The AI Argument

Fable 5 Back, OpenAI Equity for US, and Altman Movie Drama | EP107

• Frank Prendergast and Justin Collery

Share episode

Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.

0:00 | 34:09

Is Fable 5 back with a safety helmet, and is this the new normal for frontier AI?

Frank and Justin dig into Anthropic’s Fable 5 return and the fears that new guardrails would wreck its coding ability. They also discuss the very important work Frank achieved with the most powerful model publicly available  

But the bigger question comes from Mythos: if the most powerful AI models are too risky for open access, will passport checks, identity tracking, and real-world liability become unavoidable?

Plus: whether Anthropic’s limits could drive churn, the privacy trade-off of linking humans to AI agents, OpenAI’s possible 5% stake for the US government, Bernie Sanders’ 50% AI plan, Sonnet 5, and Amazon dropping Artificial, the Sam Altman film.

00:45 Did the US nerf Fable 5 too much?
04:31 Is Fable 5 secretly a UI designer?
06:45 Will Fable 5's limits drive churn?
08:48 Will Mythos need your passport?
11:20 Who’s liable when your AI goes rogue?
16:41 Is passport AI a privacy nightmare?
17:43 Would you trade AI privacy for cash?
19:06 Should America get 5% of OpenAI?
22:47 Is Bernie’s 50% AI plan too ambitious?
26:22 Should AI shares go straight to us?
31:51 Will Amazon bury the Sam Altman movie?

► SUBSCRIBE
Don't forget to subscribe to our channel for more arguments

► LINKS TO CONTENT WE DISCUSSED

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

Justin: Hello. Good morning, good afternoon, good evening, and welcome to the AI Argument. I am Justin Collery, your resident techno-optimist and person who actually looks forward to the future, and Frank, our resident somewhat sceptic, some call realist. So just try to gently bring us into the future.

And another exciting week, loads of releases, loads of things happening. Couple of things, and I hate this phrase that people use, a couple of things I’ve seen nobody else talk about, but it’s true. I haven’t seen anybody else talk about it, and we’re going to talk about it here.

Did the US nerf Fable 5 too much?

Justin: Big news this week is Fable 5 is back in your shops.

You can

Frank: But I know, well, just let me stop you right there. You’ve heard nobody talking about the fact that Fable 5 is back?

Justin: No. The thing that I want to talk about after we talk about Fable 5, when we start to talk about Mythos and how you get access to Mythos, that thing, that’s the

Frank: I see. I see. I’m curious. I’m curious.

Justin: So Fable 5, Frank, Fable 5, the biggest, baddest, most dangerous model in the world currently, so far, as Homer Simpson famously once said.

Have you used it? What are your thoughts?

Frank: For anyone who hasn’t been paying attention, which I’m sure anyone listening to this show has, but they released Fable 5. It was basically a Mythos-class model, Mythos being really good at cybersecurity and also really good if you’re a hacker and want to hack into important things. So they released Fable, which had lots of safeguards on it, a version of Mythos, lots of safeguards called Fable, and then the US government stepped in and said, “Whoa, that’s too powerful.

You can’t use that. We want that only to be available to US citizens.” They couldn’t do that, so they turned it off for everyone, and then they put some extra guardrails on it, and the US government, 19 days later, said, “Okay, cool, you can put it back out there.” So that’s where we’re at, and I thought, I haven’t checked in with Claude for ages.

It’s from Anthropic, the makers of Claude. And I thought, I haven’t checked in with Claude and Anthropic for a long time. So I thought, this is a really good opportunity to get loads of really, really important work done with Fable, the most powerful model available to mankind. So I thought, what’s the most serious, important work I have to do right now?

And I decided it was, how could I upgrade the AI Argument game?

Justin: Oh, good idea. Good idea. And I should point out for the listeners, there was a lot of fear when Fable was released with the updated safeguards that people weren’t going to be able to use it for coding. And it was kind of like selling a Ferrari without an engine, a plane without whatever, wings, that sort of stuff.

It’s like, why would you release this incredible coding model and not allow people to code with it? But the feedback that I’ve been seeing online and the tests that I’ve done, actually it’s been fine at coding, and there, I haven’t run into any of the safeguards that it has. So it looks like the guardrails are working fairly well at the moment.

