricke 18:10

There is tremendous opportunity to do that, but it would require powers and folks who want to push, you know, the right governance of this technology to recognize this is a conversation about power. It is not one that is mostly just about technical management of the systems. So when I see these kind of proposals, these frameworks where it's, bringing a few countries together, to look at, uh, you know, setting up, monitoring of, systems and, having guardrails, potentially. But yeah, fine. That's okay. But there's a much bigger, opportunity and an imperative, which is you have a lot of countries who are not the US and China who will be locked out of the future shaped by the technologies which are being built in the US and China. most of the deep tech ecosystem and the value that is gonna be unlocked by advanced AI, quantum computing, nuclear fusion, robotics, on and on and on, these are going to deliver massive gains to a relatively small sliver of people, in the US and China. And for everyone else, they need to figure out how to make themselves into an inconvenient block of people who have a seat at the table. And what does the solution look like? The most hopeful example I've seen this year of what true middle power cooperation looks like to manage an AI-shaped world, it's got nothing to do with AI. It was Canada being lined up to be the first associate member of the European Union when Mark Carney spoke to the European Parliament a couple of months ago. And Australia and New Zealand are now rumored to be in talks as well with the European Union to also become associate members. That is an example of an ambitious cross-cutting agenda of global collaboration, which now everyone who wants to, have a world where AI and the benefits are shared with everyone, they should be pushing for that to expand to include all those countries you mentioned who right now we're not doing vaguely enough to bring them to the table. India, Brazil, Japan, lots and lots of other countries who believe in democracy, believe in the values that we care about, and they don't wanna be ruled by Washington or Beijing

Ben Whitelaw 20:11

Yeah. Yeah, that makes sense. I mean, Paris is a voluntary... This, Paris call is an, a voluntary agreement. It is I think surprising that those, as you say, are still, normative. I think there's probably I would expect much more from an EU perspective in terms of the regulatory drive. But we've seen in other countries and jurisdictions this week that that's actually you know, moving in different directions. So Australia announced yesterday that it was, taking a very strong regulatory approach in response to a number of, incidents, including OpenAI hacking some of its government infrastructure earlier this year. So, uh, Anthony Albanese had said he was taking on all of the recommendations of a report from that incident and was moving quickly down a regulatory r- uh, avenue for those frontier models. In New York, we're also seeing the, the RAISE Act, which is a, statewide piece of legislation in which the frontier models have to, again, declare a lot of their safety frameworks, incident reports on a public portal for other people to, to read and to understand, I guess, is the first port of call. how do you expect that trajectory to change over the next kind of 12 months? Are you... Is that the kind of only, only route that we're gonna see towards a regulatory model for frontier labs?

Dex Hunter-Torricke 21:20

I think you're gonna see a huge proliferation of different regulations and strategies used by a lot of governments. But, like, the main sort of common theme is there is a massive counterattack now that is building, which is going to be pushed in countries all over the world. The blue wave was just the start. plenty of people now have, got fed up with the agenda coming from tech. The, Trump administration has entered its lame duck era and is groaning under the weight of its crises, domestic and international. And so, governments, regulators, communities, activists, people now are looking for much harder agendas against the tech industry. And, governments who absolutely are not radical, they have been wanting to gain the economic, benefits from AI and to get the proceeds of growth. They are signing up to that because governments also are political creatures, and they recognize that, backing up big tech and following the, agenda of the last four years, just continuing that, is a losing political hand. And so Australia, has had a particular role, I think, in pushing, ideas, policy proposals that are much more ambitious, which then end up getting replicated. Of course, the, youth social media bans, that was, like, one particular example. I think it's quite interesting actually why Australia ended up in that position. We could talk about that for hours. But I think there's probably something in there about being a market, you know, which sort of is under-invested in by big tech. they tend to put most of their sort of public policy apparatus and lobbying efforts into a fairly narrow sliver of Western European countries and the UK. and then other markets, they, don't take them that seriously. They're smaller markets, they're less lucrative, so they have less influence. And those markets tend to then, react more strongly, to tech. But the, uh, data center regulations that are being developed in Australia, for example, I was there, a few weeks ago and I met with a bunch of, you know, members of the Australian Parliament and others who are pushing those through. They are some of the toughest proposed laws on, managing data center build-out in the world. And what I fully expect in Q1 is that as Democrats, take over Congress, Australia will end up, becoming a powerful, template for how to manage data centers and that infrastructure in the United States. And I think that is gonna set off a wave of replication of the Australian proposals around the world

