Tech News Weekly 450 Transcript

Please be advised that this transcript is AI-generated and may not be word-for-word. Time codes refer to the approximate times in the ad-free version of the show.

Jacob Ward [00:00:00]:
I'm Jake Ward, guest hosting Tech News Weekly. And this week you're gonna hear from multiple journalists about some very interesting and amazing stuff. Amazon scanning Twitch people to feed into the AI monster, the ways in which the tech boom may not in fact reward the cities that wind up hosting it, and the ways that the European Union is finally showing all of us what is AI and what isn't. So all that's coming up on Tech News Weekly.

Jacob Ward [00:00:34]:
This is Tech News Weekly, episode 450, recorded Thursday, August 13th, 2026: Amazon Training AI Models on Twitch Streamers. Welcome to Tech News Weekly. I'm Jake Ward. I'm standing in for Micah Sargent, and it's hard to do. That guy is awesome, and, uh, I'm going to do my very best to try to keep up the, the good vibes and smart energy that he brings to it every week. I am a tech journalist. I work at— I'm a contributor at CNN, and I run something called theripcurrent.com.

Jacob Ward [00:01:07]:
And each week, as you know, we bring in other journalists to talk through some of the stories of the week, some of the biggest stuff that we're seeing. And our first guest this week joining me to share what each of us consider the story of the week is Amanda Silberling. She is senior culture writer at TechCrunch. Amanda, you and I have never met. awesome. Like, I love— yeah, I love meeting you for the very first time, uh, in this format. And we don't really know each other's stories here, so this is going to be a genuine conversation of discovery, which I'm excited about. So great to see you.

Jacob Ward [00:01:38]:
Thank you so much for being here. Tell us a little bit about what you do at TechCrunch so people get oriented as to who you are.

Amanda Silberling [00:01:43]:
Yeah, so I write about the intersection of tech and culture, basically just all of the ways that technology is changing how the world around us looks, which right now is a lot of AI, um, a lot of just I've found it in particular in the last year or so that like people that don't work in tech like just want to talk to me about AI and like what do I think is going on with AI. And I think that always sort of is a good indicator of the things I should be looking into because I think my audience that I try to write for is like normal people who don't know what's going on in tech. But then also sometimes I think that people that are so engrossed in tech can be a little blind to what, like, the rest of the world feels like. So I'm basically trying to be that gap.

Sharon Goldman [00:02:40]:
Right.

Jacob Ward [00:02:41]:
That's a good bridge. I like that a lot. So let's talk about your story of the week. What was it for you?

Amanda Silberling [00:02:47]:
Yeah, so I guess speaking of, uh, people having really strong feelings about AI, so my story of the week is that yesterday Twitch, which is owned by Amazon, announced that It is now going to start using creators' streams and videos that they post on Twitch to train Amazon's generative AI. And the way that they framed this story, I think, is part of it, where Twitch knows that its community is creative people that generally it's like they're playing video games or like People are doing all sorts of different creative stuff on Twitch, and generally people that are in those industries are more hesitant to embrace LLMs because they see that they've often been trained on content without creators' consent. And there's just a lot of like understandably tense emotions going on in those spaces. And the way that they announced it was that they said that they were rolling out a feature that would allow creators to opt out of having their streams used to train generative AI. And my initial reaction, as well as I think a lot of other streamers' initial reaction, was like, oh, I didn't realize Twitch was even doing that to begin with. And they might not have— like, what they said was kind of confusing, but it seems like more so this was not really explicitly happening and that they rolled it out as a Look, we're giving you an opt-out option. Isn't that great? But of course, creators very quickly, like, saw through that. I say I'm not sure about the training on the content before because there was a livestream that the head of community at Twitch, Mary Kish, and the chief product officer, Mike Minton, did, like, very shortly after this announcement.

Amanda Silberling [00:04:52]:
to just address creators' questions. And somebody asked, have my videos already been used to train Amazon's generative AI?

Jacob Ward [00:05:01]:
Great question.

Amanda Silberling [00:05:02]:
And Mike Mitten said, I'm not sure.

Jacob Ward [00:05:05]:
Wow. Right.

Amanda Silberling [00:05:06]:
So that's— yeah.

Jacob Ward [00:05:09]:
For you, when you think about the cultural— so Twitch, I mean, I'm thinking about how huge some of the personalities on there have become as cultural forces, right? Like Hasan Piker is like the voice of progressive politics in the West, right? He travels not, not just all over the United States, but now like all over Europe talking about, about politics, right?

Amanda Silberling [00:05:34]:
Yeah.

Jacob Ward [00:05:35]:
And like, for that guy, it's such an empire what he's built on Twitch that the idea that, that suddenly they could create, you know, they could use all of his incredibly well-performing content to do something else. Like, for you, like, if you If you were going to try and explain to people who don't spend a lot of time on Twitch just how, like, like what it is culturally and how important to these creators it is to be themselves, to be a human creator on there. Like, I don't know, help us think through a little bit, like, what is it, what would it mean to suddenly like fill that thing with AI slop?

Amanda Silberling [00:06:12]:
Yeah, I mean, I think that like people use Twitch for like people like Hasan Piker where they're getting news from this. They're like thinking through the news, through what people are saying. But then even in terms of like your typical video game streamer, it's like some of the highest earning creators are Twitch streamers that started by just playing video games on camera. And that sounds like it would be boring, but it's obviously— I mean, there's a huge industry around it. So people, people want to watch other people play video games. And I think also it's kind of A form of content that maybe why it's so popular is that you're experiencing it passively. Or like when I've watched streamers before, usually it's like I'm cooking dinner and I want to put something on in the background. And like, it's kind of like, I think especially in the pandemic, it blew up a lot because people were like, oh, I feel like I have a friend here, but I can't actually be with people.

Amanda Silberling [00:07:15]:
But in terms of, I guess, what Amazon is doing with that content to generate other content, I don't— I don't know if it's that they're trying to automate Twitch streams and like make AI streamers. But I do think that's something unique to Twitch is that this is almost entirely like very high percentage of this is people that are sitting like we are right now, looking at a camera, talking into a microphone. And there's undoubtedly like enough hours of Hasan Piker or of like any streamer, not even just like the really big ones, that there are so many hours and hours and hours of them talking and looking at a camera that you could feasibly use that to deepfake really well. Where I think when you think about like training AI models and how much sheer content you need, there's probably more footage of someone like a big Twitch streamer just looking at a camera and speaking than there is of like, like Donald Trump, maybe. I don't know. I mean, that might be—

Jacob Ward [00:08:27]:
Well, but you're right. And it creates this like opportunity for Amazon to create, like if you're gonna create like a customer service agent or a virtual receptionist or any of these other things in which it's like a person facing a camera speaking to you eye to eye. That's right. It's an incredibly faithful way of training that. You know what it makes me think of too is like all of these companies and all these acquisitions we've seen where, you know, I'm not even sure the people who bought them, right, would understand what they're going to do with it eventually. But like, you know, I think about like WhatsApp and Meta where now like just recently I had to like connect those accounts in a new, you know, my Instagram and my WhatsApp.

