In this episode of Futuristic, Cameron and Steve reunite after a three-month break to reflect on how far artificial intelligence and robotics have come since the launch of ChatGPT in November 2022. They chart the wild rise from OpenAI’s first conversational model to today’s trillion-dollar valuations, integrated browsers, agentic coding tools, and the dawn of humanoid robots. Along the way they weigh the “bubble” narrative, the myth of a job apocalypse, and the cultural impact of AI on everything from creativity to capitalism. The conversation pivots from nostalgia for the early internet to speculation about the next three years—when everyone, they predict, will be working with personal AI agents and, perhaps, living alongside household robots. The banter swings between philosophy, tech history, humour, and a few pulled hamstrings.

Futuristic recording – Oct 30, 2025

Cameron: [00:00:00] Welcome back to the Futuristic Podcast, episode 46. It is the 31st of October, 2025. The first time I’ve seen your face on a screen except on TikTok, Steve Sammartino. Since August 4th was the last time you and I did an episode, I did one with my old friend, Nick Johnson, principal of Toowoomba Ankin School on August 23.

But we have not done a podcast for nearly three months, Steve. Not because there’s been nothing happening, just because there’s too much happening, and you’ve been too busy.

Steve: Sounds so needy. I apologize wholeheartedly, Cameron. And, uh, the fact that you didn’t delete my phone from your contacts

Cameron: No

Steve: is revelatory, and I like it. It’s good. You’re a good man. And

Cameron: revelatory.

Steve: it and I missed it. Our chats because [00:01:00] every time I chat to you I get a little bit smarter. So I’ve been in

Cameron: Same, same.

Steve: and I’m glad we are going to, uh, unpack.

Cameron: I’ve got less hair than the last time we talked. Not because it’s falling out, but just because it’s hard to tell with, there’s glare coming through my window. But, um, I am, uh, Sean Light so bright.

Steve: And I just asked, when did you go gray? ’cause you got a really good gray mop there. When did it, were you one of

Cameron: Well, I’m, yeah, I, I started going gray at 23. I’m white. I’ve been white for a lot longer than I’ve been gray, but Yeah.

Steve: Blonde is area part of the Aryan Nation,

Cameron: Or as the Mormons like to say white and delight them. If you wanna get into heaven, you need to be white and delight them.

Steve: Do you?

Cameron: Hmm.

Steve: me. All right.

Cameron: Steve? Um. In terms of what to talk about today. There’s been a lot of news in the last [00:02:00] three months and too much to catch up on, obviously, and I thought singers were coming up to the third anniversary of chat, GPT, which hit the world in November of 2022.

It might be a good time to stop and reflect where have we come from, where are we today, and where do you think we might be three years from now? What do you think about that as a model, Steve?

Steve: Perfect. It is the perfect time for a review. Three years. Things work in threes. It is the perfect time to review it because it’s been a pretty radical three years and maybe in some ways we’re in the [00:03:00] trough of disillusionment now. A lot of people are starting to ask questions. The bubble word comes up, which comes up with every technology revolution. And bubbles aren’t bad anyway. If a bubble bursts, the beneficiaries are usually the people because there’s been an excessive investment in capital and that can benefit all of us because the infrastructure gets built out and they overinvested and we need that. It happened with broadband cables and early internet and, and now it’s, it’s happening again potentially.

And that’s good ’cause you want overinvestment because the beneficiaries are usually the wider populace when it comes to and technology Revolutions.

Cameron: Yes. Look, there’s obviously a ton of investment,

Steve: Mm.

Cameron: ton of hype going into all things AI and robotics we’re to, [00:04:00] NVIDIA’s just became a $5 trillion company open. AI is suggesting they might IPO at a trillion dollar valuation. I mean, it is, uh, like bonkers stuff, absolutely bonkers. But I thought I would start this by pulling up the blog post that OpenAI put out in November, 2002.

It was November 30th, so we’re about a month away from the five year anniversary. It was pretty simple. It said, we’ve trained a model called Chat, GPT, which interacts in a conversational way. The dialogue format makes it possible for chat GPT to answer follow-up questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.

Chat GPT is a sibling model to instruct GPT, which is trained to follow an instruction in a prompt and provide a detailed response. We are excited [00:05:00] to introduce chat GPT to get users feedback and learn about its strengths and weaknesses during the research preview usage of chat. GPT is free. Try it [email protected].

Then there were some samples of ways to use it. Limitations Chachi PT sometimes write plausible sounding, but incorrect or nonsensical answers. Well, I’m glad they fixed that.

Uh, but. These were simpler times in November, 2022. Do you remember when you first heard about it and when you first used it?

Steve: I was actually using GPT, uh, the, the model before, it might’ve been a 2.5 or a three, uh, through one of the APIs. ’cause the APIs were released earlier. wasn’t Grammarly. I’m trying to remember the name of it now. And it was, and it blew my mind the first time I saw it. Uh, when I went on to Chacha, bt I do remember it [00:06:00] being a super revolution, and we can remind listeners that they had a hundred million users in the first month, fastest adopted consumer product in history. also they’ve got 983 million monthly users now, which is pretty radical. You know, maybe one quarter of the internet. And a,

Cameron: Wow.

Steve: of people don’t have, call it unlimited data like we have in western markets. So you’d pretty much say if you’re in a Western market, you’re on it. Uh, it’s decimated search. I, I did say a statistic, which blew my mind on search. Now, if you’re a first page search on Google, first page search item, traffic is down. Reportedly just came out, uh, yesterday, 79%. And that’s two reasons. The first one is the AI dropped down summaries that happens in Google, which to their credit, they’ve adopted the reality of where they are.

Unlike Kodak and said, look, people are going this way. We just have to adopt it and work out the business model later. [00:07:00] But also, I know that I use chat GBT 80% of the time when I once would use search. And I simply ask it for the live feeds of, tell me where you got it from, what happened today. And you know, those small prompts to make sure the data you’re getting isn’t just a hi historical construct. And it, it really has changed the way we use the internet fundamentally. And even though the last, I think the last maybe. Three or six months. A lot’s been happening, but nothing fundamental. I would say easy to forget how far we’ve come in those three years because the first version of chat, GBT when it was launched in, in November of 2022, remember that was an 18 month old database, and then it was six months old, and then now it’s live. It didn’t have code bases, didn’t have Dali, didn’t have image recognition, didn’t have video. It’s easy to forget that [00:08:00] it isn’t just chat GBT. It’s actually been really dramatic and we’ve almost been spoiled by the fact that every iteration, most of them, for 80% of that time in the past three years has been incredible and wowing.

And maybe in the last few months, because we haven’t been wowed with the most recent iteration of Cha JBT, then all of a sudden, oh, bubble Bubble came out.

Cameron: I’ve got an article here from the 5th of December, 2022 New York Times by Kevin Ros. He says, like most nerds who read science fiction, I’ve spent a lot of time wondering how society will greet true artificial intelligence. If and when it arrives, will we panic start sucking up to our new robot overlords, ignore it, and go about our daily lives?

So it’s been fascinating to watch the Twitter sphere try to make sense of chat, GPTA new cutting edge AI chat bot that was open for testing this week. Chat GPT is quite simply the best artificial intelligence chat bot ever released to the [00:09:00] general public. It was built by open ai, the San Francisco AI company that is also responsible for tools like GPT-3 and Dali two, the breakthrough image generator that came out this year.

Um, goes on to say that, uh, AI chatbots are usually terrible, but chat GPT feels different, smarter, weirder, more flexible. It can write jokes, some of which are actually funny, working computer code and college level essays. It can also guess at medical diagnosis, create text-based Harry Potter games and explain scientific concepts at multiple levels of difficulty.