Ben Whitelaw 23:38

Hmm. Interesting. we, we've talked a lot on the podcast about how the Online Safety Act and the social media ban has, has been replicated, you know, sometimes not always, uh, effectively, and sometimes some of the kind of most ineffective parts have been copied. but I think you're right that, being the, the first mover on this does have benefits. My, my concern, Dex, is that with the focus on transparency around these companies, on, these Frontier labs declaring and submitting information about what they're up to when it comes to safety and how their, how their guardrails are protecting users. You know, we, we have seen that before. You know, the, DSA and the OSA are predicated on transparency, you know, on the idea that you have to submit risk assessments, ahead of feature launches. You have to produce audits. you have to respond to RFIs from the regulators. There is a lot of information about, out there in the world about how, these other major tech companies do safety. I'm not sure it has moved the needle in terms of our understanding of the platforms or has given us the kind of leverage to argue that they should do things differently. and obviously enforcement has been a big question this year. and I predict will be a big issue next year as well as some of the investigations of these companies go into, sometimes a second or a third year. You've talked a lot about how, you know, there is this risk of us regulating AI in a, in the same, same way as social media or of making the same mistakes when it comes to tempering the power of these companies. Are we using kind of the same model and hoping it goes better this time round when it comes to regulation?

Dex Hunter-Torricke 25:07

100%. how much did, uh, NVIDIA's risk registry, uh, stop Jensen doing all the things he did? you can have anything you like on paper, but ultimately what we need to pay attention to is outcomes and action. And the action which we should be prioritizing is something that goes well beyond a legal, a very narrow regulatory formalistic model of action. It is political will and a much more ambitious agenda that should be turning up in the crafting of policies, and legislative priorities at every level massively. Like, that would be a serious sign that we are handling what is perhaps the defining, issue of our time. fiddling around with risk, management frameworks in the background is a sure sign that the debate has yet to properly enter the mainstream politically

Ben Whitelaw 25:56

Yeah. let's talk about that just for, for a minute before we move on, because my sense and my thesis is that not enough people understand the, potential risks of AI. There are people kind of very close to me who I'd expect to be somewhat familiar with even just the near-term risks, not even thinking about the, existential ones, which a small group of, the AI safety community are worried about. how do we get more people to know about the risks of AI or to engage in ways that go beyond data centers? What are the levers we can pull?

Dex Hunter-Torricke 26:27

I mean, it's in a way the simplest one, and it's one that, I perhaps have a bias on this because I spent my career on it. It is about telling the story. the ability to translate a set of issues which, if it's done badly, look really wonky and technical into something that is profoundly universally human and speaks to the future of everything that people care about, that is the art of, uh, how to do this thing. And the tech industry is really bad at this, which is why they just got absolutely whooped in the, you know, midterms. But the issues that resonate with people, in a way, are timeless ones. They want to know that, they can have a good quality of life for themselves, to deliver an even better one for their kids, that they can live in places that, are nourished and nurtured. They want to know that there is a good future for all the other people, the communities, the country, the things that they care about. And they also want to, you know, that they're being treated fairly. They've got a fair shot at the future. like they're not sitting around thinking, "Wow, GDP numbers are really, really good." Like that means absolutely nothing. It's entirely possible for a country to have extremely good GDP growth and the proceeds don't get shared by lots of people and they don't feel it. And this is something that Silicon Valley, in my entire time there, seemed culturally incapable of understanding. Feelings are facts too. They're actually the most powerful facts of all. They shape human decisions and choices far more than any amount of numbers you can assemble in your Excel spreadsheet. And so this is the nature of the moment, right? If we want people to take a true transformational societal agenda for AI seriously, to come up with an agenda that says, "This is how we manage AI safely, responsibly. The enormous rewards from AI are shared with everybody, not just, you know, 1% of the population," then the way we talk about AI and how we translate it into then the priorities into the policies, that has to be very urgent and human and probably, dare I say, populist much more than the sort of wonky, kind of weird proposals and the, the tenor of the conversation that a lot of the discourse gets, you know, stuck in. And that's why, I'm all over the world talking to, people and communities and doing that work. I don't consider it a, trivial diversion, from the, core work. The ability to literally activate serious mainstream conversation about this stuff, that is how we win this thing

