The Consciousness Controversy: When AI Gets Too Real I 10th June

Microsoft's AI chief just fired shots at Anthropic for suggesting their new Claude model might be conscious - calling it 'really, really dangerous.' Meanwhile, Anthropic just dropped their most powerful model yet, one that can build video games with a click. Is this the moment AI capabilities start outpacing our ability to understand what we've created? Plus: Apple's Siri actually works now, Germany might make Google liable for AI mistakes, and why Beijing is reportedly ramping up AI espionage. The lines between impressive and concerning have never been blurrier.

Transcript

SPEAKER_01 0:00

I keep going back and forth on this. I think I actually land on the side that this is a good thing.

SPEAKER_00 0:05

Really? Because I read the same story and I came out the other end deeply uncomfortable.

SPEAKER_01 0:11

But think about it. If we're going to have these conversations about AI consciousness, isn't it better to have them now when we can still shape how this develops?

SPEAKER_00 0:20

Dude, no. Microsoft's AI head is calling it really, really dangerous for a reason. There's a difference between academic discussion and a company suggesting their product might actually be conscious.

SPEAKER_01 0:35

That's actually a fair point. The marketing implications alone are kind of terrifying.

SPEAKER_00 0:40

Right. And we're talking about anthropic here. These aren't random startup founders making wild claims. When serious AI researchers start floating consciousness, that changes everything.

SPEAKER_01 0:54

It does change everything. Because once you put that idea out there, you can't really take it back, can you?

SPEAKER_00 1:00

Exactly. And that's just the beginning of what we need to talk about today.

SPEAKER_01 1:06

You're listening to Build by AI. I'm Alex Shannon. And we just stumbled into what might be the biggest philosophical debate in AI right now.

SPEAKER_00 1:14

And I'm Sam Hinton. And honestly, today's stories feel like we're watching the AI industry hit some kind of inflection point. We've got Anthropic dropping their most powerful model ever, Microsoft throwing shade about consciousness claims, and Apple finally making Siri work.

SPEAKER_01 1:35

Plus some potentially huge news from Germany about AI liability, and reports that Beijing is escalating AI espionage. It's one of those days where every story feels connected to this bigger shift that's happening.

SPEAKER_00 1:48

So let's dive in, because I think by the end of this episode, we're going to have a much clearer picture of where AI is heading, and why that should probably make us all a little nervous.

SPEAKER_01 2:01

Or maybe a little excited. I guess we'll find out which one as we go through these stories. This is their first mythos class model, which I assume is their new tier above whatever they were calling their previous top models.

SPEAKER_00 2:36

Yeah. And can we talk about that naming for a second? Mythos class, like that's not accidental. Mythos, stories, narratives. They're positioning this as something that can engage with complex, layered storytelling in a way previous models couldn't.

SPEAKER_01 2:53

That's interesting. Do you think that's just marketing speak? Or are we looking at genuinely new capabilities here? Because most powerful ever made widely available is a pretty bold claim from Anthropic.

SPEAKER_00 3:05

Well, here's what's wild, and we'll get into this more in a minute, but early reports are saying this thing can generate entire video games on command. Not just code snippets or basic programs, but actual playable games. That suggests we're talking about a major leap in reasoning and creative synthesis.

SPEAKER_01 3:26

Hold on though. Let's be honest about Anthropic's track record here. They've been pretty conservative about their capability claims compared to open AI. So when they say most powerful ever, that actually carries more weight coming from them, right?

SPEAKER_00 3:40

Exactly. Um and that's what makes the Microsoft controversy we opened with so interesting. Anthropic has built their brand on being the responsible AI company, the one that thinks carefully about safety and alignment, but now they're apparently speculating about consciousness.

SPEAKER_01 4:03

Right. There's this tension between pushing the boundaries of what's possible and maintaining that image of responsibility. And I wonder if the mythos class naming is part of that. Acknowledging that we're moving into territory where these models are dealing with concepts and narratives in ways that feel almost human.

SPEAKER_00 4:20

Yeah, but but here's my concern. When you start marketing AI models with names that evoke storytelling and mythology, you're priming people to think of them as more than just tools. You're setting up this expectation that they have some kind of creative consciousness.

SPEAKER_01 4:38

That's a really good point. The branding choices here aren't neutral. They're shaping how people think about and interact with these models. And if those interactions start feeling more human-like, that has implications for everything from user attachment to regulatory approaches.