And, uh, you know, I always like to go back and read news articles from the early days of the internet, 93, 94, 95, and see how people were talking about it. Um, he finishes this article though by saying. Personally, I’m still trying to wrap my head around the fact that chat, GPTA chat bot that some people think could make Google obsolete and that has already been compared to the iPhone in [00:10:00] terms of its potential impact on society isn’t even open AI’s best AI model.

That would be GPT for the next incarnation of the company’s large language model, which is rumored to be coming out sometime next year. We are not ready. Now what they did come out with a week ago or so was chat, GPT Atlas, their chromium web browser. Are you using Atlas?

Steve: No, I haven’t. I did have a look at it, but I haven’t gone to using it. Are you?

Cameron: Yes. I’ve been using it exclusively as my browser since it came out and basically, you know, if you type something in. You open it up and you type something in the, the, the URL field, which would normally pull up a Google search. It pulls up a chat, GPT answer. So chat GPT is built in. It’s the default answering model.

You can also, you know, use it on a [00:11:00] sidebar and access all of your chats to, with chat, GPT. Uh, anything you plug into the browser also turns up in your chat GPT history if you interface via the app or your phone or whatever. So it’s totally integrated into the backend.

Steve: because that’s the last thing I want, Cameron. I don’t even want to talk about it.

Cameron: Uh, but yeah, look, it’s, it’s. I, you know, I, I’ve seen a lot of negative feedback about it, uh, in the forums, but, um, I think it’s great. Like it is the default thing I go to now. It’ll give me, you know, um, if I just type in a general search, it’ll give me Wikipedia links, new source links, uh, whatever I want. Um, it’s a search engine that understands me.

It’ll, quite often if I ask a general question, it’ll say, it’ll gimme an answer. Then it’ll say, if you’re thinking about this in terms of doing a podcast episode for one of your history podcasts or your investing podcast, here are some of the things you might wanna think about. So it’s search answers that are [00:12:00] already built, personalized, built around what it knows about me, what I’m trying to do.

Um, yeah, so. That’s a big thing. Like we’re, I, I don’t go to Google. I do literally do not go to Google anymore. ’cause this is now my browser is just chat, GPT, uh, integrated, you know, uh, uh, native integration into chat GPT. Anyway, I’m getting off, I’m getting off, uh, the track. So, November 22 came out and my recollection is that I was shocked and impressed that I could have a conversation with a chatbot that could converse in English.

It could understand what I was talking about. It could understand a question, it could answer my [00:13:00] questions. It could write coherent English responses, sponsors, uh. About anything. They weren’t always correct. They still aren’t always correct, but it was like just a massive shift in language. User interfacing with a computer, uh, blew my mind at the time.

Yeah.

Steve: That’s, that’s the key element, Cameron. The big shift when chat EPT came out was it was the first successful consumer oriented natural language processing computation device where you could or talk to it and get a result that was coherent and not only coherent, a PhD in every single subject. And just like normal human PhDs, they say stupid things too occasionally. So feel like it, it really was when the abstraction of computation got removed and we could [00:14:00] interface in a human way with computational devices. And that was the revolution there. Even the word chat, because chatbots had been around or felt like a bit of a misnomer, almost was named poorly. And didn’t really give the gravitas to what it was because my opinion, and I know yours because we’ve discussed it over the time, is we already have artificial general intelligence. To me, that was the launch of a GI. It is general, it is better than most people at most things, the Cameron Riley definition of artificial general intelligence.

And we’ve had it for three years. And that’s quite telling too, because we’ve had artificial general intelligence for three years and where exactly is this job apocalypse that everyone keeps talking about? It ain’t here. And by the way, it ain’t coming anytime soon or ever. And I know that and I’m even more steadfast on that three years down the track that, uh, not about to replace everyone.

Certainly people with [00:15:00] jobs that have variety in them.

Cameron: Well, let’s talk then about where we are late 2025. Um, you say there’s no jog, job apocalypse, but big tech companies, Amazon just announced they’re laying off, I think 40,000, was it 30,000 or 40,000 people? And it’s not because they’re replacing them all with an ai, but. It is, as I understand it, because they need to cut costs so they can sink more money into building data centers to run ai, and they have to find the money from somewhere.

And the first place to go is these 40,000 employees. So it’s a side effect of AI not directly being replaced by ai, but Microsoft’s been letting thousands of people go. I think some of them are being replaced by ai. [00:16:00] Um, it, we, but, you know, we’re a long way from seeing the apocalypse, and I think the reason is the AI is still not good enough yet.

I, I had a, a, a friend come over for a coffee yesterday and, uh, last time he was over a week or two ago, he’s a, he’s a web designer slash user interface designer guy. Bit of a coder. Um, I said, are you using Cursor? And he said, no, uh, I’m not, I’m not using, I said, it’s your company works in the energy sector, building solutions for the energy sector.

I said, how much AI are people using coding in the background? He goes, uh, not much. I said, check out Cursor. Now, I I, I dunno if you and I, we probably haven’t talked about this ’cause it’s been months, but for people who don’t use Cursor aren’t coders, cursor is a, a development developer tool, an IDE as we call it an interactive developer environment development environment.

It’s basically where [00:17:00] you write code and it’s had AI integration now for a year, 18 months. But one of the things that they implemented a couple of months ago is basically ag agentic development processes. So now I will, and I use it all day, every day, and it’s not. Full time all day, every day. But it’s working on projects behind the scenes.

So I will be trying to write a piece of code for running my business, right? I’ll say, here’s what I wanna do. Um, help me build this. And it’ll start writing code. And then it will test the code using linter. It’s like a testing environment and it has full access to my hard drive, my project folders. And it will write the code, build the code, test the code, the code will fail.

It’ll go, oh, that didn’t work. Lemme try again. That didn’t work. Lemme try again. Oh, I see what I did there. And it’ll [00:18:00] run in the background for 10, 15, 20 minutes trying to get a working version of it. It’ll finally say, yeah, I think this works. I’ll test it, it won’t work. I’ll say, have a look at the console.

And then it’ll read the console. Oh, okay. I see what the problem is. And it goes through another cycle of debug test, debug test, debug test until it thinks it’s got a working model. And I’ll rinse and repeat, and this will sometimes go on for hours or days, and at the end of every day, I will export my conversation with it to my project folder.

And then the next day, if I start a new session, I’ll say, we’ve been trying to solve this problem. Go and read the conversation that we had yesterday. It’s in my folder, it’s a markdown file, and let’s pick up from where we left off yesterday. The other interesting thing that it does is, uh, when the context window, if you’ve been having a chat with it for three or four hours and you’re [00:19:00] still going around it, the context window, you know, it gets so large that it can’t remember what you talked about hours ago, and then it starts to repeat itself.

Steve: it does. Yeah.

Cameron: What, what cursor automatically does is when the context window is filling up it’s sum, it stops, summarizes the conversation, creates a new chat session, uses that summary as the prompt for the new chat session and kicks it off again. So it, that’s all, it just automatically reha, uh, what cleans out its context window and starts again.

So it’s amazing. But he, so this friend came yesterday. I tried using Cursor and it, it got like 90% of the way to building a solution for me and then kept failing on the last 10%. And I just got in this loop where I couldn’t get it to fix the last 10%. And I was like, yeah, yeah, I know that feeling. And I think that is the problem is it’s still [00:20:00] almost useful in that sort of closed loop way, but it’s still not.