Amanda Silberling [00:09:09]:
Mm-hmm.

Jacob Ward [00:09:11]:
that I'm like, oh wait, there's a connection here being made that I didn't even sort of think about when these two first came together. And with Twitch and Amazon, it's the same kind of thing. I wonder what the original Twitch leadership thinks when they look at this and think, oh, now suddenly we're like a database for these guys. There's all this unforeseen stuff that comes out of mergers that we're starting to sort of see in the AI world, it feels like.

Amanda Silberling [00:09:35]:
No, yeah, for sure. I think that people in the Twitch community were sort of like, so Amazon owns Twitch now, but you haven't really seen that come into play a ton yet, except that like, if you have Amazon Prime, you can subscribe to someone on Twitch for free. So creators are always like, use your Prime subs and then I get free money and you don't have to spend any money. But now it's like— Also, I guess Twitch is kind of pegging this on Amazon, which I'm I mean, I don't blame them. Like, it probably is something that's coming from Amazon more so than Twitch because this kind of data is more valuable to Amazon.

Jacob Ward [00:10:13]:
But yeah, I mean, all of those companies, like, you know, Amazon and all of them are like, they're desperately trying to make them some plan for revenue on the back of all this expenditure they're, they're making on data centers and the rest of it. They need a way to make money off it. And so, yeah, I think that's— And, and the training data is such a important and valuable thing in a landscape where they're— if all your models are all trained on the same data, then there's no differentiation. In this case, they get a big differentiator with everything coming out of Twitch. That's really interesting.

Amanda Silberling [00:10:43]:
Yeah.

Jacob Ward [00:10:44]:
And do you think that there— is it at all clear to you how and whether someone really can opt out of this? And in your experience, do you think that this kind of opt out, which clearly means that you are opted in unless you say otherwise. Does that tend to— do people tend to know about that? Like, do people tend to opt out?

Amanda Silberling [00:11:09]:
Yeah, that's an interesting question. And one that was what a lot of people were mad about was that it is opt-in by default. And the fact that they even gave people the opportunity to opt out, which this is kind of like a corporate talking point, but it's like she kind of has a point that like the Twitch head of community was like, at least we're letting you opt out because other platforms don't. And it's like, you have a point, but also people can still be mad about this.

Jacob Ward [00:11:37]:
Yeah. Are these the standards we're holding ourselves to? That's right.

Amanda Silberling [00:11:40]:
Yeah. So then the CPO, Mike Minton, said, this is a direct quote that I'm reading, why is it not opt-in? That's what everyone's spamming in chat. I get it. Let me opt out versus Or let me opt in versus making me opt out? Well, there's an honest answer. If this was opt-in, nobody would opt in.

Jacob Ward [00:12:01]:
Yeah, right, right. Isn't that funny? It's saying the quiet part out loud, right? Exactly. Because nobody actually wants to be part of this in their conscious— right, in their waking world. Like, they're not— no one's saying, you know what I want to do today is hand all of this content I've put over to these systems. It's interesting too because this is happening on the— and this is going to kind of lead us into my story of the week, like It really pushes back against what a lot of companies are doing, and I would argue being forced to do around trying to be a little more explicit about their AI plans and what's getting sucked in as training data and what is slop and what is not. And so clearly there's still such business advantage to hoovering up as much of people's activities as possible. It's just a good reminder that like, that's still going to be the priority for quite a while for a lot of these companies.

Amanda Silberling [00:13:03]:
Yeah, no, it's definitely kind of— it's like a big treasure trove of footage from the perspective of Amazon. And then from the perspective of creators, it's cool. Now I'm going to be used to train Amazon's AI unless if I opt out from this thing, which I mean, if you know about it, then you probably know enough to opt out of it. But not every creator is really keeping up to date with like the Twitch patch notes they call their updates. Like, yeah.

Jacob Ward [00:13:32]:
John Ashley, our esteemed producer, points out that during the 2024 elections, there was AI-generated streams on Twitch of Biden and Trump having mock debates that were given prompts by chat users. So, you know, we can imagine many, many use cases for what they're going to do with this, right? From Amazon Studios, the movie makers, through to, you know, some sort of customer service agent some kind of weird AI-driven virtual visual Alexa, but could very easily be AI-generated people on Twitch chatting it up who never have to sleep and can speak any language. That could definitely be in the offering.

Adam Rogers [00:14:11]:
Oh man, yeah.

Jacob Ward [00:14:11]:
So Amanda, we're going to do my story of the week now. And mine has to do with— so when I worked at NBC News, I tried to pitch the Today Show a story and I And I said at the time, it's— there's a story about how California is about to have more strict privacy regulations than Europe. And the 2 producers laughed in my face. And one of them said, that's the worst pitch I've ever heard. And I was like, oh no, why? And they're like, that's so boring. You just said California, privacy regulations, and Europe. And I was like, oh, never mind. So I learned how to— not to pitch that kind of thing anymore.

Jacob Ward [00:14:51]:
But—

Amanda Silberling [00:14:52]:
Yeah, it's a tech nerd thing.

Jacob Ward [00:14:53]:
It's a tech nerd thing, right? So here's my exciting story about European tech regulations, 'cause they're actually a really, really big deal, which is—

Amanda Silberling [00:15:00]:
They are.

Jacob Ward [00:15:00]:
They're a really big deal 'cause they're the only ones we get. So as of August 2nd, according to the EU AI Act, and if you wanna make a Silicon Valley executive angry, talk to them about EU AI regulations. The EU AI Act requires under something called the Transparency Code, it's actually a, it's weird, it's kind of a voluntary code, but you can be fined for violating it. You are required as a model maker, right? So this is Google, Microsoft, Apple, Amazon, OpenAI, Anthropic, all of them. They're required if they serve European customers to make it such that they in some way make the AI-generated content detectable. And so what's been quietly happening the last few months is that these companies are beginning to make that happen. And it reads from the outside, if you don't know about the EU regulation, as if it's some kind of cultural movement. Substack is making it so that you have to do this, Suno, the rest of them.