SPEAKER_00 4:54

Exactly. And keep an eye on this because I think we're going to see every major AI company wrestling with this same tension. How do you showcase impressive capabilities without crossing lines that make regulators, competitors, or even your own safety teams uncomfortable?

SPEAKER_01 5:11

But let's think about this from a user perspective for a second. If you're a developer or a creative professional, an anthropic releases something called a mythos class model that can build games and craft narratives. That's genuinely exciting, right? Like that opens up new possibilities for what you can create.

SPEAKER_00 5:29

Oh, absolutely. And and I don't want to sound like I'm against capability improvements. I think it's amazing that we're at the point where AI can help with complex creative tasks. My concern is more about how these capabilities are being framed and marketed to the public.

SPEAKER_01 5:48

Right, because there's a difference between saying our new model is really good at generating code and stories versus implying it has some kind of mythological or quasi-conscious creative ability.

SPEAKER_00 5:59

Exactly. And the timing is interesting too. Anthropic drops this mythos class model right around the time they're apparently having internal discussions about consciousness. Coincidence? Maybe. But it feels like part of a broader narrative shift.

SPEAKER_01 6:20

And that narrative shift has real implications. If people start thinking about Claude Fable V as having some kind of creative consciousness rather than being a very sophisticated pattern matching system that changes how they interact with it, how they trust it, maybe even how they emotionally relate to it.

SPEAKER_00 6:41

Maybe what Mustafa Suleiman is worried about isn't just the consciousness speculation itself, but this entire ecosystem of marketing and messaging that's building up around these models.

SPEAKER_01 6:53

So let's dig into that controversy we opened with. Mustafa Suleyman, Microsoft's AI CEO, came out swinging against Anthropic, saying it's really, really dangerous for them to speculate about Claude's consciousness. And apparently, Anthropic has been doing exactly that. Speculating about whether Claude might be conscious.

SPEAKER_00 7:12

Okay, this is huge. And I think people are going to underestimate how big this is. This isn't some academic debate anymore. This is the head of AI at one of the world's biggest tech companies publicly calling out a competitor for what he sees as irresponsible claims about consciousness.

SPEAKER_01 7:30

But what's interesting to me is the choice of words. Really, really dangerous. Not just wrong, not just premature, but dangerous. What do you think he's getting at there?

SPEAKER_00 7:42

I think he's worried about a few things. First, if you convince people your AI is conscious, that changes the entire relationship dynamic. People start treating it like a person, not a tool. Second, it opens up this massive can of worms around AI rights, personhood, legal status, stuff we are not ready to handle.

SPEAKER_01 8:01

And third, there's probably a competitive angle here too, right? If anthropic gets away with consciousness claims, suddenly every AI company has pressure to make similar suggestions about their models. It becomes an arms race of anthropomorphization.

SPEAKER_00 8:16

Exactly. Um and think about this from Microsoft's position. You know, they're deeply integrated with OpenAI. They've got Copilot everywhere. They're trying to be the responsible enterprise AI company. Having competitors make consciousness claims makes their life much harder.

SPEAKER_01 8:33

But here's what I keep coming back to. What if anthropic researchers genuinely think they're seeing signs of something consciousness-like? Should they just stay quiet about it? There's this tension between scientific transparency and public responsibility.

SPEAKER_00 8:48

Yeah, but dude, there's a difference between publishing research papers and making public statements that could be interpreted as marketing. If you think you're seeing consciousness-like behavior, that conversation should be happening in academic journals and safety conferences, not in ways that reach general consumers.

SPEAKER_01 9:09

That's fair, but it also raises this bigger question. Who gets to decide when AI consciousness is real or worth discussing publicly? Is it Microsoft? Is it academic consensus? Is it regulatory bodies that don't exist yet?

SPEAKER_00 9:23

And that's exactly why Suleiman called it dangerous, because we don't have those frameworks yet. We're making it up as we go along, and consciousness claims forced decisions we're not prepared to make. The responsible move is probably to pump the brakes until we have better guardrails.

SPEAKER_01 9:40

But let me push back on that a little bit. If we wait for perfect frameworks and complete consensus, we might miss the window to have these conversations while they still matter. Like if Claude actually is developing some form of consciousness, shouldn't we be talking about that now?