It’s, there are still hallucinations. It still makes mistakes. It still gets stuck in these loops, so we’re not there yet. It, it can’t replace a human. It can replace maybe some like low level developers, but it can’t replace a good developer yet.

Steve: So the Amazon, going way back to the jobs and the replacement, I mean, that’s the interesting part is yeah, there’s a huge investment and you need to offset the cost of that investment by removing employees. And it’s not overtly dissimilar to getting robots into a car factory. Funding that investment and having some short term pain where you get people within the administrative or marketing or other commercial realms to offset that, that that has happened quite frequently.

It would be good to see a study on how that has happened [00:21:00] historically, but I don’t think there’s an increase in job losses directly related to the technology itself. And I would be surprised if this automation exceeds the impact that robotic and manufacturing automation has had. I re I really would, and I, and I, I actually don’t think it’s gonna happen and I actually don’t think that even if it can finish a project and that last 10% that you’re speaking about can happen, I actually don’t think it’s gonna replace us because. Agency is the thing that humans want, we change our mind frequently. Uh, a friend of mine, Nick Hodges, he talks about it and I wrote about it last week. He calls it the takeoff and landing problem, is that we want to set the course for the plane. Watch where it’s going, and you just said, then you, you are using the console or the control panel to [00:22:00] help it on where it’s going and guide the AI that’s flying the plane.

But you are sort of guiding it and then you bring it into land on what you finally want. I think that’s where we’re gonna go. And I, I would actually that where we are right now with AI is that it hasn’t reduced the workload for anyone at all. Almost what it has done is enabled us to get more done more quickly.

I know that I get things done a lot quicker now, whether it’s thinking of a blog post, working on a presentation by working closely with the AI as an agent, as a. An idea generating tool that as something to synthesize my thoughts, to gather information and data on my behalf, that might take me five hours.

That takes me five minutes. I’m not doing less work. And in fact, I’ve had my biggest year ever this year financially, and I think it’s because I’m getting more done and being able to satisfy clients’ needs far more quickly and able to do more keynotes and write more things [00:23:00] because I’ve got this incredibly powerful general AI at my disposal.

But everyone I

Cameron: An accelerator.

Steve: yeah, an accelerator and everyone I

Cameron: Hmm.

Steve: the informational realm, whether they’re tradies or people working in commercial roles or entrepreneurs, they all just say that they’re all getting more done. No one’s doing less work. I, I don’t think, and productivity’s been the big missing link that we haven’t had.

I think this is it, but I, and even if the AI can do everything for us, I don’t think we’ll want it to. We actually want to make decisions. We want to have the final call. That’s why you have board meetings and someone comes in and the big cheese makes the decision. People don’t want to outsource their agency.

In fact, the entire capitalist premise money itself is about increasing agency to make independent decisions on your behalf. That’s the actual thing that we want, right?

Cameron: Yeah, but what we want isn’t necessarily what the same thing as what our bosses want or our shareholders want. You know, I, you know, if I, I [00:24:00] still think we’re gonna see low level jobs going. I, I’ve already, like, I wonder about the impact. It’s.

Steve: them go, but not more than we have. Like, it’s gonna be the same pattern that we’ve, I I don’t think it’s gonna like be this massive overnight deluge. Oh, Agis arrived. That’s it. Agen can start, finish envisage the project. Guess what the project is before you even start it. And it’s already done it.

And just, I just don’t think that’s gonna happen.

Cameron: I wonder what the impact is on like gig economy jobs. You know, I, two years ago I had somebody who was editing my shows for me. I had somebody who was transcribing my shows for me. Now I do both of those things myself with AI tools, um, editing a show used to take a.

Steve: those singular things. And this is the point that I think a lot of those pieces of the puzzle, you have, this guy does the edit, this guy does the transcribe. Well, [00:25:00] now you’ve gotta, I can do those bits and, and, and it is incumbent, it is incumbent upon those people who have singular jobs, jobs without a wide scope.

Like we need to be as horizontal as possible what we do so that we have all of the pieces of the puzzle so that we don’t get replaced. The more singular your role, the greater risk that you’re at, like translators, the classic one. Oh, like, must be a tough time to be a translator right now.

Cameron: Well, I, I do think the big question is when are we gonna have ais that are reliable and can be trusted to. Three nines, four nines, five nines. Never gonna be a hundred percent, but trust enough or have AI working to check each other’s work in a system where you can give it, uh, initially a relatively low level job and be [00:26:00] confident that it will deliver a, an output that it’s at human quality level or better.

Steve: Yeah, I think we will have that, and maybe we do have that now, but where the pilot’s flying the plane, right? And we are telling it, it needs to go off and do this. We’re managing agents in the same way that we manage employees, or we are managing ai, doing processes that we would’ve done ourselves or low level employees would’ve done. Some of those low level employees get replaced in the same way that, uh, mail room people, it’s a terrible example from 1989. They get replaced now that we move to email. And so low level functional roles get replaced, but I don’t think complex roles that have a multitude of inputs, outputs, and stakeholders and micro projects get replaced. people who are tiny, thin project tier get replaced.

Cameron: But if you replace all of the low level people, what’s the job of the person who’s currently managing all of the low [00:27:00] level people like corporations are

Steve: the

Cameron: hierarchies, right?

Steve: Yeah. They are, they are. The,

Cameron: I.

Steve: the job of the person who is managing low level people is someone who manages a bunch of ais and agents and actually becomes a really important role because those agents are delivering work that then goes up the hierarchy and you become a, a, an, an orchestrator of the task that AI bots and agents do on our behalf. In

Cameron: So you’re no,

Steve: the

Cameron: you’re no longer managing humans. You’re managing ais.

Steve: bots. Yeah, that’s right. but I do think that some of that hierarchy will remain because people like to be in charge of people, and I still believe that 90% of jobs are bullshit jobs that no one really needs. Anyway, all those Amazon jobs, those 30,000 Amazon jobs, they’re all just like, oh, what are you on Billy Blog’s podcast, promotional operator, strategic, uh, special projects guy on 400 grand a year and whatever, and bye-bye.

Thanks for coming. Like, that’s, that’s what happened. For sure. [00:28:00] Like they’re real bullshit jobs. And the great reminder is what happened during COVID, we really find out. We found out the jobs we really needed during COVID, didn’t we? Well, we need the person who can clean the streets and the medical people and the food, and we really found out what we really needed.

And all of the bullshit jobs, those 30,000 Amazon, that was everyone working from home and doing nothing and getting paid more, which is insane, but that’s the world.

Cameron: And we increased the salaries of those people doing the really important jobs a hundred times because we realized how important they’re,

Steve: So if I can propose a question to you, Cameron, and I

Cameron: Hmm.

Steve: am writing a post on this right now about ownership. So let’s imagine that you get a robot and this robot. Can do everything that you can do and can do your job on your behalf. Let’s say, uh, the robot could be an AI or it could be a humanoid Cameron that looks and acts like Cam with short gray hair that he got at the age of 23, [00:29:00] but this robot is in a 23-year-old.

It’s a new robot. It’s just arrived and it can do all the things that you can do on your behalf or the things that you don’t want to do. super efficient. If you could get that robot and it could do all of your tasks or even if you’re an employee, and let’s imagine you have this weird sense of, uh, autonomy where you can deploy whatever resources you want in your job.

You get to make that choice. You could bring in a bot to do things for you a little bit like how you go on the internet and find things to do your job. If you could own that bot. course you would. You would say yes to it every day because you could be more efficient, get more done, maybe free up your time, do other things maybe not work as much, right?