Jacob Ward [00:16:05]:
But it's all, I would argue, because of the regulations. Now, I go on and on about this at theripcurrent.com because I was looking for parallel historical episodes in which an important regulation forced companies to basically make the invisible visible. The one that I really bumped into and really nerded out on, got really focused on, was in 1937, a school in Texas exploded. This is a horrible story about where almost 300 people died in this sudden explosion. A woodshop teacher turned on an electric sander and the spark ignited the whole school and all these people were killed. The fuel that was ignited was an incredible amount of natural gas that had pooled in the walls and the crawl spaces of this school that the school district had sort of bought on the cheap from this pipeline nearby. This is East Texas where the oil fields are all happening. The school explodes.

Jacob Ward [00:17:09]:
All these parents who are working nearby come and try and dig kids out. I mean, it was just a nightmare. Walter Cronkite guy covered this as a young person and said it was the scariest, worst story he ever experienced. He covered Nuremberg, and yet he thought this was the worst ever. So it turns out that the reason that that all happened is because natural gas doesn't have an odor. You can't detect it in its natural form. When it comes out of the ground, it has no odor. So Texas began to require a malodorant be added to the gas so that it would give off a rotten egg smell.

Jacob Ward [00:17:42]:
This stuff called mercaptan is now the global standard, right? If you're in your kitchen, this happened to me just the other day. Our pilot light went out, and, and you could tell immediately because it smells awful.

Adam Rogers [00:17:53]:
Yeah.

Jacob Ward [00:17:53]:
And, and the rule is you gotta do that in sufficient quantities that I will be revolted and smell it hard and, and call the utility long before there's an explosion risk, right? And so it feels very parallel to this where, where, you know, you could— you just adding a little something to this So that people can detect it is, it turns out, you know, very doable and very common sense and being required by, you know, it's really only happening because the law requires it. And, uh, and so anyway, so I'll pause there and, and get your reaction to this.

Amanda Silberling [00:18:35]:
Yeah, I mean, that is a terrifying story about the school, but yeah, I mean, it makes sense That's why dangerous odourless gas is— you could see how that became a recipe for disaster. And I think what you're getting at is that that's sort of the vibe of AI right now, where we don't have super clear methods for determining what is or isn't AI.

Jacob Ward [00:19:06]:
That's right.

Amanda Silberling [00:19:07]:
And thankfully, at least it's not going to blow up Yeah, it's not blowing up schools.

Jacob Ward [00:19:11]:
No, no, exactly.

Amanda Silberling [00:19:12]:
Hopefully.

Jacob Ward [00:19:12]:
Sure, sure. But it does all kinds of things, and we don't know about it, right?

Amanda Silberling [00:19:18]:
Yeah, I do think one of the biggest problems with the development of AI is that we just don't know what is synthetic or not anymore. And then there's obviously really serious consequences of that, like even the thing that John just mentioned to us, that there was AI-generated debate livestreams between Biden and Trump, which, I mean, at this point, I think— I don't know if the technology is there that it would look as legitimate, but like the technology is getting better and better and better. And it's by the time we have another presidential election, I wouldn't be surprised if that kind of technology is like able to look legitimate.

Jacob Ward [00:20:01]:
Totally Right: couple beers in you, you wouldn't know that that's not really them, right? Like, I'm— I just the other day was, was, you know, I'm a big music nerd, and my 12-year-old daughter and I, we nerd out on music together all the time. And I played her this song that I was really into, and she's like, Dad, dummy, that's an AI-generated song, you know?

Amanda Silberling [00:20:17]:
Oh, wow.

Jacob Ward [00:20:17]:
And I was like, what? Oh my God. And she was totally right, you know? And, and it is, it's this, it's this, uh, you know, I remember years ago, right when ChatGPT first came out, I was interviewing Margaret Mitchell, who was one of the authors behind this famous Google paper called Stochastic Parrots. That was sort of one of the early warnings about LLMs and the sort of social damage they might do. And she was this sort of big ethicist thinker about AI. And she said at the time, she was like, I just think it's totally unconscionable that you could create a tool that lets people just release this AI-generated text and imagery into the world without any way to identify it and without any way to claw it back. Because it's going to not only like pollute the media landscape, it's also going to make it such that you don't know whether the data you're pulling in for future models is AI-generated or not. Turns out there's a whole danger around what's called model collapse, where if you really feed it its own creation for too long, the model develops these like defects that can never go away. Like, there's all kinds of technical problems with it.

Jacob Ward [00:21:24]:
But I just remember thinking at this time, like, yeah, she's so right. You know, it is such an important Sort of just piece of common sense, you know. But again, like, and I wonder what you think about this, like, like, what is the business case for, like, why is it better for an AI model maker to make, to have the content be undetectable or not easily detectable, as opposed to what they're going to have to do now under these EU rules? Like, why is it better? I don't know the answer to this. And I'm just thinking about it out loud.

Amanda Silberling [00:22:00]:
I Yeah, I guess I could see that maybe let's say that you're like a corporate executive and you're writing your emails to your team or like making PowerPoint presentations or something with like using AI, then maybe even though at this point it's like there's not as much of a stigma in the corporate world around that, but there could be a stigma if you're like, oh, this guy sent an email and it's like he didn't even write it.

Jacob Ward [00:22:29]:
Totally.

Amanda Silberling [00:22:29]:
And I think that if it becomes easier to detect than it already is, then you can more easily sort of catch people in the act. And then that's when it becomes kind of an interesting, like, cultural situation of like, is there an act to be caught in? Like, for some people they're like, well, yeah, obviously I wrote this email with AI. And for some people that's like career-ending.

Jacob Ward [00:22:54]:
Totally. I did a whole piece about how there's this weird cultural moment we're in where the young— where young people are pretty straightforward, honest with, at least with their peers, with one another, about their use of AI. But people in their 40s and 50s, my age, are much more likely to pass that stuff off as their own. And I think you're absolutely right that there's like, it serves the purposes of a company to make it such that you could do that, as opposed to the world that I think we could be entering if this EU regulation has has the effect that it, it, I think, is intended to, which is it could actually create real social embarrassment about it. You know, like, you're not— you're— if you get caught, you know, passing this stuff off as your thinking or as your ideas, you know, that's really interesting. I mean, so then the last thing I'll just say here is, like, technologically, this— it's not clear that they've totally cracked how to do this, right?