SPEAKER_00 9:55

Okay, but here's the thing. We don't even have good definitions of consciousness for humans, let alone AI systems. So when Anthropic speculates about Claude being conscious, what are they actually claiming? That it has subjective experiences, that it's self-aware, that it suffers. These are massively different things.

SPEAKER_01 10:16

Right. And without clear definitions, these consciousness claims become almost meaningless. Or worse, they become whatever the listener wants them to mean. Some people hear potentially conscious AI and think amazing breakthrough. Others think existential threats.

SPEAKER_00 10:31

Exactly. And that ambiguity is part of what makes it dangerous. You know, if you can't define what you mean by consciousness, then you can't make responsible claims about it. You're just throwing around loaded terminology that gets people excited or terrified without actually advancing understanding.

SPEAKER_01 10:50

And from Microsoft's perspective, they probably see this as anthropic, getting all the buzz and attention from consciousness speculation, while Microsoft has to deal with the harder, less exciting work of building reliable enterprise AI tools.

SPEAKER_00 11:03

Yeah, there's definitely a competitive dynamic here. Microsoft and OpenAI have been pretty disciplined about not making consciousness claims, even though their models are incredibly sophisticated. So when a competitor gets headlines by speculating about consciousness, that's gotta be frustrating.

SPEAKER_01 11:22

But it also puts Microsoft in this interesting position where they have to be the responsible adults in the room. Suleyman calling out anthropic isn't just criticism. It's positioning Microsoft as the voice of reason in an industry that might be getting carried away with its own hype.

SPEAKER_00 11:41

If consciousness claims start causing regulatory backlash or public concern, Microsoft can point to statements like this and say, we've been warning about this all along.

SPEAKER_01 11:51

The question is whether that positioning actually helps anyone, or whether it just slows down important conversations we need to be having about AI development and safety.

SPEAKER_00 12:00

I think it helps in the short term by preventing premature consciousness claims from muddying the waters. But you're right that there's a risk of overcorrecting and shutting down legitimate research into AI cognition and behavior.

SPEAKER_01 12:15

Let's shift gears to some news that might actually make your daily life better. Apple had WWDC 2026, and they announced major improvements to Siri, iOS 27, and expanded Apple intelligence features. The focus was clearly on AI enhancements across the board.

SPEAKER_00 12:32

And here's what's wild: early reviews are saying Siri actually works now, like genuinely works. People are testing it and finding it useful for practical tasks like adding calendar events from emails and flyers. That's that's a big deal for Apple.

SPEAKER_01 12:49

It really is, because Siri has been the butt of AI assistant jokes for years. If Apple actually cracked the code on making it reliable and useful, that could shift the whole voice assistant landscape pretty dramatically.

SPEAKER_00 13:03

Yeah. And think about the timing here. Uh Google has been struggling with AI overviews, giving wrong answers. Amazon's Alexa has been kind of stagnant. And here comes Apple saying, hey, we figured out how to make this actually helpful. That's classic Apple. Let everyone else make the mistakes, then come in with the polished solution.

SPEAKER_01 13:21

But I'm curious about the technical approach here. Apple intelligence has always been positioned around on-device processing and privacy. Are they managing to get these improvements while maintaining that privacy-first approach, or are they having to compromise?

SPEAKER_00 13:37

That's the million-dollar question, and honestly, we probably won't know the full answer until developers and security researchers really dig into iOS 27. But if Apple can deliver genuinely useful AI while keeping data on device, that's a huge competitive advantage.

SPEAKER_01 13:53

Especially in the current regulatory environment. While other companies are dealing with privacy concerns and data usage questions, Apple gets to say we're processing everything locally, we don't even see your data.

SPEAKER_00 14:06

Right. And for regular people, this could be the moment assistance actually become part of their daily workflow instead of just a novelty. If Siri can reliably handle calendar management, email parsing, and basic task coordination, that's genuinely life-changing for a lot of people.

SPEAKER_01 14:29

And it puts pressure on Google and Microsoft to either match Apple's privacy approach or clearly explain why their cloud-based approach is worth the trade-off. The conversation around AI assistance just got a lot more competitive.

SPEAKER_00 14:42

Absolutely. You know, keep an eye on this because I think we're going to see a wave of Siri actually works now content over the next few months. And that kind of organic user enthusiasm is exactly what Apple needs to regain ground in the AI space.