If you could get that bot, the problem isn’t the robot or the ai, the problem is who owns it, right? Becomes an ownership issue. If the company deploys the robot and owns the robot, it gets rid of you. But if you own the robot and deploy the robot, that’s a totally different set of [00:30:00] circumstances. I know it’s a weird kind of antithetical example, and that’s not how commerce works, because the boss and the management and the owners of capital get to decide where resources get deployed.

But if you owned those, you’d have to say The AI and robotics revolution isn’t bad. It’s a question of who owns it, who deploys it, and where do the benefits get distributed? See, now I’m sounding like a communist. Now you’ve finally got me. I’ve had three months of thinking without Cameron, and I’ve come back a communist.

Cameron: Well, for a start, I wouldn’t wanna replace myself with a robot because I like what I do. I don’t want to outsource what I do. I like what I do.

Steve: Well, no, but wait a minute. Wait a minute. Yes. You like what you do, but you were just telling me five minutes ago about the bots that you were deploying to do certain parts of your job. So thi this is the point. Some people would wanna replace themselves entirely with a robot that becomes their economic engine. Let’s say instead

Cameron: Hmm.

Steve: a company, maybe you own a robot that becomes your personal [00:31:00] economic engine that

Cameron: Slave.

Steve: and it Well,

Cameron: Hmm.

Steve: wow. Here we go.

Cameron: We’re just, you’re talking about slavery basically. Yeah.

Steve: Well,

Cameron: a, an army of slaves out there working for me.

Steve: well, Kurt Cobain sang about it. He said, and it’s okay fish, because they don’t have any feelings. It’s okay you fish ’cause they don’t have any feelings. Something in the way the robots don’t think like that’s where we are. So,

Cameron: I.

Steve: I think we are gonna get to a question of ownership and deployment. You can have a robotic economic engine

Cameron: Robots. Robots is the next section of the show. Steve, I’m trying to stick to AI and then we’ll get to robots in a minute. Yeah, you’re getting ahead.

Steve: Alright.

Cameron: My question is, where we’re at with AI today? Is this where you thought we’d be [00:32:00] three years ago? Are we further ahead or are we not as far ahead as you thought we’d be?

Steve: We are further ahead than I thought we’d be. Way further ahead because humans have a terrible habit. Once something arrives and they’re used to it, they just have full disappointment and they just want more. We are greedy little, little beings. It’s like, seriously, we are so much further ahead. Chat GPT, the first, what was it, 3.5 that came out in blue.

Mines like that was crazy. Good at everything. And then now you can have live realtime conversations. It’s got live web, you can create video, you can get it to write code, like incredible. I think we’re further ahead and I think maybe in the past few months, if we imagine technology doesn’t move in a straight line, it moves in step changes.

You have a bit of a flat moment and then a boom and then a flat moment. We had quite a few booms and then a little bit of flatness in the last iterations were maybe a [00:33:00] teeny bit disappointing because our expectations are so incredibly high now. I think we’re way ahead.

Cameron: At the end of last year, the end of our 2024 season, we predicted 2025. We both said the same thing, the year of agents, because that’s what all of the state-of-the-art model companies were talking about. You and I were talking, uh, the other day and you said agents very disappointing.

Steve: Yeah, they are. So, we are further ahead than I thought we’d be three years ago. We are not as far ahead as I thought we’d be almost 12 months ago. So we, we, we, we did so well in those two years, it was so ridiculously incredible, or two and a half years, even. The first half of this year was pretty incredible that my expectations was just so high, and now I’m a little bit disappointed that agents haven’t arrived, but it might be, it might be in the next six months.

That moment happens where everyone goes, whoa, [00:34:00] remember agents couldn’t take off and land a plane. Now they can.

Cameron: I think agents have arrived,

Steve: They

Cameron: but

Steve: and they’re a bit disappointing.

Cameron: yes, it, it’s, it’s very early days. Yeah. Yeah. They’re not very,

Steve: days. We haven’t

Cameron: yeah.

Steve: jet engine phase yet. We’re still smashing those propellers and the dual wing.

I mean, we’re not there yet.

Cameron: I mean, my cursor example before is a limited example, but to me it’s mind blowing compared to where coding was with AI a year ago, where I had to. Tell it what I wanted. Copy and paste the code, put it in an IDE, run it, copy and paste the error message, put it back, find a document, share the document. Now I just give it access to my directory on my hard drive and say, go do it.

If you need something, look it up. And it does. And it finds it. And it fixes it. Still perfect. No, but in terms of where we were at the end of last year, mind blowing. [00:35:00] But it’s a limited. Example of, of agentic behavior from a coding perspective. Uh, by the way, there, there was something I would’ve talked to you about a couple of months ago if we’d done a show.

Um, there was a lot of criticism I was seeing online because Dario e Modi, the CEO of Anthropic had said six months earlier that within six months, 90% of code would be written by ai. And people will say, well, that’s not true. You know, I’m still coding the old fashioned way at blah, blah, blah. My argument is, but there’s 200 million people writing code now that weren’t writing code six months ago.

Steve: So it is. It’s maybe even more than 90%. You are right. it definitely is. And I was just. Yeah, looking at the, the idea of it coding, one thing that it has done recently is AI can release its tentacles into the web in the market and bring things [00:36:00] back now. So in a way that’s kind of agentic within the scope of the project that you’re asking it to do, rather than, here’s a project, go and do it.

But on the tools that it’s been trained on, like coding, it does very well go out and go, oh, okay. I’ve gone onto to hugging face and I’ve done this and I’ve gone to these and it brings back So it’s tentacles are starting to reach out.

Cameron: We’ve also seen this year the rise of. Text to video and text to music models, which are really accelerating very, very quickly. VO 3.5 just came out where you can do much longer videos. Saw a two came out a few weeks ago. I still don’t have access to it even though I have an invite code from my son in LA because you have to be in the US to be able to download it Still.

But the videos that are coming out are really, really super impressive, but interesting. I was in the car with Fox the other day and he said, you know what? I’m [00:37:00] sick of saw videos. He goes, everywhere I look, it’s saw a video, saw a video, saw a video. And he said, I’m actually quite concerned about the future where I’m not gonna be able to tell what’s real and what’s not real anymore.

And he said, if I had a time machine, he said, he asked me, he said, if you had a time machine, would you go into the future? Would you go into the past? And I said, well, hold, can I get back after I do it? He goes, yeah. And I go, my answer is usually the past. And when people say, where would you go? I say, I’d go back 30 years ago to talk to my dad and my grandparents when they were still alive again, because there are so many things that I wish I could tell them about.

And talk to them about that. I didn’t have the opportunity or, you know, they, they didn’t get to meet any of my children and you know, they didn’t get to meet my wife and all that kinda stuff. But I’d, I’d show them my abs and say, look at me. [00:38:00] I’m 55 and you know, I’m, I’m do kung fu Um,

Steve: though? Don’t just wait a minute. Stop, pause everyone. Holy mackerel. Riley’s in good shape. And now the podcast is X-rated. All of a sudden,

Cameron: dude, I have this incredible like kung fu body now after, and because of chat GPT,

Steve: you say so yourself, I

Cameron: well, compared to where I was a couple of years ago. Yeah, I would say incredible subjectively, not objectively, subjectively,

Steve: Yeah.

Cameron: um. But he was saying, and and I said, well, but then I’d also go into the future and see what stocks are successful and then I’d come back and invest in them, you know?

Steve: style.

Cameron: But he said, yeah, yeah, yeah. But he said I’d get back into the past. ’cause I wanna know what life was like before everything was AI and you know, computerized what life was like before iPads and iPhones and the internet and AI.