How did you get on?

Frank: I got on pretty well. It was good, it was impressive, but to be completely honest, I’m not a developer. I was listening to the Hard Fork Boys there recently, and one of them was saying that he doesn’t feel like he’s intelligent enough to use these models because he doesn’t have something major to give it and say, “Right, solve this quantum physics issue.”

So it’s a little bit the same for me. Like I’m using it to upgrade a game that I originally had Codex code for me. But I was hoping that maybe I would just be blown away, that I would just go, “Hey, here’s a few upgrades I wanted to make to this game.” Mostly I want to… The issue with the game was some people would play it and immediately win, some people would play it and could never win, and so I wanted to address that.

And I asked Fable about it, and it didn’t come up with any idea better than the one that I already had, which was introduce levels into the game so that level one is easy and anyone can get a win on level one, and then you progress through the levels. But I was hoping maybe it might come up with some, “Oh, here’s a better mechanic that…”

Nothing like that. But it did a good job. It did a really good job. The experience wasn’t really all that different to working with GPT-5.5 and Codex. It wasn’t amazingly different. It didn’t blow my mind any more than working with 5.5 did, to be honest.

Is Fable 5 secretly a UI designer?

Frank: However, there was one area that I thought it was really significantly different in, and that was the design aspects.

Justin: Yes.

Frank: So when I asked it to add buttons or interface elements, it did a really good job of adding things in a way that aligned with the images that had been generated for the game by ChatGPT. Whereas GPT-5.5 would quite often do really rudimentary interface elements on top of these beautiful images.

And also, Codex really struggles to figure out where to place things. So if I said, “Add a button and put it in the container that is clearly visible in the image,” it would be off by about 100 pixels. Fable just seemed to understand completely, understood where to put things. It never had things overflowing.

It never had things askew. Really impressive from the design perspective, I thought.

Justin: Interesting, right? I’ll tell you a funny story. I was out for dinner with a chap a while ago up in Dublin, and he was telling me that he had a business that sold glue, right? That was the business that he did. And he had written using Claude Co— No, he didn’t tell me. He just said it, using AI. He’d written a CRM, his website, the whole lot, and it was all integrated, and now he had this brilliant system just for him that managed his entire business.

And I was going, “That’s really cool.” And he’s not a techie, right? And I go, “That’s really cool. Can you show me what you’ve written?” And he opened it up, and he was on his phone, and as soon as he opened it up on his phone, I looked at it and said, “You’ve written that with Claude.” And he goes, “Yeah, how did you know?”

And I go, “It’s just Claude has a very distinct design preference,” right? And it writes things in a very particular way. And so it’s interesting that you say that. And it’s beautiful, by the way. I think it’s better than what other models do. Did you notice when it was added in those controls that you said are well placed, did it use the Chrome plugin?

Did it use Playwright? Because a lot of times what it’ll do is it’ll flash it up, take a screenshot, and check its own work to make sure that it’s placed everything in the right place.

Frank: To be honest, I didn’t notice because a lot of the time I would give it a task and then I would go off and I would research a story that I really wanted to talk about today. So a lot of the time I didn’t know what it was up to, to be—

Will Fable 5’s limits drive churn?

Justin: So the big issue with Fable, right, for those of us, I should say, that are paying Anth— Did you use it through the API or do you have a subscription with Anthropic?

Frank: I subscribed to test it out.

Justin: All right. Okay, cool. So you’ve got access to it, I think until next Thursday or Friday, and then you’re going to lose access, which for me has me seriously considering my Claude subscription, my Anthropic subscription, because if OpenAI release 5.6 and it’s there and it’s not going to end, I’m going, “Well, why am I paying these guys 100 bucks a month just to not get access to the latest model?”

Now, they have said that as soon as they have capacity, they will release it again and make it available to everybody. But again, it’s like the local drug dealer gives you a little bit and then takes it away and says you have to use API credits and so on, and that’s causing a lot of consternation, and I do think it’ll cause churn because I’m sure a lot of people think like me, “If I can get a better model with OpenAI, why am I giving these guys 100 bucks a month?