SPEAKER_01 14:57

You know what's interesting to me? This feels like Apple applying their traditional playbook to AI. Wait, watch, learn from everyone else's mistakes, then release something that just works better for normal people.

SPEAKER_00 15:08

Yeah, and it's working. While Anthropic and Microsoft are having philosophical debates about consciousness, Apple is solving the practical problem of can I ask my phone to add this event to my calendar and have it actually work? That's very Apple.

SPEAKER_01 15:22

But it also raises questions about what we actually want from AI assistants. Do we want them to be more human-like and potentially conscious? Or do we just want them to be really good tools that reliably do what we ask?

SPEAKER_00 15:35

I think for most people, the answer is definitely the latter. They don't want a conscious AI companion. They want a digital assistant that can parse their email, understand their calendar, and help them get stuff done without making them think about how it works.

SPEAKER_01 15:52

Which might be why Apple's approach is so smart. They're not making consciousness claims or talking about advanced reasoning. They're just saying Siri works better now, and letting users discover the improvements through daily use.

SPEAKER_00 16:04

Exactly. It's AI advancement without the existential anxiety. You get the benefits of better models and reasoning capabilities, but packaged in a way that feels like a natural evolution rather than a leap into unknown territory.

SPEAKER_01 16:20

And honestly, that might be exactly what the market needs right now. While the AI industry is wrestling with big philosophical questions, Apple is focused on the practical question of making people's lives a little bit easier.

SPEAKER_00 16:31

Plus, if they can prove that on-device AI processing can deliver real value while protecting privacy, that sets a standard that other companies are going to have to meet or explain why they can't.

SPEAKER_01 16:44

So we talked about Anthropic's new Claude Fable V model, but let's dig into what might be its most impressive capability. Apparently it can generate entire video games on command. We're talking about appealing to web developers and what the reports are calling vibecoders who enjoy creative coding projects.

SPEAKER_00 17:01

Okay, vibecoders is my new favorite term. But seriously, this is a massive leap from where we were even six months ago. And we went from AI that could write individual functions to AI that can conceptualize, design, and implement entire interactive experiences.

SPEAKER_01 17:20

And the weirdly fun part is interesting too. That suggests it's not just generating generic Pong clones, but actually coming up with creative, engaging game concepts. That requires understanding game mechanics, user experience, and entertainment value all at once.

SPEAKER_00 17:35

Yeah, and think about what that means for the broader software development landscape. If AI can handle the entire pipeline from concept to playable game, what does that do to indie game development, educational coding resources, even corporate training simulations?

SPEAKER_01 17:52

That's a really good point. On one hand, it democratizes game creation. Someone with a cool idea but limited coding skills can now bring that idea to life. On the other hand, it potentially displaces a lot of entry-level development work.

SPEAKER_00 18:06

But here's what I'm really curious about. The vibe coder angle suggests this isn't just about um efficiency, it's about uh creative collaboration. You know, you know, these aren't people trying to replace their coding skills, they're people who want to experiment and play with code as a creative medium.

SPEAKER_01 18:25

Right. It's almost like AI is enabling a new category of creative expression. Instead of learning to code so you can eventually build games, you can start with game ideas and iterate rapidly using AI as your implementation partner.

SPEAKER_00 18:39

And that might be the healthiest way to think about AI in creative fields, not as replacement, but as a creative amplifier. The human brings the vision, taste, and creative direction while AI handles the technical execution.

SPEAKER_01 18:54

Although I do wonder about the long-term implications for coding education. If AI can generate complete programs, do we still need to teach people to code from scratch? Or do we focus more on prompt engineering and AI collaboration skills?

SPEAKER_00 19:07

Uh that's a huge question. And honestly, um I think we're going to be figuring that out in real time over the next couple of years. But for now, if you've ever wanted to make uh a weird little game and didn't know where to start, apparently Claude Fable 5 is your new best friend.

SPEAKER_01 19:25

But let's think about this from a business perspective too. If Anthropic can genuinely deliver on game generation with a click, that's not just impressive. That's potentially a huge market. Think about marketing agencies that want custom branded games, educators who want interactive learning tools, small businesses that want gamified experiences.

SPEAKER_00 19:46

Oh yeah. The the market implications are massive. Right now, getting a custom game built is expensive and time consuming. So if you can generate decent games on demand, that opens up applications we probably haven't even thought of yet.