Amanda Silberling [00:23:50]:
Mm-hmm.

Jacob Ward [00:23:50]:
Like, there's something called C2PA, which is this particular sort of standard that's become this big industry standard that Adobe and others have been sort of pushing. But then there's also like a watermarking of text where I guess there's some kind of like, it's like a statistical relationship between the words or the letters or the tokens, I don't really know, which then gets carried over even if you cut and paste it into another document, which I thought was interesting. But the, the problem, of course, is like, what are we going to do when someone's like co-writing with AI? You know, like, does that count in this case? You know what I mean?

Amanda Silberling [00:24:36]:
Yeah, I mean, I think maybe the process of flagging these things, it could be done in a way where it's not like black and white, where there maybe is a gray area where something could flag, like, there are parts of this that Claude wrote, but there's also parts of it that Claude didn't write. And being able to decipher that, like, I think just every greater level of specificity we have is useful.

Jacob Ward [00:25:06]:
Totally. Tthis has kind of been like something that's been in uh, just discussions more broadly already is we're starting to see people like lose book deals because their publisher-

Jacob Ward [00:25:24]:
In our world, it's, it's like a hanging offense to use this stuff, whereas people who work in technical positions say that in many cases they're being punished if they don't use AI in their work.

Amanda Silberling [00:25:35]:
Yeah.

Jacob Ward [00:25:36]:
So it is such a funny cultural moment around that stuff, and there's these like tribes developing around it, it seems to me. So I'm really curious how these regulations will change that, you know?

Amanda Silberling [00:25:47]:
Yeah. Well, yeah, I mean, I guess just in terms of like, if there's something like in the cases where people have lost book deals, where if you are able to prove that an author used AI to like write a novel, then there are questions about just, is it borrowing ideas from other writers in a way that maybe is not immediately obvious, but would become immediately obvious or anything in that nature. But with something like if we're talking about writing emails in a corporate setting, at least there's a bit more leeway there. But I think just the more transparency that there is with whether people are not— people are using AI or not is better.

Jacob Ward [00:26:35]:
Yeah, absolutely. That's right. Amanda Silverling from TechCrunch, we really appreciate you being here. Tell us a little bit about how to find you, where to follow you on social media, where do we see your work?

Amanda Silberling [00:26:45]:
Yeah, I am on TechCrunch. You can find all my articles there. And I've been posting on X a little bit more than I have in the past, mostly because I've been making very silly tech-related pottery.

Jacob Ward [00:27:01]:
Oh, that's awesome. Really? That's what you're doing on X? I love that. Oh, that's amazing. What's an example? Like, what's tech-related pottery look like?

Amanda Silberling [00:27:09]:
Um, I made a mug that, like, you know, the exchange that went kind of viral that came up in the discovery of the, uh, like OpenAI Elon Musk case, or, um, or no, I think, no, it was around, um, like it was around when They were looking at documents from the attempted ousting of Sam Altman.

Adam Rogers [00:27:33]:
Yes, I was in that courtroom.

Jacob Ward [00:27:35]:
Oh my God.

Amanda Silberling [00:27:36]:
Tell us.

Jacob Ward [00:27:36]:
Yeah, tell us. I want to hear. So what was it? Yeah.

Amanda Silberling [00:27:39]:
So when Sam Altman texts Mira Murati and says, can you tell me directionally bad or is it directionally good or bad? And then Mira says, directionally very bad. See, there There we go.

Amanda Silberling [00:27:53]:
That's funny.

Amanda Silberling [00:27:54]:
Yeah, there's, uh, I posted those today. Oh my God, that was like the conversation between—

Jacob Ward [00:28:00]:
This is so good. God, this is really— I've always wondered like, what's good merch in our beat? This is good merch. That's good work, Amanda.

Amanda Silberling [00:28:08]:
Well, yeah, this is my, uh, my journalism failsafe.

Jacob Ward [00:28:12]:
Uh, yeah, that's your fallback. Good, you know, good.

Amanda Silberling [00:28:15]:
If the, if the industry just crumbles and burns, then I'm gonna pivot to the silly mug business.

Jacob Ward [00:28:21]:
I will buy your mugs, no question, and promo them myself. All right, Amanda Stirling, we really appreciate your time. Thanks so much for being here.

Amanda Silberling [00:28:28]:
Yeah, thank you.

Jacob Ward [00:28:30]:
Every week we are joined by a couple of journalists, and I got to choose them. That's because I, I'm— I, I have the power. I am He-Man. Uh, and so I brought in, uh, my good friend and never colleague, but we've been friends and peers for a long time, Adam Rogers, longtime writer for Insider, writer for Wired, writer of a couple of really very cool books, a history of booze and the science of it, science of color. These are great books. Anyway, I love what you're thinking about, Adam. Give us a little sense of yourself. Just introduce yourself to people and what you cover, how you're thinking about these things these days.

Adam Rogers [00:29:09]:
I mean, I trained as a science writer, so I mostly— and always still think of myself as a science journalist, even though I haven't covered straight science in a long time. You know, the sort of like, there's in a new journal article, scientists at Caltech report or whatever, but, uh, science and technology. So I, I covered the early days of the internet and the middle days, and then now the late days, maybe.

Jacob Ward [00:29:29]:
Here we are in the latter day. That's exactly right. So you have a piece in, uh, the SF Standard, and, and this is a thing where, as a national audience, bear with us everybody, because I think this is a really important national story in addition to being a story about the Bay Area, where you talk about the way in which money is being cast off by the AI industry, and it's landing in San Francisco in a measurable way. But here in Oakland, not even 10 miles from that where I'm standing right now, those benefits are not being seen. And there's something I think important for us to understand about it in a sort of broader national sense. But first, give us a little summary of what you found out. How did you you look at this and, and what did you find out?

Adam Rogers [00:30:20]:
Right. And as you said, it'll be important to, to say that this is actually relevant in like whatever city people are living in too. But, but in— so, uh, you know, San Francisco gets to be this— they don't like it when you use the word epicenter in San Francisco, it's bad luck— but an epicenter of, um, of, of cash coming in as part of AI investment, as often happens in San Francisco. Uh, Oakland's right across the bridge. and has been pitched in times past as varying its own center of economic engine as cities can be, but also as kind of a bedroom community for San Francisco sometimes, especially for tech workers. When Mayor and Governor Jerry Brown was mayor of Oakland, he pitched it that way, built a lot of new housing and some headquarters for buildings too. The question that we wanted to ask at the SF Standard was like, okay, well, if all this money's coming into San Francisco, is it spreading out? And if it's not, how come? Because Oakland famously has a lot of problems. Oakland's a classic deindustrialized 19th century city.