Steve: freedom and, and fluidity to it, [00:39:00] didn’t it? It really did have a different sensibility, and every era has upsides and downsides, right?

Cameron: My point is, you know, I keep wondering how his generation is going to respond to this world of AI content and the dead internet theory and all that kinda stuff. I kind of assume they will just flow into it, not question it, not think about it, and it’ll just be, well, this is what we’ve always known. But the fact that he’s 11 and is going, you know what, I, I don’t like this, this is really kind of depressing and scary.

Um, I dunno if that’s reflected across his generation at all yet, but it was kind of surprising to me that that was his current point of view.

Steve: He’s a very wise, young fellow. How old’s Fox now?

Cameron: He’s 11 and a half.

Steve: Yeah, so

Cameron: Hmm.

Steve: smart parents. I, I’m not surprised. of things. The first thing [00:40:00] is if he wonders what life’s like pre-internet, can actually still do that in the real world. It’s not that it’s not possible, it’s just that we don’t, so you can go somewhere without a smartphone and only read a newspaper and maybe just watch the TV channels that are on the TV and be in nature and be at the beach and not have your phone and all of that is still possible.

We choose not to. Right. So that’s an interesting, just side note that anyone who wishes for the past can reinvent it. By changing places and

Cameron: Log off.

Steve: about what Yeah. And what things you bring with you. Right. But I wanna pick up on the VEO and the video to text, which is another amazing thing with ai.

So yeah, again, if, the listeners haven’t got it yet, we are so far ahead of where I thought it would be three years ago. Incredibly so. And that’s another great example, but one of the terms that you introduced me to way back in the early Twitter days, one of [00:41:00] probably 20 years ago when we first met, was the Splinter net. And I feel like the Splinter net is making a comeback because what we have now is a couple of apps, are AI only apps. The Sawa app and the Meta app are both AI only apps. And in June of this year, and I’m not talking about. Malicious bots, the amount of content on the internet crossed over to be a majority of ai.

It’s now 50, it was 51% in, I think it was August. was a lot of articles that came out with the internet officially crossed over to be more non-human than human. And I think we might see apps which are human only, and a couple of people have proposed it. What? What is the new social network? How do you prove that you typed it with your finger in that moment, and the photo has to be live, or it has to be [00:42:00] only human, and there’s technological ways you can do it.

It creates. Boundaries and barriers to e entry. And that’s okay because that’s what it would be for. You would log in and go, well, this is a human with a human opinion. There are no bots on here. You can only be a verified human in some capacity because we’ve already got the AI versions of those now. And to be quite honest, after five minutes, they’re pretty damn boring. I’ve uploaded a heap of photos into the meta app and created videos. I’m like, yeah, okay. I, it’s, it’s a little bit like, I get the joke because there’s something wonderful about dogs hang out with dogs, birds hang out with birds, fish hang out with fish, and humans hang out with humans. And

Cameron: Fox was,

Steve: we’ve got this other species, the AI species, and we’ve talked about spawning another species,

Cameron: Hmm.

Steve: hang with humans.

Cameron: Fox was talking about Nightmare, but now nightmare before Christmas at bedtime last night. And he was like, we’re never gonna see that again. Like handcrafted [00:43:00] animation or stop motion animation. He said, there’ll be stuff that’ll look like it’s handcrafted stop motion animation, but ha, but it won’t be.

And I said, well, maybe people will go back to doing that, uh, because they want human generated content. He goes, but how will you know? And I said, well, they’ll be behind the scenes stuff of the animators making it. And he go, yeah, but how do you know that’s not AI generated?

Steve: call.

Cameron: like, yeah, it will be ha.

Steve: we’ve shown a behind the scenes to show how humans generat it, but we’ve generated it with AI to make you believe and feel, because we, we here at Animators Incorporated know that the feeling of the belief that it really was human when it wasn’t, is the feeling that you are really after.

Cameron: Okay. Uh, before we run outta time, ’cause I still wanna talk about robots next three years, Steve. Um, so you’ve talked about, we’ve come a long way. You’re a little bit disenchanted with the lack of agentic process, uh, progress this year. [00:44:00] Based on what progress you’ve seen happen over the last three years, where do you think we will be sitting three years from now in the end of 2028?

Steve: Of 2028, this is my prediction. All of us will be working with a number of agents. I think they’ll solve the agentic problem. I don’t think agents will be just let loose and just ping you in the morning and say, Hey, you know how you had a few conversations with Cameron and then Billy and Mary and you’ve been thinking about doing this startup over here?

Well, I’ve just launched it all overnight and here’s what I’ve done. Hope you like it. I don’t think that’s ever gonna happen. we’re still gonna set the course of the AI agent, but I think it’ll be able to do everything. All of us will be managing a suite of ais or a singular AI that can do a, a number of projects simultaneously. A little bit like the movie Her, where she’s having, you know, [00:45:00] 11,064 conversations simultaneously. We’ll be managing those. I think leading edge consumers and companies will start to have a lot of robotic humans, humanoid robots walking around in offices, retail spaces, factories, warehouses, and homes

Cameron: Don’t skip ahead. Robots is the next section. Steve, stick with ai. Steve?

Steve: straight ai. straight ai. You mean screen-based? Ai.

Cameron: Yes.

Steve: Okay. I think all of us will be working with a suite of agents, both in corporate and private settings that will be, will be guiding on projects that can do all of the technical things that we can’t do, whether that’s video editing, writing code, managing projects, mathematics, engineering, all of those things.

And I think it’s gonna be a real period of emancipation where your qualifications are almost irrelevant. And the only qualification you need is to have good taste, to have desires, an [00:46:00] entrepreneurial ethic, and know what projects you wanna undertake.

Cameron: Yeah, I, I, I tend to agree. I don’t expect to see a slowing down of progress. I expect to see, uh, ramping up. Uh, I think things are gonna continue to build exponentially on the version that came before. I think AI is gonna play a role in enhancing and developing itself to greater and greater extent AI’s, coding ais and taking the learnings and, and factoring them into the next iterative, uh, step forward.

I do think, um, you know, with the on, with the text to video generation and the content, a lot of it is AI slop, as people call it, or just the same idea repeating itself. And I do, you know, I am waiting for the first [00:47:00] generations of truly innovative and exciting content to emerge out of those, a kind of music that we’ve never heard before, a kind of video content that we’ve never seen before that takes all of these things and really just does something truly exciting and truly innovative and spawns a whole new area of content.

Um, you know, I remember with podcasting when we started at 20 years ago. You know, people would say, well, it’s just radio and you, it’s just radio on a, you know, on a, a,

Steve: Radio,

Cameron: portable device. Right? And that was to a certain extent, true. And I used to say, one day we’ll figure out how to do something that’s truly unique.

And without blowing my own horn, I was one of the people who did that because all of a sudden I did a hundred hours about Napoleonic history. N no [00:48:00] one ever in the history of humanity had talked about Napoleon for a hundred hours in a media format like that.

Steve: Well,

Cameron: You couldn’t do that on,

Steve: it. You couldn’t do it because there you couldn’t.

Cameron: even in cassette tape pack days in the eighties, no one did a hundred cassette tapes.

Hour long. BASF cassette tapes on Napoleon or on Caesar or Alexander even hour long, or now, you know, two, three hour long. Podcasts are quite commonplace now. I oh oh,

 a cramp in my leg. It’s not a heart attack. Just everyone relax.

Steve: attack.

Cameron: Oh,

Steve: relax. Karate man. The black belt himself has abs, but unfortunately his gray hairs and age give him te terrible cramps. It’s not a

Cameron: oh,

Steve: [00:49:00] Everybody relax. He will live through this

Cameron: oh my

Steve: I

Cameron: god.