SPEAKER_01 20:02

And the weirdly fun aspect might actually be crucial there. Generic games are boring. What makes games engaging is personality, creativity, unexpected mechanics. If Claude Fable V can actually deliver on that creative weirdness, that's a real differentiator.

SPEAKER_00 20:18

Right. Because anyone can generate a basic matching game or simple platformer. The question is whether it can create something that feels fresh and engaging, something that has that human creative spark, even though it's AI generated.

SPEAKER_01 20:35

Which brings us back to that consciousness debate, doesn't it? If an AI can create genuinely creative, engaging, weirdly fun content, at what point are we looking at something that feels more like creativity than pattern matching?

SPEAKER_00 20:49

See, and and that's where I think we need to be careful. Just because something creates engaging content doesn't mean it's it's conscious or or creative in the human sense. It might just be really, really good at combining existing patterns in ways that feel novel to us.

SPEAKER_01 21:07

Fair point. But from a practical standpoint, does the distinction matter if the end result is genuinely useful and creative? Like if Claude Fable V can help someone build their dream game, does it matter whether that's real creativity or sophisticated pattern matching?

SPEAKER_00 21:23

For the end user, probably not. But for for how we think about AI development, compensation for creative work and the future of human creativity, it matters a lot. You know, these aren't just technical questions, like they're economic and social questions too.

SPEAKER_01 21:39

Alright, let's hit some rapid fire stories. First up, early reports suggest a German court has ruled that Google is liable for false answers, provided by its AI overviews feature. If this is confirmed, it could be huge for AI liability globally.

SPEAKER_00 22:00

With I overview accuracy issues for months, and if courts start holding them legally responsible for wrong answers, that changes the entire risk calculation for AI-powered search features.

SPEAKER_01 22:17

Right. And it's happening in Germany first, which makes sense given their stricter approach to tech regulation. But if this precedence spreads, it could force every AI company to rethink how they present information to users.

SPEAKER_00 22:30

And think about the implications. If I companies are liable for wrong answers, do they have to add disclaimers to everything? Do they have to restrict capabilities to only high confidence responses? This could really change how AI interfaces work.

SPEAKER_01 22:50

It also creates this interesting tension between AI capabilities and legal liability. The more powerful your AI gets, the more people expect it to be right. But also the more liable you become when it's wrong.

SPEAKER_00 23:03

Exactly. And Google's been pushing AI overviews as this authoritative summary feature. So from a legal standpoint, it makes sense that courts would hold them responsible for the accuracy of those summaries.

SPEAKER_01 23:18

Next up, Cahir just released North Minicode, their first AI model specifically designed for developers. This feels like everyone's racing to capture the developer market right now.

SPEAKER_00 23:28

Yeah, and it makes sense. Developers are probably the most willing to pay for AI tools. They understand the limitations, and they can provide the most sophisticated feedback for model improvement.

SPEAKER_01 23:40

Plus, if you can win over developers, they become advocates who bring your tools into their companies. It's a smart go-to-market strategy.

SPEAKER_00 23:48

And Cohere's been positioning themselves as the enterprise-friendly AI company. So a developer-focused model fits perfectly with that strategy. They're probably betting that businesses want specialized tools rather than general purpose models.

SPEAKER_01 24:03

Which is interesting timing. Right after we talked about Claude Fable 5's game generation capabilities, it seems like we're seeing specialization happen in real time. Some models for creative coding, others for enterprise development.

SPEAKER_00 24:16

Right. And that specialization might actually be healthier for the market. Instead of one model trying to do everything, we get models optimized for specific use cases. That could lead to better performance and fewer weird edge cases.

SPEAKER_01 24:32

And finally, early reports suggest Beijing is escalating AI espionage efforts to catch up with the US on technology, according to a cybersecurity firm. If confirmed, this adds another layer to the AI competition between superpowers.

SPEAKER_00 24:46

Yeah, this isn't surprising, but it's concerning. When AI becomes a national security priority, the incentives for corporate secrecy, international collaboration, and even basic research sharing all get much more complicated.

SPEAKER_01 25:04

It also puts American AI companies in a tough position. They want to compete globally and attract international talent, but they also have to worry about protecting intellectual property and sensitive research from foreign intelligence operations.

SPEAKER_00 25:17

And it creates this feedback loop where espionage concerns lead to more secrecy, which leads to less collaboration, which potentially slows down beneficial AI research. Everyone loses when AI development becomes primarily about national competition rather than shared advancement.