Adam Rogers [00:31:15]:
Oakland was and remains a hub of Black life in the United States, was one of the destination points of the Great Migration, used to have things like multiple professional sports teams and railroads, and one of California's big ports still has that. And also classically, a lot of problems with relationships with the Oakland Police Department and the people who live there. And so just trying to figure out, well, if the money's coming in, where's it going and is it coming in? And what we found is that really not yet. There are some knife fights going on about finding a place to live. Housing costs go up there in Oakland, the East Bay around it, as much as they do in San Francisco, but in very specific and constrained places, still in the single-family homes in nice neighborhoods. And nice gets defined in different ways, but it's still not much philanthropy. Still not, at the time I was writing, very few businesses moving their headquarters there. And what the history has been in the Bay Area is not that the new tech companies move their headquarters to Oakland or to other East Bay cities, but that the companies that get displaced from office space are the ones that move to the East Bay.

Adam Rogers [00:32:36]:
So your big, your OpenAIs buy the— rent the millions of square feet in San Francisco, and then support companies, insurance companies, and the other businesses in the world, because there still are those, are the ones that move. And that hasn't started to happen yet either, the post-pandemic push against offices. Offices remain vacant in San Francisco as much as they do in Oakland. So Oakland's hopeful. Oakland's trying to develop itself as a place that's more authentic with its own city. It's not just trying want to depend on San Francisco castoffs. They built a lot of housing there, much more so than San Francisco in a lot of ways. Um, and they're trying to make sure that their, their economics are good for companies to come and move there.

Adam Rogers [00:33:15]:
That may be starting to happen. There was just some reporting today that, um, Tesla's opening an AI facility in West Oakland, which is one of the industrial neighborhoods of the city, right adjacent to a place where Jeff Bezos has an AI facility as well. So maybe that's changing.

Jacob Ward [00:33:29]:
That is so interesting. I mean, I, I also I mean, as a longtime Oakland resident, I'm a big booster of this place. And it's like part of the kind of character of living here to be upset at how the riches of the tech boom have driven up costs for everyday people without necessarily bringing in the kind of opportunities for everybody that you would want. So when you think about the AI, in San Francisco specifically, we're seeing a a wealth hitting that place. There's all these stories now about $2,500 for a room without a closet kind of rents. I just heard people talking the other day about $105 cup of coffee at this one new fancy Japanese coffee place.

Adam Rogers [00:34:17]:
Good Lord.

Jacob Ward [00:34:17]:
I mean, it's crazy, right? Crazy things. For me, I fly a great deal of the time around the country and my costs for flying now are so much higher than they used to be. And that's Strait of Hormuz stuff, that's other things. But San Francisco specifically, SFO flights have become really expensive.

Adam Rogers [00:34:38]:
Yeah.

Jacob Ward [00:34:38]:
In part because that's the people that are flying in and out, right? I wonder if you think there's a difference. Is there a measurable difference yet that we've seen or are beginning to detect in how money comes out of this generation of AI companies as opposed to the way it came out of the Netscapes and the AOLs?

Adam Rogers [00:34:56]:
And this, I think, is going to be the key determinant of what a boom looks like, both locally and then nationally. Because not to be a New York Times pitchbot or whatever, but what do we mean by boom? What does that mean? You want that to mean something good, like there's more development, people are making more money, they're spending more money in places like restaurants and museums, the city can build more more civic art.

Adam Rogers [00:35:23]:
You want a boom to be more than just, well, now housing costs a lot and it displaces the people who used to live here. And right now, the way that economic boom, even in San Francisco, so forget about the East Bay for a second, still looks like there are a few people who are getting paid very high salaries and they're bidding up the cost of housing. But are they buying more food and restaurants? Are more restaurants opening?

Jacob Ward [00:35:48]:
Right.

Adam Rogers [00:35:48]:
Are some of the restaurants charging more? For sure. Are the cultural institutions seeing new people on their boards and more philanthropy? No, they're closing. Theaters are closing. Museums are in trouble. SFMOMA is constantly in debt. I think the de Young still is. And when I talked to people for the story about Oakland at these cultural institutions, universally they said, we can't get tech people to donate money to us. They don't buy— they don't even buy art for their own homes.

Jacob Ward [00:36:14]:
Wow.

Adam Rogers [00:36:14]:
Much less donate to the museums. And this is places like the de Young, which is— the de Young Museum in San Francisco is one those old money institutions. Every medium-sized and big city in America has one of those where that's where the old rich people in the city—

Jacob Ward [00:36:28]:
Right. That's where the Monet show comes or whatever. Yeah. Right.

Adam Rogers [00:36:31]:
Exactly. And that doesn't happen even here. They're having trouble getting those folks, the new money, to participate in what used to be old money, to send kids to the public schools so that there's more of a tax base. And some Now, you've probably seen some of the reporting about looking for money to come out of the IPOs. If OpenAI and Anthropic go IPO, they're going to be new millionaires, and all these NGOs and nonprofits are hoping to get money. But there's no evidence that these folks give money to any kind of philanthropic cause.

Jacob Ward [00:37:07]:
Yeah.

Adam Rogers [00:37:07]:
And I don't mean to be tarring with a broad brush. What do I know? Maybe all the new millionaires are going to be like, finally, we can we can go spend money to keep a school open in the Tenderloin or something.

Jacob Ward [00:37:16]:
Yeah.

Adam Rogers [00:37:16]:
I could be totally wrong, but they haven't done it yet. They haven't done it in San Francisco, much less in Oakland. And that's different. Even that old, that first tech boom you and I remember, the 2000s, 2010s tech boom even, saw people trying to get their names onto hospitals and people trying to start schools for lower-income folks adjacent to the daycare for the where they opened their campus and stuff. That hasn't happened this time.

Jacob Ward [00:37:44]:
Unclear whether it will. Structurally, it may also just be a function of how these folks sort of imagine working. Just the math of it doesn't necessarily lend itself to there just being enough jobs to fund the kind of stuff that you're talking about. There's this famous instance in which Sam Altman was on a podcast with Alexis Ohanian in January 2024, and he says that he's got this bet going with his friends, his tech CEO friends, about which financial quarter will see the first billion-dollar one-person company. And he's talking about that like that's cool, like that's a good thing. But from an economist's standpoint, and certainly from an urban economist's standpoint, that is a nightmare because I don't care how many of those companies— I mean, there may be a few of those companies, but we don't all get to be a gardener for that one person. We don't all to be the driver for that one person. That's really problematic.