Steve: that is karma. Coming back to Cameron for showing off his abs,

Cameron: Just gonna

Steve: first on

Cameron: stand up and stretch my

Steve: Cameron could not

Cameron: oh

Steve: hamstrings popping as they have.

Cameron: oh

Steve: Cameron,

Cameron: oh.

Steve: this out. It’s hilarious.

Cameron: Oh my God, that hurt.

Steve: If you’re not watching this on YouTube and you’re listening right now,

Cameron: It was right,

Steve: out his hamstring.

Cameron: right there. Oh, I just tucked my leg under my chair.

Steve: attack. Has

Cameron: I,

Steve: Has anyone died live on a podcast?

Cameron: that’s a good question. I don’t know.

Steve: a podcast? We’re

Cameron: Christ. That fucking hurt. Oh, I need to go have some electrolytes and magnesium. Um, potassium.

Steve: two

Cameron: Um,

Steve: podcast with Napoleon that [00:50:00] you did that wasn’t possible on the, it’s just radio on internet was, first of all, you found a long tail of customers without any promotional costs, where they kind of found you, you know, distributed around the world, which wasn’t possible before

Cameron: yeah.

Steve: had breadth of

Cameron: Yeah.

Steve: in the physical world, whether you’re burning CDs and shipping them around, there are the cassettes.

That wasn’t possible either. And so the content, the topic, the distribution, and the audience, all of those things weren’t possible. And you were a real innovator within that realm where, Classic long tail. One size fits one, you know, the, the world is weird. And now we can all find each other.

Cameron: So anyway, my point just without wanting to suck my own dick, but the point was new things do come outta these things. It takes time sometimes, but new things do come out of a w worthy new things. I mean, again,

Steve: And,

Cameron: says me,

Steve: we haven’t found them yet with the AI

Cameron: you

Steve: the text video. I did see one small iteration that, know, [00:51:00] gave me a three minute smile. And then I got on with my life, which was animations with ironic political bents arriving on TikTok, where

Cameron: right.

Steve: talk about, uh, the Little Mermaid and then like turn it into some political thing.

And it had all of the imagery in Disney and it was pretty smart and cool, but it was, it wasn’t a fundamental in the road where it’s an entirely new thing.

Cameron: So there’s a lot, and I think we’re gonna see, um, wider, wider deployments of AI into everyday life and business operations. More decisions, more workflows. More operations will be partially or fully AI driven rather than just AI assisted as they are today. I do think AI agents will not just generate code or content, but will be able to execute multi-step tasks with some maybe real world effect bookings and purchases is kind of little bit, [00:52:00] sort of boring for me, but real operational decisions.

Um, you know, what I’d love to be able to do today is say to my ai, you know, I want to, I want to. C create a new marketing campaign for my investing podcast in America. Can you go out, find all of the right audiences, find the people that are listening to investing podcasts, that are interested in buffet style value investing.

You know, track down their emails, their socials, uh, what they’re listening to, go and create ads for other podcasts. Execute it. Do the deals, negotiate the deals, execute it, just go off and do it like, and have it, you know, go and deliver all of that kinda stuff. I think models are gonna become more efficient.

I think they’re gonna become cheaper to train and deploy. I think we’re gonna see, uh, even more stuff coming out of China. Oh, by the way, um, side note, but great. I dunno if I already [00:53:00] sent you this on a text, maybe not. There’s a new book I just, uh, heard about by a Chinese American China scholar called Dan Wong.

It’s called Breakneck, and it’s talking about China and where it’s going. And one of the greatest insights that explains China versus the US and the rest of the world was this. He said, China is run by engineers. America is run by lawyers.

Steve: you sent me that quote,

Cameron: I sent you that quote.

Steve: Yeah. Which, which, yeah,

Cameron: I listened to a podcast interview with him. I think it was a New York Times interview, and he, he said, I lived in China for, I think it was like 2017 to 2023, and he said. Not only are Chinese cities better than American cities, but you go out into regional China now, and the regional cities are better than big American cities.

They’ve got better roads, better energy, better water, cleaner water, better [00:54:00] wifi, better internet. Everything’s better. Because he said, you know, dong Xiaoping made a conscious decision to elevate engineers into the Communist Party, into the pulp Bureau. And that’s continued. So they have an engineering first mindset.

How do we build the future rather than a, you know, democracy run by lawyers that’s like, well, here’s what we’re gonna allow and here’s what we’re not gonna allow. And you just all figure it out in the middle. They’re like, here’s what we’ve gotta build. Let’s go build it. And he said, and sometimes they get it wrong.

They build the wrong thing or too much of something or whatever. But it’s that engineering mindset and which is why they’re now doing more patents than anywhere else in the world. Um, and he also said competition is even more cutthroat in China than it is anywhere else, because there’s more of them.

Steve: to America’s one of the most uncompetitive countries in the world. It, it

Cameron: He,

Steve: is

Cameron: he said, but if you look at it, and he’s talking about what happened in [00:55:00] the tech sector over there, he said, like he said, he interviewed one company over there that started off as a Groupon clone and survived, but they said they were one of 5,000 Groupon clones operating in China at one time. But he said the, the thing is, because of the way it works, the people and the state get all of the benefits.

The entrepreneurs themselves fight it out for scraps because there’s so much competition.

Steve: that’s what capitalism should be. It should be that corporations fight and that there’s no barriers to entry and we can continue to have that fight so that the consumers get the best outcome. That’s the idea of capitalism, which lurches the entire economy forward. But you don’t have capitalism in America anymore.

I mean, again, there’s no such thing as pure capitalism, pure communism, any of that. We’ve

Cameron: Hmm.

Steve: that neither of those exist, but America has became become far less capitalist, as has Australia in in the last couple of decades. Incredibly so. And I think a big part of that is [00:56:00] because software has copyright, uh, laws baked into it, which reduces competition potential.

It’s one of the major issues. And then you’ve got the aging population in these economies as well, where. To get into power, you need to protect the power structures that exist. within China, the idea that they’re doing better outta the advantage is they don’t have legacy infrastructure. Uh, when you’re starting at ground zero, it’s easier to design a new futuristic city than when you’ve got existing roads and systems and vested interest of who owns the capital.

It becomes a far more difficult place to innovate around,

Cameron: Hmm.

Steve: you have protection versus creation.

Cameron: Mm-hmm.

Steve: is in creation mode

Cameron: Hmm.

Steve: in protection mode.

Cameron: Yeah.

Steve: that that’s a core issue. And, and know, you use the word poly bureau, that’s actually the core issue in America now is not technological.

It’s not that they don’t have enough entrepreneurs or enough [00:57:00] clever people or the wherewithal to do it. It’s that we have the wrong political structure. The same in Australia. It’s why no one can afford a house. Because we are protecting existential power structures and it’s getting worse in America too, and certainly with tech and ai

Cameron: So getting my point was just gonna be, I, I expect to see more AI leading edge stuff coming outta China. More chips, more data centers, more models. I, I do think China’s really going to. Become way more important three years from now than it is at the moment. It’s already caught up in many ways, uh, to the us, but I think it’s gonna, we’re gonna continue to see rapid progress coming out of China.

Let’s move on to robots. Steve, we’re an hour in, can we, can we do robots

Steve: We can, we definitely do robots. Real quick,

Cameron: quick?

Steve: last thing that happened with AI was I really liked the idea when [00:58:00] Deep Seat came out and really blew everyone away, they had an open source code base. The fact that it showed, it was the first to show the thinking of the ai, the thinking process,

Cameron: Mm-hmm.