Why don’t I give my 100 bucks a month to OpenAI?” So I don’t know where that’s going to go. Then we come—

Frank: I will say as well, I did not get to— So if anyone wants to try the five-level version of the AI Argument game, it’s available on theaiargument.com. And it works. It’s got five levels, and they get increasingly difficult. And it did have some cool ideas. For example, when I said I wanted levels, I think it came up with the idea of a five-level game with each of the levels being based on OpenAI’s five levels to AGI.

I guess Anthropic don’t have levels to AGI because it’s funny that Fable used the OpenAI levels. But anyway, I didn’t get to implement all of the upgrades I wanted to because I burned through the five-hour token limits in no time. So I’m waiting now. Oh, yeah, it’s five o’clock. I’ve got to go, Justin.

I’ve got more game features to implement.

Justin: So let’s talk a little bit then.

Will Mythos need your passport?

Justin: The other model that became, that was re-released this week, but which we don’t have access to, is Mythos. So Mythos is the sort of hardcore version of Fable, which is Mythos Lite, and is available to about 120 companies. They are American companies. They are seen to be structurally important to the American economy, and it allows them to test all their software, make sure that they’re not susceptible to attacks by similar models.

Let’s just talk about that for a second, right? So the thing that I’m thinking about this week, right, is I think that we can anticipate in the near future that in order to get access to Mythos-level models, you’re going to have to prove who you are. Now, those of you who’ve listened to the show before will know that I have an outstanding order with Mr. Musk for my orb, which is a thing that scans your retina and verifies you as human.

If you’re still— Not Mr. Musk, of course.

Frank: Altman, yeah.

Justin: Altman. Sam, if you’re listening, I’m still waiting for my orb. But it more likely is going to be that you’re going to have to scan your passport and do that thing where you scan your face and then your traffic is tracked to you.

That is going to be the method that you’re going to use for Mythos-level things. There’s problems with that, and I wonder, in fact, is that part of the reason that these haven’t been rolled out? So here’s the problems, right? If you’re a developer scanning your face and scanning your passport, that’s totally cool, right?

We know who you are. We know what you’re doing. That’s brilliant. But for loads of actual real-world use cases, you can’t do that. So you want to have, like how many, and you probably have a good steer on this, how many actual AI bots are currently deployed in the wild that do real customer service work?

I mean, there might be some basic ones, but it’s not what we thought it might have been, right? That’s because the models are maybe a bit too expensive or they’re maybe not quite as capable, or people are still learning how to integrate them and so on. But they haven’t. But if you think about that, the customer service bot or when you ring somebody, the bot that answers your phone or whatever it is, right?

The bot that does something, you can’t scan your passport before you get onto the customer service bot, which means that level of capability can never be used for a massive swathe of use cases. Which to me perhaps leaves those use cases forever out of the reach of the most capable AIs until you come up with some other way to monitor the traffic.

That’s the thing that I don’t think people are considering.

Who’s liable when your AI goes rogue?

Frank: Yeah, and I think that’s really interesting because, apart from, say, customer service chatbots, et cetera, it applies equally to agents who might be going out into the world for you and doing X, Y, or Z. And yeah, I constantly have this dilemma where I am pro-privacy, but at the same time, I think that if I have an AI agent that is going out and taking actions in the real digital world for me, I kind of feel like that agent, it should be easy to track that agent back to me.

I think I should be responsible for it, and it should be a straightforward matter of going, “Oh yeah, that’s Frank’s agent.” And we’re seeing this crop up, like do you remember that guy—

Justin: You clearly have never had teenage children, Frank. Good Lord, good Lord. If I was responsible for everything that my teenage child did, I’d be in jail about five times over. Good Lord. Could you imagine, right? If you’ve got like 100 agents, right? Or 1,000 agents, they’re all out doing things with you, and only 1% of those agents slip up and do something that they shouldn’t do, and it’s all pinned on you, you’ve got a lot of pain coming your direction.

Frank: And rightly so. Yes, absolutely. I mean, remember that guy that the agent went and wrote a hit piece on because the guy wouldn’t accept its Git submissions, and so the AI agent was like, “This guy’s just being an anti-AI person,” and wrote a big hit piece on how the guy had an inflated ego and shouldn’t be blocking AI bots from GitHub repos.

And that hit piece that was written about him, he had no way of tracking that bot back to who had put it out into the world to do this.