Jacob Ward [00:38:43]:
That's really problematic.

Adam Rogers [00:38:44]:
Especially if that one person is doing everything by DoorDash and their second home where they spend most of their time is out of the city. And if they're not participating in the cultural life or fabric of the city, and if there aren't more of them— the Wall Street Journal just today called this a jobless boom.

Jacob Ward [00:39:03]:
Right.

Adam Rogers [00:39:03]:
And I'm I've been struck by just— it's kind of— is we— you and I have talked about it as a tech boom. It's— if it's more of a financialization boom, Hmm. then that doesn't have the same impacts. It doesn't mean new— it means new technology companies can't spin off. Like, part of the promise here has been like, well, if everybody gets, um, access to generative AI tools, then they'll be able to spin off new businesses more rapidly. So there's this hope that it's going to be an entrepreneurial boom, and I, I don't know that that's Maybe that's happening. There are people who are saying they're starting business with fewer employees because the AI software can do all of these things, SaaS services and customer service and stuff. But that doesn't look boom.

Adam Rogers [00:39:47]:
That's a boom-like object, I guess.

Jacob Ward [00:39:49]:
Right.

Adam Rogers [00:39:49]:
But it doesn't have the economic ripple effects that we would come to expect from a California boom. And California depends on these booms. And this is what makes makes it not just a local story. California depends on these booms because Proposition 13 means that it's very hard to raise property taxes on both homes and commercial buildings. So they need lots of new businesses and new people moving here to buy new places to live and buy or rent new places to work. That has to happen for the boom to work. And when that happens in California, historically, San Francisco booms have spread nationally. That's been true since gold and since silver and since finance and Fruits and vegetables.

Adam Rogers [00:40:26]:
San Francisco is a cyclical city and those booms always— railroads is another one. They spread out nationally and then maybe they contract as well. So the question is whether this one can do that if it's structured the way— if it's just a boom-like object, if it's a jobless boom, a financialization boom, a boom that doesn't even spread to the East Bay, can it spread any farther than San Francisco?

Jacob Ward [00:40:45]:
That's really interesting. So to close this out, and again, trying to think about this nationally, I've been seeing— so we We have this somewhat new mayor, Daniel Lurie, in San Francisco, who, as I understand, has been trying to pitch these companies on creating a pool of money to improve the city and is supposedly getting nowhere. When you think about cities across the country that are going to presumably see a boom in this kind of company, if we're even calling these companies, and if we're even calling this a boom, do you think that there's a lesson yet for, let's say, the leaders of those companies— I'm sorry, of those cities, the mayors of those cities, for like, do we have any idea how they should be talking to these companies or preparing themselves for this, or at least what expectations they should maybe manage about this stuff?

Adam Rogers [00:41:39]:
Yeah. I mean, what even is their CapEx? Like, what are the companies— an AI company spends— a big AI company spends money on building a data center, which doesn't bring any jobs, doesn't bring much taxes to the cities that they build in. It's one of the reasons that local people oppose the construction of data centers, among the environmental consequences and all the other stuff, but also, well, they don't get anything in return. So are offices— is it going to be like Google where Google opens a new office in big cities all over the country and then 50 people work in it? They don't seem to be doing that. So what can you ask for? How do you lure them to the city with promises of, we won't charge you for— we'll give you a tax tax break. We won't charge you for your utilities as much. We'll build a facility that has the power needs that you do. Are they going to— is Elon Musk going to build Optimus robots in every town? Is that the promise? He's not even building them here really yet.

Adam Rogers [00:42:32]:
What things can they open? I think one of the things that, like Daniel Lurie in San Francisco, and Oakland does this too, is figuring out ways to make the city a better place to be for people.

Jacob Ward [00:42:44]:
Yeah.

Adam Rogers [00:42:45]:
And so that's, that, that's the kind of demand side instead of the supply side. Can you build a lot of nice protected bike paths? Can you build more housing so it's cheaper to live there? Can you make it easier for a small family-owned restaurant to open, um, or make it easier for them to park the truck that they were— like, all that stuff that makes it more pleasant to be in a place. Because San Francisco is still, and Oakland is, very pleasant to be in. Oakland is an amazing city, right? It sure is.

Jacob Ward [00:43:11]:
It's great.

Adam Rogers [00:43:12]:
The people are great, the food's great, the place is good, it's great to walk around there, biking, it's wonderful, right? So, you, you know, maybe you work on that.

Jacob Ward [00:43:20]:
Right. That's really, that's really smart and interesting, right? Maybe the sort of secondary effect of this, even if the primary effect is sort of disappointing in terms of the AI boom, maybe the secondary effect is that it is going to help us, you know, or help city leaders understand like, oh, the thing that really is going to define my legacy is not the office space I created. It's the cool block parties I got going. It's the farmers markets. It's the pop-ups, all of that stuff.

Adam Rogers [00:43:49]:
What's the boom supposed to bring if not that? That's what's supposed to be the secondary effect. Well, maybe you make that the primary effect. You work on doing that, and then maybe the companies come. Maybe, maybe. I'm being hopeful.

Jacob Ward [00:43:58]:
I love it. I love it. Well, I really appreciate you bringing this to us. It's super interesting. Adam, tell us a little bit about where people can find you, what What else have you got coming up? What, what can we do to promote what you do?

Adam Rogers [00:44:07]:
Uh, I got a passel of features coming. They're all— they all got to come out of the keyboard in front of me and turn into story. But I got some stuff coming in Scientific American. I got some stuff coming in, uh, Mother Jones and in Alta. It's good. There's some good stories coming, um, that I'm looking forward to, but I unfortunately don't have links to them yet because someone has to write them.

Jacob Ward [00:44:25]:
All right, good. Well, we'll put you into our Google News search and Google News alerts. Appreciate you then. Okay, Adam Rogers, thanks so much for being here. you.

Adam Rogers [00:44:33]:
Yeah, thanks, Jake. Take care.

Jacob Ward [00:44:35]:
So for our final journalist of this Tech News Weekly episode, we've brought in Sharon Goldman. Sharon is a longtime tech journalist and founder of Ground Level AI. groundlevel-ai.com is where you can see it. Um, Sharon, great to see you. Thank you so much for being here.

Sharon Goldman [00:44:49]:
Thanks so much. Great to see you as well.