Steve: how quickly

Cameron: Mm-hmm.

Steve: AI turned around and released a model of for

Cameron: Mm-hmm.

Steve: I think China being competitive is good for Western markets because it keeps us on our toes.

Cameron: Now we have lots of really state-of-the-art models coming out of China that are, if not as good, almost as good as the state-of-the-art models coming outta the us. So robots in November, 2022, Steve Humanoid robots were largely curiosities in a lab. Um, high cost demos of walking or balancing or some simple manipulation, but not doing general tasks in human environments.

They were wired up. Tethered, usually lots of external sensors or human teleoperation. The business cakes was weak and they didn’t have ai, so they weren’t very [00:59:00] smart. It was all about, look, it can stand up and balance or stack boxes. Robots and AI were sort of separate tracks three years ago. Today, of course, it’s just a robot sort of delivery mechanism for AI in many ways.

They’re all gonna, we see it now. It’s all gonna be integrated from the get go. And you know, one of the big news stories this week was the neo humanoid robot going on presale $20,000. You can pre-order a neo and a very, very well, not very impressive, a somewhat impressive launch video for it that came out this week of it going around your house, doing all of your chores for you while you’re out of the home.

Wall Street Journal went and met with them and, uh, behind the scenes, 99% of everything that you saw in the launch video was a human tele operating the robot. The robot’s not real. It doesn’t exist.

Steve: be [01:00:00] behind it. There’s

Cameron: Yeah.

Steve: in the humanoid robot dancing around.

Cameron: It’s just, yeah, horrible. But that’s where we’re at there pre-selling this thing that doesn’t even exist yet. Trying to get ahead of the market.

Steve: exist. I love a presale.

Cameron: What about your robot that you ordered? Have you got it yet?

Steve: No, I, it arrives in December, K Bott, is, you have to code it. You can’t verbally and visually train it,

Cameron: Right.

Steve: which is gonna be tricky. Uh, but still, I’m, I’m gonna pre-order one of the Neos too, like why not It’s pre-order. It’s like a few hundred bucks.

It’s definitely worth doing. I love what you just said, Cameron. Humanoid robots are a delivery mechanism for ai. That’s beautiful, man. And that’s exactly it. And I think this is the big, I think this is a bigger shift than screen-based and voice-based ai, because they’re multimodal. Now that AI can talk, have visual, verbal reasoning, [01:01:00] we can explain things, we can do demos.

It becomes the, the, uh, let’s say the embodiment moment of ai. And I think it’s akin to what happened with the car because we had horse and cars for hundreds of years. We’ve had robots not for hundreds of years, but for quite a long time, uh, static robots that can do tasks within factory settings. What we’ve done with ai, I believe, is developed the brain and the nervous system put inside the humanoid robot to get a deployment. And we need humanoid robots because we’ve got a human shaped world. And humanoid robots for me are a chance for grand emancipation of humanity where we can escape the screen and the tech behemoths potentially, because a lot of the humanoid robots aren’t coming from the traditional big tech, which I like and is quite exciting.

And if we have open source l LMS LLMs, we can plunk that into a humanoid robot that we train with [01:02:00] our perspective, with our ideas, a little bit like the way we train our children and show it the tasks that matter in our work context or in our domestic context. That’s the exciting part of humanoid robots.

And so long as they’re open source and or we can tinker with them in the same way that we can tinker with a car. We used to buy a car and you own it. You can lift up the bonnet and soup up the motor and put spoilers on it and change it the way you want. That’s what we need in ai. That’s the missing link. Right. And And I, and I feel like. Humanoids is the big, big thing. I think that, I think the robot economy, and I wrote about that a few weeks ago. I think we’re gonna have this huge robot economy. The nighttime economy is 13% of GDP and it didn’t exist in Australia until 1917 because there was no electric lights. If we can deploy robots to do things on our behalf, they become a new economic engine for everyone. I’m like, I’m really excited about that, but it’s got to be open source where we can play with it, not where the [01:03:00] code base and it’s all controlled from afar and across the cloud and they upload and download. Potentially, we could teach our robots something and then sell license. What we taught our, so you can have the, the nicest trimmed hedges in your garden. ’cause your robot is, is the champion at that or whatever it is you want to teach it. For me,

Cameron: Stop being so capitalist, train it and then give it away. Give the code away. Make it freely available. Open source the code.

Steve: Okay, great. Do that. Well. Well, well, well, either an economic engine or let’s say an abundance engine might

Cameron: A world where your robot sees a black or helicopter and you say, can you fly that thing? And it goes. I can now, yeah. Downloads the

Steve: right. Yeah,

Cameron: right? Like

Steve: but let, let’s

Cameron: Trinity.

Steve: Either economically or, or, or either a fashion, uh, that it opens up abundance because things can be done for you and on your behalf and within this ecosystem that all of [01:04:00] us own and develop. ’cause I think that’s what happened with airlines, with cars, with every technology until we inside, you know?

Cameron: So where we’re at today with humanoid robots, I talked about where we were at three years ago. Where we’re at today is they are available for sale. You’ve bought one primitive model, but you’ve bought a humanoid robot. Do you think three years ago you would’ve expected that you would’ve bought a humanoid robot by late 2025?

Steve: Now this is a massive surprise. And even the price point, we’re talking 16 K cam and the new Neo is 20,000. Uh, I, yeah, 20,000 us. So let’s just say they’re the price of a small car. I think that’s a fair synopsis. And, and

Cameron: And

Steve: anyone, anyone would buy a humanoid robot if it came at the cost of a small car, because I think it’s gonna have maybe even more utility than a car does.

Cameron: I don’t think we will end up buying them. I think we’ll [01:05:00] end up having them on a monthly subscription, but

Steve: Ooh, here.

Cameron: Yeah. But well, not all of us are as rich as you are, Steve. I can’t afford to spend $20,000 on a robot, so if I don’t get it on a subscription, I’m not getting one. But anyway, that’s another story.

Steve: but if you, no, wait a minute. Let’s, let’s hold up there. If you would buy a car for $20,000, I do not see why you wouldn’t buy a humanoid robot, which can undertake all of the domestic tasks. And let’s assume it can do everything. Just like an ai as a PhD in every subject. Why? Why wouldn’t you, in fact,

Cameron: ’cause you don’t have the money. ’cause I don’t have the money, Steve.

Steve: sell the car and get the robot, the robot will take you further and it can

Cameron: It’ll pick me up and carry me to kung fu

Steve: you to the bus stop. It will give you a piggyback

Cameron: of the family too. Yeah. Yeah.

Steve: what I need to get. When my robot comes, I need to hop on its back and go. You think you can’t afford a robot? Well, I sold my car and I’ve got a robot. Where am I going? Now he’s piggybacking me to the train. Stop, bring [01:06:00] on rail.

The electrical transport system.

Cameron: ranting. Where are we? So there are humanoid robots being deployed in factories, uh, in an Amazon and places like that.

Steve: has them doing, um, certain parts of, uh, moving panels within the cars.

Cameron: Yeah, Tesla are saying they’re gonna ship 5,000 units in 2025, but Tesla says a lot of things that they don’t deliver on. Uh, you know, I’ve seen forecasts that the mar, the humanoid robot market is projected to be about us $30 billion by 2035. So that’s 10 years away. But where it’s at today is there’s a lot of improvements.

Like they’re able to walk around without being tethered up without teleoperation, do simple things, better balance, more robust hardware, better finger manipulation. They can pick up objects, navigate human spaces. A little bit more capably, [01:07:00] but a long way from something that’s, you know, science fictiony robots.