Ben Whitelaw 28:47

Yeah. And, and, and the ways that people can get involved as well need to be clearer, I would say. It's, it's one thing to kind of know these risks or to understand the kind of trade-offs of, having this technology at our fingertips. But for the average person, particularly now, when there is a kind of cost of living crisis in many places, you know, there is a political, s-stalemate in many countries, knowing what to do in the course of your day-to-day is massively important. And, I still don't think we're, we're seeing that when it comes to, to online safety, but also A-AI as well. The best example is the social media ban in here in the UK, where, you know, it was a very strong, rollout of a campaign directed at a set of, society who, who, are famous for, caring, and that's kind of middle-class parents, let's be honest. Um, and they were able to kind of set up a campaign and develop a theory and get into kind of the political, halls of power within two years and get a ban basically r- you know, from zero to 100. And so it, it is possible, and I think I would love to see more of that playbook replicated in other areas as well, because I think, y- as you say, there isn't much time and, and we need to kind of, s-start now or, or yesterday even.

Dex Hunter-Torricke 30:00

I mean, it's interesting to me, you know, I counter this all the time throughout my entire career in tech, how many experts and well-meaning people who want to drive the right solutions, they don't understand how laws, regulations, or power really work. most of the critical decision-makers who will be the people who ultimately pass the things we want, they know nothing about AI. They don't care about AI. Like, this is like one out of the 50 things they just have to deal with, which they barely, understand. I was in a meeting, a few months ago with a politician who literally didn't know what an AI agent is, and that person happened to be one of the most influential European policymakers who's out there talking about AI, like, all the time on TV and at events. And I, I won't say who it is, but, like, that's a normal interaction. And there's lots of politicians, right? They're going through their day managing 50 crises, most of which are all about getting reelected. if you can't turn what your agenda is into something that is so simple it fits into, like, three bullet points, game over. And so, you know, in a way, why is it that, the social media, you know, set of experts, the people who wanted to see management of that well over the last 20 years, they weren't the folks who ended up proposing things like, bans on infinite scroll or, curfews in the evening for, young people using social media or even just the age bans. We talked about some of that. I'm, exaggerating, but, like, crystallizing it into a simple set of policy proposals that could be quickly adopted, that work wasn't done by them in the end. It basically happened very, very quickly at the end by a set of-- a small set of activists working with a small set of policymakers, and people adopted it instantly after 20 years of arguing because nobody could just tell them what to do. And most of the regulators and politicians I meet every day now, they just wanna know, "What should I do? Like, give me the policy and I'll go push it right now. I don't like Zuckerberg or Elon or any of these people, but, like, no one's telling me how to actually manage AI well."

Ben Whitelaw 31:54

Yeah. Interesting. And, and so, you know, this, this is almost certainly gonna be a focus for next year. I think we, we are set up for an absolutely, massive week, um, in, in some of these regards. You know, you mentioned before we met Kim on the podcast that every week feels like a year. every day is more, more than enough stories for us to do an hour-long episode of Ctrl Alt Speech. So I, I appreciate you coming on to the podcast with that background, that sense. let's do a couple of quick reflections or quick couple of the stories that you've, you've seen on your travels, and which, are worthy of mention, but we won't go into detail. I'll start by just talking about, a company that has come up a lot in our last couple of years here on Ctrl Alt Speech, a company called Sama, who, is famous for kind of doing content moderation for one of your old employers, Meta, as was then Facebook, and who had large teams in parts of the world, including Kenya, and was criticized for its approach to, wellbeing support and mental, mental health support for those content moderators. There was a revolt. There's court cases still happening in Kenya, one of which is just about to kind of start again. they have announced this week that they, will be liquidating, as a result of some of those lawsuits. they had pivoted into data labeling, for large AI models and frontier labs, including OpenAI. And it's interesting, I think that they, they have been a, kind of fall guy of some of what we've seen over the last couple of years in regards to, a greater awareness about how content moderators are treated and AI labelers are treated and, and some of the unionizing around that. So I think that's worth flagging. I think we'll, we'll share that link in the show notes. Time have done a really good report there. it's definitely worth going to have a read. what else has kind of cropped up for you this week, Dex, that's worth mentioning to listeners?