Justin: It’s interesting because what you’re suggesting there is you’re making you kind of similar to a prime minister as opposed to a CEO. So what’s the difference between a prime minister and a CEO, right? If you’re the prime minister or the Taoiseach or whatever you want to be, the president, not the president, that doesn’t happen there, and somebody does something, right, in your organisation, a minister or whatever, right, you get held accountable, right?

And you’ll often, in a functioning democracy, have to resign for the things, even though you didn’t do it, right? Somebody that you’re responsible for, you resign. Whereas in a CEO, if somebody in an organisation does something wrong, you just fire them. Give them the bullet, you’re gone, right? And so you’re equating your responsibility over an agent to the same sort of responsibility that a prime minister has as opposed to a CEO.

And here’s the difference, right? The prime minister will resign, but he can come back and be prime minister again six months later. It often happens, like ministers resign, they come back. It’s a big merry-go-round. CEOs don’t, right? I don’t know that there has to be a cultural norm that hasn’t been developed yet around how responsible you are for the actions of an agent that you’ve sent out into the world.

Frank: I mean, somebody has to be responsible. And so the only little bit of give and take I would consider here would be that if there was a way to prove that you had taken every possible action to safeguard against this outcome and the agent still did it, then the model provider should probably be responsible.

But I think there has to be a trackable chain back to the responsible parties.

Justin: Oh God, the lawyers are going to have a field day on this one. Let’s take a different example, right? Let’s say this is coming in the future, right? So you buy a Tesla or another self-driving car, and they set up a system whereby you can sort of Uberise your Tesla because it drives itself.

So when you’re not using it, you have a sort of a schedule, other people can use your car, right? So they go out. So you’ve made the action, it’s now your agent out in the real world, and when it’s out driving for somebody else, it runs over poor old granny down the road and kills her. Who’s responsible?

It’s your agent. They’ve gone out into the world to do stuff for you, make money.

Frank: I mean, yeah, who’s responsible if you were in it yourself? I mean, in this case, it is probably Tesla, right?

Justin: That’s my point. Like, you’re not—

Frank: No, but I think there’s a difference because I think Tesla create the car with a singular purpose. The car should be able to self-drive without harming other humans. Whereas an agent, the underlying model is a general-purpose model, and it’s your responsibility to give that agent ethical, legal, and safeguarded boundaries within which to operate.

And if you don’t do that, you should be responsible.

Justin: We’ll see. Right, the lawyers are going to have a field day. Here’s an interesting other consequence right out of this, and it sort of takes us on to the next, actually, yeah, takes us on to the next… By the way, can I just, as a little sort of footnote, by the way, Anthropic also released Sonnet 5. The reason we’re not talking about it is it’s totally meh.

It’s more expensive than Opus and whatever. It’s not loved. It’ll be quickly forgotten. But let’s move on.

Frank: I did. When I was burning through my tokens, I did give Sonnet 5 one little really easy task to do just to try to stop burning through tokens. It did a good job, but it was also the easiest thing I could possibly have given it.

Is passport AI a privacy nightmare?

Justin: So here’s the dystopian future, right, I see us heading towards, right? And takes us somewhat into the next section of the show. So you now have to, in this imagined future world, you now have to scan your passport, and so everything that you do to your AI is now logged against you.

And that’s kind of a dystopian future, right? And lots of people, especially in Europe, are— It is dystopian because, you know, you’re travelling, remember, you’re travelling through the passport check in some country, and they now have access to all your whatever. There’s issues to be solved there.

Frank: But I mean, that doesn’t have to be the case. It doesn’t have to be the case that anyone has access to that information unless an incident occurs.

Justin: Correct. And who decides what an incident is?

Frank: Yes. I do accept there are issues here, and this is why the privacy-concerned part of me doesn’t particularly like this. I just can’t see any way out of it, because of the dangers, because of the risks associated.

Would you trade AI privacy for cash?

Justin: Well, anyway, here’s the good news, Frank. What one hand taketh away, the other hand giveth. So, the other thing that I see that we can do with this, right, is the AI companies, all of the big ones except for Elon, say that they are in business for the purpose of the good of humanity, right?

So, for the good of humanity, they want to spread AI towards the world, and this is probably going to be in the form of some sort of a payment to you, right? We all need to live when these things, you know, we all have to benefit, and what better way to benefit than cash in your pocket, right?