You know, still a long way from really being something that you can use to get things done in the home. But now when I say a long way, a year, two, three, um,

Steve: Maybe

Cameron: know, but we’ve,

Steve: things are non-linear. This is non-linear

Cameron: it’s non-linear. That’s right. And a lot of these things turned out to be really hard, like getting the dexterity of them and the balance and all that kind of stuff.

Their, their ability to walk or move quickly. Like most of them, you see, they’re really slow moving. It’s like, I’ve just eaten. Too many gummies and everything’s moving in slow motion.

Steve: Uh.

Cameron: Uh, but we’ve come a long way in the last three years. And the big thing, of course, not just the, the form factor, uh, improvements, but AI in the chip set, AI in the brain, that’s, as you said, is the, the [01:08:00] big leap forward in many ways.

Um, there’s a lot of work that’s being done by NVIDIA and companies like that for building virtual environments to train the robots to do something, and then they can take that code and just stick it in its head and it already knows how to put away the dishes or cook a dinner or whatever it is.

Steve: humanoid robots are gonna benefit from cognitive surplus in the same way that the internet did where we had all of this connection and the knowledge bank went up exponentially. soon as one robot knows how to do one thing, theoretically, if we have the right model and open source nature of it, then all humanoid robots, like you say, download it and it knows how to do it. So the rapid onset of the ability of humanoid robots should be. As, as soon as one knows how to do it all, know how to do it. And, and, and that’s, [01:09:00] that’s where I think you get that exponential improvement on capability. So long as the balance, strength, dexterity within the finger movements then everything changes.

But then you get the second and third order effect where all of a sudden, and I’ve been espousing this for some time now, is that the advantage of low cost labor markets starts to get eroded. And manufacturing and production of many things starts to become possible again in high cost labor markets.

Cameron: Yes, I see that happening too. So the question is, where do you think we’ll be three years from now, late 2028 with robotics?

Steve: I think someone in your street will have a humanoid robot. In every developed market, and it’ll be a curiosity for two, maybe one or two or three years until people see the incredible utility in workplaces and or your home [01:10:00] and ev, and then everyone gravitates towards getting their model T

Cameron: Yeah,

Steve: quickly after that.

Cameron: I think you’re right. I think three years from now I’ll expect to see them out and about In what businesses? Factories in small numbers.

Steve: yeah, you’ll

Cameron: Yeah. Uh, elder care facilities, maybe hotels, places like that. Doing service delivery work.

Steve: which they’re in now. I was at the Gold Coast airport and I had a robot going around cleaning the floor. I’ve been in a hotel in Shanghai that had a robot that goes up the stairs and delivers your food. Again, it’s on wheels, but it’s not far of a shift

Cameron: Hmm.

Steve: that capability into the human eye. Uh. Walking element and once you see it, the fear gets removed. People see the utility and, and you just have to go to it.

Cameron: Well, there’s my local ZA restaurant has a robot that delivers your [01:11:00] food to your table from the kitchen. But yeah, I’m not talking about that. I’m talking about genuine humanoid form factor robots doing generalized tasks. I do expect that you, if they’re already on presale in really early models, three years from now, we should have, I, I still don’t think they’re gonna be Jetsons, but I think we’ll have robots that are able to do dozens of tasks around the house or in the business.

And prices and availability won’t be at a level where they can be mainstream yet, but they will still, they will start to become something you’ll see out and about more often by the end of 2028.

Steve: Yeah, the really good ones, the figure bots are well over a hundred k. Some of the other bionic ones that come from Uni tree, they, it’s sort of anywhere between 30 and 50. So

Cameron: Yeah,

Steve: expensive, but we’ve gotta remember that the price will halfen, the capability will [01:12:00] double. It’s, it’s gonna be, uh, production efficiency in Moore’s law capability element inside that.

I think this is bigger than AI on the screen.

Cameron: who’s the guy who wrote about um, 10,000 hours

Steve: That was Gladwell.

Cameron: Gladwell? I was gonna ask Chachi T but you answered before I could even type in the question. I just wanted to finish by talking about self-driving cars. I saw a great clip of him couple of months ago. I dunno if I sent it to you, if you saw it.

Steve: No you

Cameron: explaining why we’ll never have self-driving cars

Steve: Oh,

Cameron: on mass, he said, because.

They are designed to stop, if anything gets in their way to prevent accidents. And I’ve seen videos since then of people in streets in LA or somewhere like that, just all standing around and sitting on top [01:13:00] of it and it can’t move. It can’t go anywhere because it’s not allowed to run over people. He said the reason people stay off the streets today is because if you step into a street, you’re probably gonna get hit by a car and killed or hit.

If the roads are just full of self-driving cars, no one’s going to stop for a self-driving car. They’ll just walk out in the middle of it and it will just have to stop and no one will get anywhere and nothing will get done.

Steve: Yeah, it’s,

Cameron: his argument.

Steve: and it’s a, it’s an interesting idea,

Cameron: It is, but I, I have my counter argument to his argument.

Steve: which is

Cameron: It is just a legislative thing. All of these cars have cameras and you’ve got, you know, facial recognition in the cameras.

Steve: right.

Cameron: your phone is giving off a chip. If you step in front of a vehicle or [01:14:00] interfere with a self-driving vehicle and you can’t justify why you did that instant,

Steve: it on purpose. Yeah.

Cameron: instant, fine.

You know, it’s a, it’s a one stop legislative. Well, that, that’s on you. Yeah. You, you’re fucking, you’re fucking with society. Immediate. $500 fine. Right.

Steve: Yeah. And also the fact that Waymo is operating incredibly well in every market it’s been in.

Cameron: Well, that’s because if there’s a, if, if 1% of the cars on the street are self-driving autonomous vehicles, people aren’t gonna just go and f fuck with the traffic. Generally speaking, if 95% of them are self-driving vehicles, people might just go, it’s gonna stop for me. I can just, you know, cross the street whenever I want.

But we already have pedestrian crossings. We already have light systems that say when you can and when you can’t cross the street. Yeah.

Steve: would, you would hope that self preservation [01:15:00] and not putting, just taking the piss or putting adverse, you know,

Cameron: But this was,

Steve: trust

Cameron: my point was.

Steve: Yeah. Yeah.

Cameron: was on stage going, and this is why we’ll never have it. And within 10 seconds, I was like, no, fuck you, you idiot. It’s like, it’s one piece of legislation that just kills that dead in its tracks. So yeah. So much for Gladwell as a sociologist or whatever, he markets himself as

Steve: people’s work anyway. You know? He never had

Cameron: Oh, allegedly, allegedly.

I don’t wanna get sued for defamation on this podcast, Steve. Allegedly.

Steve: does

Cameron: Fuck,

Steve: of his

Cameron: allegedly.

Steve: and he just takes everyone else’s research and just, just

Cameron: Well, that’s what

Steve: public.

Cameron: writers do. That’s what I did.

Steve: Yes.

Cameron: Yeah, I, I just said, according to Steve Santino wrote a whole book about it, psychopath epidemic. Look it up.

Steve: That’s

Cameron: Uh uh.

Steve: it.

Cameron: No, I’m saying I have, I’ve stole the idea [01:16:00] from you. I didn’t, but yeah,

Steve: you didn’t,

Cameron: probably stole it from someone. All right, Steve, we should go. That was great fun.

Steve: It was so good.

Cameron: Let’s not live it. Three months. Yeah. Let’s do it again in three months. Three years. Do it in three years and see how we went.

Steve: We might not be here.

Cameron: Alright. I.