Jacob Ward [00:44:51]:
Yeah, so I reached out to you because I saw that you had been at Black Hat, the big cyber conference. And clearly, the sort of the hot topic around AI recently has been the capability of AI to scare the bejesus out of everybody in terms of its hacking abilities. And then that all, of course, came to a head with this recent Hugging Face attack by OpenAI's creations. And you had all kinds of experiences, it sounds like, interviews at Black Hat. And so I was so interested to just hear Like, what is the mood on the ground in that place about AI generally, when it comes to that world and cybersecurity, and then specifically around the Hugging Face hack and how that was sort of allowed to happen? So tell us what it was like to be there and what it's like in this AI era.

Sharon Goldman [00:45:44]:
Yeah, absolutely. Well, I will say that if you go to a cybersecurity conference like Black Hat, you'll find that cybersecurity folks tend to be pretty relaxed and on the chill side, not terribly panicky about things. But I will say that this OpenAI Hugging Face briefing in which 2 OpenAI researchers came and kind of gave the first detailed breakdown of what had actually happened with that Hugging Face exploit, you know, there were definitely some people who told me that they thought it was a little bit of marketing spin and rolled their eyes. But most people were definitely concerned about it and they took it very seriously. If anything, they just wanted more details. They felt like the briefing didn't give as much detail as they would like, enough data, enough logs to fully understand what happened. But they were definitely concerned about it, not just for this specific situation, but what will happen, what will happen coming down the line.

Jacob Ward [00:46:41]:
So what, what do you think? What's the vibe? vibe from folks there in general about like the capabilities of AI, right? Claude Mythos, their big product, got that company into a lot of trouble with the White House, which seems to have sort of panicked around its capabilities. There was reports of like top executives from big banks getting pulled out of dinner to meet with Scott Bessent, the Treasury Secretary. You know, like there's— that thing really set off panic in the sort of mainstream circles. Are Are black hat folks, like you say, they're very chill a lot of the time. It's a laid back scene, but do they, generally speaking, was the vibe that you got that they are, that they feel that there are new capabilities coming online because of this?

Sharon Goldman [00:47:23]:
Absolutely. I mean, one of the reasons I was so interested in going this year was because of what happened with Mythos and Fable. I'm like, I wanna hear what these folks are saying because people I speak to are very concerned concerned about it. And I will say that, you know, this was my 3rd time at Black Hat. And the first time I went 2 years ago, AI was not nearly as front and center as it is now. So this time, I mean, AI really sucked the air out of the room of every single session and briefing and hallway exchange you might have, but for good reason. People are very concerned about it.

Sharon Goldman [00:47:57]:
It's really a combination of the speed of the change in capabilities, like for models like Mythos, as well as the scale that defenders need to get to, to kind of match that. Because Mythos can be used for good and for bad. So you can use it for the— the attackers can use it, but so can the defenders. At the moment, of course, there are only some companies that have access to Mythos. But the idea is that all security organizations will have to rise to the occasion, both in terms of speeding up their processes and scaling up their defense to sort of match what's, what's coming down the pike from attackers. So every single CISO, every single security organization is talking about that now.

Jacob Ward [00:48:40]:
That's so interesting. I remember going to— I went to a conference of CTOs, like big Fortune 500 CTOs once. It was like this sort of— it was this very crazy gathering of like very high-level people, uh, and, you know, CTO of Walmart, CTO of, you know, Chase, like these various places. And that crowd, I just remember discovering like just how like well-informed those folks were about global politics, about, you know, the state of, of the— this was before the Ukraine invasion, but they were— they knew all about that. Like they, they were so hip to geopolitics geopolitics as, you know, because they needed to know what part of the world is under financial or geopolitical stress and therefore is most likely to spawn an industry of attackers because they don't have the money but they do have the technology, blah, blah, blah. And I can't even kind of imagine how they're going to adapt, yeah, to a world in which you don't necessarily need much in the way of expertise to stage a very effective attack. Like you say, the last time I checked, the United States had 1,000 Like scaling up, I would imagine, must be a source of real terror to people that are— that have in the past been able to at least rely on the idea that you— it's just a, you know, the limiting factor was just like how many hours a person could stay awake. And that's not true anymore.

Sharon Goldman [00:49:57]:
It is true. On the other hand, I will say that I've spoken to a couple of experts who have pointed out that attackers also have budgets. Attackers also have— limitations. So they're, you know, in a, you know, attack on a system, a company system, there are still many steps involved. And in the same way that defenders kind of have many systems that they have to defend, the attackers have to get through many systems. So it's true that, you know, this is scary and organizations have to step up, but it's not true that like right— like you'll notice And you'll notice that we haven't— it's true that the OpenAI Hugging Face exploit was concerning, but there wasn't actual damage done in particular. And you'll notice that up until now, we haven't had this kind of like crazy hack that's taken down critical infrastructure yet. So, you know, there are— and these tools in a way have been around for over a year.

Sharon Goldman [00:50:56]:
Like last year at Black Hat, I had interviewed I interviewed the Anthropic Frontier Red Team about their cybersecurity capabilities and their models, and they were already talking about this sort of thing coming. So, you know, I think that it's definitely a concern, but the defenders are working to catch up. And if they can automate more, you know, just like the attackers are, there's a possibility that they can catch up. Yeah.

Jacob Ward [00:51:25]:
Yeah. So when you look at the reaction at Black Hat, specifically to the Hugging Face attack, tell us more about that. Like, was this— is the sense that there was a sort of a malpractice here? I mean, you know, the way that the OpenAI attack on Hugging Face has been covered, it's hard to see, you know, whether this was just a matter of like, Is the attitude that this thing is so sophisticated it was going to get away from somebody at some point, or if this was, you know, they set this thing up in a kind of loose set of constraints and walked away when they should have kept tighter control over it? Like, what was— is the attitude that this was a good and honest mistake or that this was something that they should have seen coming?

Sharon Goldman [00:52:10]:
I feel like it was a very mixed bag. There were definitely people who seemed to think that, you know, they gave a little side eye, a little sigh of relief that this thing was finally out of control. Cynical, thinking that this was really kind of a marketing play, a PR play. Even at the last keynote, they call it a locknote, there were 4 folks from the Black Hat Review Board on stage from security organizations. And one of them told me that— told the crowd that this was kind of like, you know, felony marketing kind of thing. You know, they're kind of excited to tell us how scary it is and how cool it is.