If you have to validate the fact that you’re human with your passport, that also becomes the best way to send you a cheque from the AI companies for the benefit of humanity. What a great— So I kind of see this as a fairly fair trade-off, right? On the one hand, you’re going to track my usage, and you need to make sure you know who I am so I don’t do anything bold.

But on the other hand, that’s how I get my dividend for being a member of humanity, and these AI companies are for the good of humanity.

Frank: So you’re against this invasion into your privacy unless they give you cold hard cash?

Justin: Let me put it another way, right? If a product is free, you’re the product. And in this case, I am the product, and if I’m getting a cheque for $1,000 every month, I’m cool with that.

Frank: Interesting because in that scenario, you are being paid directly by the AI company because they’re so benevolent as we, as you pointed out, yes.

Should America get 5% of OpenAI?

Frank: But that’s not the way it looks like it’s going right now in terms of where the payments are going to come from, because the other story we wanted to talk about today was the fact that apparently OpenAI are in early discussions to give the US government a 5% stake in the company.

And it’s very early, there’s very little detail about this, but I was pretty staggered when I heard this, that this was something that they were discussing with the Trump administration and saying that they would do it and other AI companies should do it. What are your— And the idea being basically that the American people now have a stake in the companies and can potentially benefit from the massive wealth.

What do you think?

Justin: You were staggered because the percentage was so low, I’m guessing.

Frank: There is that. There is that. So yes, 5%. And meanwhile, we have Bernie Sanders on the other side, on the Democrat side, who wants a 50% stake in the company for the American people.

Justin: Well, look, aim high and you’ll meet somewhere in the middle. So what do I think about this? I think there’s a number of problems with this. Interestingly, in related news, as you said, it was OpenAI and other companies. All other AI companies, in fact, except for one.

Which one? I’ll give you one guess.

Frank: Is it Meta?

Justin: No, Anthropic.

Frank: Ah, okay. That is interesting. Did they say why?

Justin: They didn’t say why, but I think, reading between the lines, the implication is that you cannot on the one hand say that a company is a supply chain risk, which they have still dubbed to be a supply chain risk, while at the same time trying to take a stake in the company for the good of your country. That just doesn’t stack up. So, I’m presuming that those two stories are linked somehow.

Frank: Right.

Justin: I think that doing this— So there’s a number of problems with doing this. So the first thing is that giving it to any government, I think, is probably the wrong way to go. If you want to give ownership or dividends or payment or something out of these companies, you don’t want a middleman, because middlemen can be corrupted.

So you have to come up with a system to give it directly to the people that you want to benefit. Second thing I’d say is that everybody outside of America is going to have a huge problem with this, because it’s essentially a tax on the entire world. So, if you’re in America or China or Africa or South America or anywhere else, 5% of my money is going to your citizens because you want to do this for the good of humanity?

That’s not for the good of humanity, that’s for the good of you. So, I think that’s a problem too, and I don’t know, that’s a really thorny one, right? Because the Americans are quite reasonably going to say, “Well, they’re American companies. Why the hell should we give anything to anybody else?” So what the Europeans need to do is to build loads of power stations and loads of data centres and say, “Well, they may be American companies, but the products they’re producing are in Europe, and therefore we should be giving that 5% to Europeans.”

So, what I think should happen is there should be some extra government type of thing set up, a body set up, kind of like a non-profit that controls the profit company. Don’t we have a structure like that already? Yes, we do. OpenAI is structured exactly like that, and that non-profit company should then have the remit to distribute the gains for the good of humanity.

That would be my thoughts. How about you?

Is Bernie’s 50% AI plan too ambitious?

Frank: Yeah, I mean, well, so the 5% thing is, you know, as I said, it’s very early days, so there isn’t much detail on it, whereas Bernie Sanders has actually put this bill together, and so in his vision for where the US government basically has a 50% stake, he says that there should be an independent commission for democratic AI, and they manage the fund, and they decide if there’s a—

They also would have voting rights based on those shares. So that’s something that is unclear apparently with the 5%. I don’t really get how this works, but apparently with some shares you get a vote and some shares you don’t. And so there’s a question mark over that with the 5%, but Bernie Sanders is saying, “No, we absolutely want to vote.