Dex Hunter-Torricke 33:31

The other one that's interesting to me is SpaceX, and, there's like an endless glut of news. but the one that's really interesting is they're getting ready for the next launch of Starship. I think that's gonna be perhaps as early as tomorrow. I haven't been paying attention to the latest launch window, but, in the next few days we're gonna be expecting to see that launch. And that is a critical moment for the company, getting, the new Starship rocket to a point where it is fully reusable. The entire valuation for SpaceX hinges upon that happening quite quickly and being able to then, send up much larger payloads in an economic way. obviously orbital data centers, the future of Starlink, Starshield, et cetera. So that will be one to watch. It's the first time they're gonna try and catch both the first stage booster as well as the, ship component of it and have those both go back to the launchpad

Ben Whitelaw 34:22

and that has a kind of interesting, I guess, trickle-down effect, right? 'Cause if that goes badly, it's been, already been a bad week for tech stocks, but we, you know, that's likely to have an impact, although Musk seems incredibly resilient for, um, even for the worst comms disaster and crises. but there is likely to be kind of downstream effects of that for them as a company and for the set of companies as a whole. it also could mark one of the worst weeks for big tech potentially ever, would you say? I mean, between Nvidia, between the midterms, and then potentially a, not ideal SpaceX launch, you know, that's, an amazing kind of set of circumstances.

Dex Hunter-Torricke 34:57

Absolutely. I mean, you remember when DeepSeek, first hit the public consciousness, and at the time it was the largest single, fall in stock market value for a single company, NVIDIA, in one day. You know, it was about, I think, 400 billion or so in, in market cap. And now we've had vastly greater, destruction of valuations in the space of a week driven by the midterms, driven by, Jensen. some of the effects are a little unpredictable as people try to flee into other tech stocks to get away from NVIDIA. So actually some companies are up today, but, SpaceX is one of the most valuable companies in the world by market cap. It is the other big infrastructure player. If, you end a week where you've had Jensen resign, you've had the effects of the midterms being digested, and you then blow up the critical rocket, which holds up a $2 trillion valuation for the company, then, truly, that is a historic week. And, um, the memes write themselves

Ben Whitelaw 35:53

Yeah. You can look forward to some more, uh, more AI-generated songs with the, with those CEOs. interesting. So we'll, we'll include, some links about, SpaceX and about Sama in the show notes. we're gonna wrap up our free episode now. We're gonna move into the special Patreon section of this week, and we're gonna be doing a bit of a deep dive on a report which many people would've seen, maybe not gone deep in the weeds on, which is a Meta report. a very notable kind of addition to this question of when artificial general intelligence will happen, when super intelligence will happen, not the Donald Trump version, but the actual, version of it. And what that means for you know, some of the urgency that you've been talking about, and which the center, is very kind of conscious of, Dex. So we'll dive into that now. We'll try and kind of make sense of it and, If you haven't already become a Patreon supporter, now is a great time to do that. go to patreon.com/ctrlaltspeech. Let's jump in.

Announcer 36:44

That's the end of this week's free edition of Ctrl Alt Speech. To hear the extended episode with additional coverage, become a Patreon supporter at patreon.com/ctrlaltspeech. That's C-T-R-L Alt Speech. If your organization is interested in sponsoring the podcast, contact us via our website at ctrlaltspeech.com. Thanks for listening