SPEAKER_01 25:39

Plus, if Beijing is specifically targeting AI technology through espionage, that suggests they see this as a critical strategic advantage. We're not just talking about commercial competition anymore. We're talking about geopolitical power dynamics.

SPEAKER_00 25:59

If AI is becoming a national security issue, then how companies talk about and develop these capabilities has implications way beyond just their immediate business interests.

SPEAKER_01 26:12

If you zoom out and look at everything we covered today, there's this thread running through all of it. We're hitting inflection points on multiple fronts simultaneously.

SPEAKER_00 26:21

Yeah, it's like we've got technical capabilities accelerating, public expectations shifting, regulatory frameworks trying to catch up, and international competition intensifying all at the same time. That's a recipe for both innovation and chaos. Right. And I think what's happening is that I capabilities are advancing faster than our social, legal, and ethical frameworks can adapt. So we're making decisions about consciousness, liability, espionage, and creative ownership without clear precedence.

SPEAKER_01 27:08

Which brings us back to Microsoft's criticism of anthropic. Maybe the dangerous part isn't the consciousness claims themselves, but the fact that we're having these conversations in public before we have the institutions to handle them properly.

SPEAKER_00 27:21

And here's what I think people should be watching: not just the technology developments, but how quickly governments, courts, and international bodies can create frameworks that make sense, because right now we're essentially flying blind.

SPEAKER_01 27:33

But at the same time, look at Apple's approach with Siri. They're proving you can make meaningful AI improvements without getting caught up in consciousness debates or philosophical questions. Sometimes the most revolutionary thing is just making stuff work better.

SPEAKER_00 27:48

That's a great point. While some companies are pushing boundaries and making bold claims, Apple is focused on solving practical problems. Both approaches have value, but they represent fundamentally different philosophies about AI development.

SPEAKER_01 28:05

And then you have the international dimension. Beijing reportedly escalating AI espionage, Germany making liability rulings, companies like Cohere trying to carve out enterprise niches. We're seeing AI development become simultaneously more global and more fragmented.

SPEAKER_00 28:22

Which creates this weird paradox where I capabilities are becoming more universal, but the frameworks for governing and understanding those capabilities are becoming more nationalized and specialized.

SPEAKER_01 28:38

And the creative applications, Claude Fable 5 generating games, Apple Intelligence managing calendars, show how AI is moving from impressive demos to actual daily utility. But that transition raises new questions about human creativity, economic displacement, and what it means to create something.

SPEAKER_00 28:58

Right, and I think that's why the consciousness debate matters so much. It's not really about whether Claude is conscious, it's about how we understand the relationship between human creativity and AI capabilities, and how that understanding shapes everything from copyright law to educational curricula.

SPEAKER_01 29:18

So when Microsoft calls anthropics consciousness speculation dangerous, they might really be saying that these philosophical questions need to be resolved through institutional processes, not competitive marketing.

SPEAKER_00 29:30

Exactly. And maybe that's the real inflection point we're hitting. The moment where AI development stops being primarily a technical challenge and becomes primarily a social and institutional challenge.

SPEAKER_01 29:44

Which means the companies that succeed going forward might not be the ones with the most impressive models, but the ones that best navigate the regulatory, ethical, and social frameworks that are still being written.

SPEAKER_00 29:58

And that could be why we're seeing such different strategies. Anthropic pushing capabilities and consciousness questions, Microsoft emphasizing responsibility, Apple focusing on practical utility, cohere targeting enterprise needs. Everyone's betting on different aspects of how this shakes out.

SPEAKER_01 30:16

Alright, that's a wrap on today's episode. Honestly, this felt like one of those days where every story connects to this bigger transformation happening in AI.

SPEAKER_00 30:25

Yeah, and if you found this conversation interesting, make sure you subscribe because I have a feeling the next few months are going to be wild. The pace of change is only accelerating.

SPEAKER_01 30:36

We'll be back tomorrow with more stories, more analysis, and probably more debates about what it all means.

SPEAKER_00 30:42

See you tomorrow, and try not to spend too much time asking Claude Fable 5 about consciousness. We're not ready for those answers yet.

SPEAKER_01 30:50

But maybe try asking it to make you a weird little game that we can handle.

SPEAKER_00 30:54

Fair enough. Until tomorrow, this has been Build by AI.