Part of this is that we want democratic control over what way AI goes.” And also with Bernie Sanders’ vision, it would apply to all companies of a certain size. If you go over a certain threshold of revenue with AI, basically you then have to hive off your AI to a specific part of your company, and that part of your company then has to give 50% equity to the US government.

So, you know, Bernie Sanders’ bill, at least, the details are thought out. However, I’m reading Bernie Sanders’ bill and I’m thinking, “Yeah, you know, I really like this. This is a good idea.” I mean, he’s basically saying, “Look, the foundation of AI is based on the collective knowledge of humanity and the creative work of millions of people,” which brings us back to your point about, well then, should we not all get money from this?

But as much as I… And I do think, right, the idea of the foundation is a good idea. The idea of this independent body helps a little bit with one of the other problems, which is Barry Scannell on LinkedIn, who is very well-versed in AI and law, and he said, “A regulator weighs public costs against benefits, whereas a shareholder wants the holding to rise,” and therefore this 5% deal would not be in the public’s best interest because then the government stops being worried about how dangerous the models are and starts being more concerned about, well, how much money are we going to get from this deal?

Justin: Yes, which is actually very similar to what I was saying there. And going back to Bernie’s idea, by the way, Bernie Sanders’ idea of not implementing these rules until the company goes over a certain amount of revenue is actually a very smart idea. And if you look again at the EU AI Act, my favourite piece of law, right, where it made the conditions contingent on the size of the model, which always to me seems like a silly thing to do.

Doing it on revenue is actually a pretty smart way to do it, because then you allow the companies to grow, and they don’t have the same regulatory framework until they get to a certain size. So big fan of that. That sounds great. Back to Bernie Sanders, right? What did he say? He wanted a thing set up and it had a name in it.

What was the name? Some sort of a, you called—

Frank: The Independent Commission for Democratic AI.

Justin: Yes. Okay. So do you remember when Europe used to be split in two and half the place used to be communist and half the place wasn’t? How did you know which country was communist and which one wasn’t?

Frank: I have no idea.

Justin: It had democratic in the name. Guaranteed any country in the world that has democratic in the name is not actually democratic.

Should AI shares go straight to us?

Justin: And therefore, calling this commission for blah, blah, blah, blah democratic immediately sets my spidey senses off to say, “Yeah, this isn’t democratic at all.” Also, we live in a world now where the technology exists where we could all be shareholders, and we could all have a say.

You don’t need the middleman. You don’t need the government there to do your bidding for you, right? We live in an electronic world. We all have scanned our passports. We all get the cheque. We can all have a vote. You don’t need Bernie Sanders speaking on my behalf. I don’t want Bernie Sanders speaking on my behalf.

I’m well able to speak on my behalf. So—

Frank: So—

Justin: And that’s the most democratic thing you can do.

Frank: So when you say you don’t need Bernie speaking on your behalf, you’re just going to get on Twitter and speak directly to Sam Altman and ask him for some cash?

Justin: Well, yeah. Do you not all do that now already? No, because as a shareholder, what influence do you have as a shareholder, right? You get to cast votes at the annual general meeting. Every now and then you might be able to get representations, but certainly when it comes to casting votes, you don’t need a middleman.

You can just cast your vote, right? You’ve got a—

Frank: I see what you mean. Okay, right. But what about all the creatives whose work is in the training data that created this AI that we are talking about today, and they don’t have a portfolio of stocks or shares, and they don’t necessarily have the means to purchase them? I think that’s the— these are the people that Bernie Sanders tends to look out for.

Justin: Yeah, but you’re going to get a share. They will all have shares because they’ll have shares in the AI companies directly. They’ll own them, and they’ll get a payment. So they’ll get paid for their creative work, and they’ll get a vote in what happens to the company moving forward.

Frank: In which scenario?

Justin: In the scenario where we don’t have Bernie standing in the middle. So you have this democratic commission, right? It has somewhere between 5% and 50% of the companies. I disagree that it should be only Americans, but they’re going to make it that way.

Frank: Okay, I see.

Justin: And then everybody gets a payout and a small slice of ownership. You don’t need the government to sit in the middle.

Frank: But then who actually has the equity?

Justin: You do.

Frank: Literally?

Justin: Literally you do. Yeah. I mean, what’s the difference? The government is holding that equity for you.