Jacob Ward [00:52:44]:
There is something to that, right? Like, let's, let's stay on that for a second because that's exactly right. Like, it is interesting, like, even the language in the blog post from OpenAI, they referred to it as unprecedented, right? And I was like, boy, that's kind of markety talk, you know?

Sharon Goldman [00:52:58]:
Like, particularly at the end of the briefing, they talked a lot about how defense really needs to step up. However, it's not a coincidence that at Black Hat and DEF CON, OpenAI had a booth to talk about their product, cybersecurity product, that they're putting out for defenders. So there's definitely that.

Jacob Ward [00:53:20]:
Okay.

Sharon Goldman [00:53:20]:
But there was also a lot of real concern from people. And again, people were also concerned that it was— yes, it was transparent, but was it transparent enough? There weren't really enough specific details necessarily that you could really see what happened and what you could learn from it. So I feel like there were people saying that OpenAI needs to do more and all of these companies need to do more to be transparent. And that would give other organizations a chance to continue to do more research about these capabilities.

Jacob Ward [00:53:57]:
And are people kind of, you know, does Do you think that community is— and I don't mean to make you speak for them as a monolith, but do they get— there's an attitude in the circles that I travel in here in Silicon Valley that this stuff has to be made, it has to progress. There's no— the phrase people always love to use is there's no putting the toothpaste back in the tube, which I think is nuts because we put oceans of toothpaste into tubes every You know, like, that's crazy. But whatever. Like, there's this idea that, like, it had— the show has to go on, and thus we must just adapt. Whereas you also, like, around the same time as these hacks were taking place, and you had, like, this big coalition of OpenAI people coming together led by Jensen Huang of NVIDIA, you also had this, like, 1,300-person letter go to the White House basically saying, please slow us down. Like, we can't help ourselves, was essentially the spirit of this letter signed by all the major players and all the top scientists at these companies, basically saying the competitive pressure we feel means we can't kind of control ourselves really. Stop us before we strike again is sort of the vibe I got. And I'm curious, like, does the— like, at Black Hat and in general, because I know you move in these circles too, like, what's your vibe But what vibe do you get from people about like, has this gotten out of control? Or is this just the natural and understandable extension of the pace at which the industry is moving?

Sharon Goldman [00:55:30]:
I mean, you do have a— I'm sure you know, there's a lot of different tribes that frequent the AI circles in Silicon Valley in particular. Here at Black Hat and DEF CON, you're really, you have a little bit of that, but you're definitely more into a lot of the traditional cybersecurity people, a lot of people who are ethical hackers or were hackers generally. So, you know, on the one hand, I'd say that they're a little bit more like, really, you're not slowing down because your competitors and your enemies are not slowing down. You know, they're really interested in, you know, doing what they can to defend. But at Black Hat, there is is there was an increasing presence of, you know, AI, quote unquote, safety folks, which were a lot of the folks that signed that letter, you know, that aren't necessarily security people even necessarily. They, you know, might be researchers who are very concerned about what they call existential risks of the technology. And cyber sort of is at the front end of that because, you know, people aren't concerned just about the cybersecurity capabilities on their own, but what that could mean for further capabilities that they believe to be an existential risk to humanity, whether that be bio risks, like someone could create a pandemic.

Jacob Ward [00:56:52]:
Sure.

Sharon Goldman [00:56:52]:
Or things like that. But again, I feel like the traditional crowd at Black Hat and DEF CON is much more practically minded. You know, they're looking how to— how things break and how to hack things. So, you know, I think they were more looking at this as, you know, we want to know more about this exploit and this hack because we want to know how to defend ourselves.

Jacob Ward [00:57:18]:
And it sounds, if I'm understanding you correctly, as if they did not feel they got the details that they needed from OpenAI about this.

Sharon Goldman [00:57:26]:
No, I mean, you know, they got summaries and they got, you know, on the slides there were some details, But I think they're talking about much more deep, like actually release the logs. Like, you know, I don't know if you saw the briefing, but they had some snippets from the logs, things that the agents were saying to each other and stuff, but not the full logs. And I guess traditionally when you would have this kind of hack, you would release all of that and let everyone see it.

Jacob Ward [00:57:52]:
Right. It says something about the sort of opacity of the industry that they they are not sharing that stuff. And I think we just forget how closed and secretive these companies really are around this stuff, even when in theory they're trying to let the community that could help them figure this stuff out in on some of their internal problems.

Sharon Goldman [00:58:13]:
One thing that I do think is arising out of this is a call for, you know, a real national policy on AI and security. Regulation around it. You know, that's something that, you know, before Mythos, certainly I don't think we were really hearing that quite as much from DC. But, you know, certainly with Mythos and Fable being shut down, you know, there was talk about that. And now I feel like there's even more people calling for that kind of thing, which is interesting. It really— there's a real overlap now between AI security and policy, and the potential of regulation. And that's something that's fairly new. I don't remember hearing as much about that last year.

Jacob Ward [00:59:00]:
Well, Sharon, we really appreciate you being here. Sharon Goldman from Ground Level AI. Tell us a little bit about Ground Level AI and where we can follow you, what you're thinking about these days.

Sharon Goldman [00:59:07]:
Yeah, Ground Level AI. I've been a longtime reporter on AI. But right now, I'm really kind of focused on— I call I call it when AI meets the real world. And that's things like this, AI and cybersecurity, AI infrastructure, you know, whether it's data centers or, you know, digital infrastructure, AI policy, AI in the enterprise. So I'm really focused on that. And it's at groundlevel-ai.com, or you can just look at Substack and you'll find me.

Jacob Ward [00:59:33]:
Tremendous. Sharon Goldman, thank you so much for being here. We really appreciate you.

Sharon Goldman [00:59:36]:
Thank you, Jacob.

Jacob Ward [00:59:37]:
Thank you. All right, you guys, we have done it. Another episode of Tech News Weekly. And I really just want to thank all of you for being here. This has been a tremendous experience to— it's always so nice to be with you here. And I'm going to shamelessly plug my own work. theripcurrent.com is where I do what I do. I mostly think about the ways in which technology, especially AI, is shaping behavior.

Jacob Ward [01:00:04]:
In the coming weeks, I'm going to be looking at the social media trials that are coming up. There's a big one here in Oakland that starts on Monday. another one in front of the same judge that starts in February that are gonna truly determine the future of social media. I think they're hugely impactful cases. And so I'm gonna be following those closely along with all of the other ways that all of these tech forces are invisibly shaping our behavior and our experience. So thank you so much for being here. This is Tech News Weekly, and we will see you next time. Thank you.