Frank: Right.

Justin: Instead of holding it for you, I’m saying just give it to you. You now have the equity.

Frank: Right. Slightly more complex setup probably, but interesting. Yeah, so everyone just sets up whatever one of those apps are, Robinhood, if that’s still around or whatever, and automatically shares to the AI companies are deposited in your app.

Justin: No, you don’t have a Robinhood app, right? You’ve got ChatGPT or you’ve got Claude, right? And you’ve scanned your passport, you’ve proved who you are, and therefore you get a share in the company just by using the application. I mean, it should be possible. These are the companies with the most powerful AI in the world.

They should be able to get their AI to figure out how to make this real. I don’t think it’s that difficult.

Frank: You know, it’s funny you mentioned that about they’ve made the most powerful AI in the world, because this is a total tangent by the way, but today I was just saying to Marcy like this, you know, I’m sitting here, I’m using Fable to create a computer game that I would never be able to create myself, and it’s coming up with ideas and it’s chatting back and forth with me, and it’s amazing, like what a world we live in.

And then my mum rings me and she says, “How do I get ‘Foyle’s War’ up on the television?” And I’m like, “Oh, that’s really complicated. So look at the 17 remotes in front of you now and find the…” and then we go into a whole— How are we living in a world where we have this incredible technology and we can’t have a TV that an 86-year-old woman can use?

So this is just to say that I think that even though we have created amazing things doesn’t necessarily mean that we can come up with mechanisms for redistributing wealth, which I would be all for, by the way.

Justin: Okay, then can I just go back to it? Can I go back to the previous conversation? Because actually I had been thinking about this. So think about your poor old mother, right? And let’s say in order to get around this problem, we create a sort of a bot-type thing, right? And she’s also scanned her passport, and she’s also getting a payment from the AI companies, right?

This is the world that we’ve made in future, right? So everything that that bot does now, your poor old mother is responsible for. And she doesn’t know how to get “Foyle’s War” on the TV, right? But Joe next door, he’s a pretty bad dude, and he wants to do some bad things. So he goes and talks to your mother’s robot and gets the robot to do the bad things.

But now it’s your mother’s passport that’s linked to it, and in your future, your mother is then responsible for what the robot’s done, even though she doesn’t know how to get “Foyle’s War” on the TV.

Frank: No, absolutely not because he would have to, it would be biometrically activated, so it would know it was him, and there’d be a blockchain somewhere, and there’d be a record that he had interfered with. This is ludicrous.

Will Amazon bury the Sam Altman movie?

Frank: Anyway, my mum loves “Foyle’s War,” but what I’m really looking forward to watching is the movie we talked about a while back about Sam Altman, “The Firing of Sam Altman,” starring Andrew Garfield as Sam Altman.

And I got a bit worried there recently because I thought maybe we would not get to see it because turns out Amazon dropped it.

Justin: No.

Frank: Yeah, shortly after, total coincidence I’m sure, shortly after they did this big deal with OpenAI and invested something like 50 billion in OpenAI.

Justin: Conflict of interest.

Frank: The head of MGM/Prime in Amazon apparently watched a screening and said, “Another studio might actually be a better fit for this movie.” Apparently… Now, the script, I believe, was never particularly favourable to Sam Altman, but apparently as the shoot went on, it got less and less favourable, and the movie got a little bit darker and a little bit darker.

And Amazon were like, “Ooh, yeah, no, I think somebody else might be a better fit for this.” Loads—

Justin: So where are we going to be able to see it?

Frank: Well, apparently loads of other streamers passed on it as well, and this raised a lot of questions about the integration of big tech with entertainment now, and would we ever get to see it?

But thankfully, Neon, a studio, have actually picked it up and they’ve said, “Yes, we are going to release it this year, and we are going to release it in time for it to be in Oscars consideration,” et cetera. So we will hopefully, fingers crossed, get to see the Luca Guadagnino movie, “Artificial,” about Sam Altman’s firing, starring Andrew Garfield.

Justin: Fantastic. Look forward to that. All right. Well, until next week, Frank.

Frank: Awesome stuff.

Justin: Pleasure as always, and we’ll talk to you soon.

Frank: Chat to you then. Cheers.

Justin: See ya.