The Impending Crisis: AI, Job Losses, and the Future of Work

David Brown: 00:00:00

Which for me is the ultimate irony, considering that AI from its core is built off of other people's IP and it's trained off of everyone else's ip. And yet then they have the temerity to complain that somebody is trying to steal their IP using the same type of tool that they use.

And it's just the hypocrisy of it absolutely. Just drives me wild.

Alan King: 00:00:29

Well, good morning and welcome to AI Evolution Podcast. We have an unusual episode today, actually, because I have Ben and Dave.

Ben Harvey: 00:00:39

Hello.

Alan King: 00:00:40

We're in the same room. Look there. We can actually reach out and talk to you. First time. I know we're not. We're not staring at different screens, but there we are.

So we are literally going to have, as you can see, a Fireside chat this morning, and I think we're going to keep it sort of fairly loose because not often we get together like this. So this is a good opportunity just to kind of have a more of a freeform approach to the podcast.

So we haven't got a specific topic this week, but we are going to talk about a range of different things that we've seen that have been going on in the AI space over the past few weeks and have a bit of a. Bit of a Fireside Chat about those.

David Brown: 00:01:15

Sounds good.

Alan King: 00:01:15

Dave, do you want to kick off this time? I think you've got something you want to talk about.

David Brown: 00:01:19

Yeah, sure.

So in the last couple of weeks, I've noticed some articles coming out where I think it was anthropic, was complaining bitterly that other AI companies have been using their AI tools to essentially attack their system, to learn how all of their algorithms work so that they can then take that IP and to make the other AIs work better.

Which for me is the only ultimate irony, considering that AI from its core is built off of other people's IP and it's trained off of everyone else's ip. And yet then they have the temerity to complain that somebody is trying to steal their IP using the same type of tool that they use.

And it's just the. The hypocrisy of it absolutely just drives me wild. And this is. It is where everything is going right.

And we talked about this a little bit, and I don't know if we want to expand this whole conversation into the death of ip, but I do find the whole situation quite comical in one way, if it wasn't so annoying in another, because for me, personally, hypocrites are the things that I can't stand the most. I'M quite happy if anybody wants to do anything they want to do other than, like, killing someone.

Pretty much whatever you want to do, if you're not harming anyone else, do whatever you want to do.

Ben Harvey: 00:02:50

Yeah.

David Brown: 00:02:51

Just don't criticize other people for it. Right. Like, you just own it. If that's your thing, then do your thing. And, you know, you have to be okay with that for me.

And so the whole hypocrisy of it, I think, is.

Ben Harvey: 00:03:04

Is, I think, especially as you think some of these companies are trained their data on illegal torrents of scientific papers, music, films, and, you know, and there's going to be big lawsuits around that. I mean, the hypocrisy is off the scale, really. Yeah, it is ironic. Very ironic.

Alan King: 00:03:24

It's very funny. I mean, when you think, you know, as you say, the scale of the sort of copying that's gone on.

And I think they've always worked on the basis, these companies that, you know, well, we'll just do it and then see if we get sued rather than actually sort of start with an ethical principle and think, is this actually the right thing to do? I remember seeing the.

I think she was CEO at the time of OpenAI a while ago, and she was being interviewed and she was sort of saying, you know, they asked her directly, have you trained on YouTube? You know, and she just would skirt around it. She wouldn't say no, obviously, because, you know, because they had probably. Probably allegedly.

But, you know, at the same time, you know, it was kind of like, well, you know, and. But you just know that OpenAI or Anthropo, whoever, it's going to be the moment there's something against them.

They're going to be the first to complain bitterly about this, you know, well, they're using our systems. Are they using our ideas? Or they've stolen this. And. But I think this is. We see this all the time, don't we? You know, in life, in general, anyway.

I mean, we were talking about this last night and I said that. But the, you know, the Radiohead example, where, you know, Radiohead are a band you would consider to be sort of, I guess, ethical. Right.

You know, they sort of, you know, they would have principles, you would think, and all that, based on the way they write their songs and stuff. And, you know, they'd obviously copied the Holly's song that yout Breathe when they made Crete.

Whether they did that knowingly, you know, but certainly, you know, the Hollies got credited for the. For that. For the Creep song.

Eventually, in the wrong on the writing credits, you know, and then, you know, a few years ago, Leonardo raiders a song which sounds uncannily like creep and they sue her. And you so think, well, you know, hang on a minute. Like, can you not see the irony here?

But no, it's like that, that to them would seem to be the perfectly normal thing to do. So think, well, maybe, maybe it isn't. So, so I don't know, I think it's human nature a little bit.

And you know, you, you hear the stories about how these, you know, LLMs were built and all the books that were scraped and the, you know, the Internet, I suppose, is one thing, right? Because that stuff that's out there to some degree that people have put out there.

But when, and you know, Reddit and those sort of threads, but when you, when you're actually, I think it was Meta said, you know, they sort of bought all these illegal books, right? So they hadn't gone and bought the books that were legally available and then scraped them.

They'd actually found a sort of library of, of illegal books where, where, you know, all this stuff was being stored and no one had paid any copyright or anything, you know, and scraped that instead. You know, it's, it's kind of. Yeah. You know, and you think of the.

Ben Harvey: 00:05:49

Volume of cash that's gone into these things, actually paying someone for copyright for some of these things would have been the ethical starting point and doing it correctly. And then, and we've often pointed these things out over the last 18 months as we've been talking about this.

But yeah, I mean, the irony and the hypocrisy that is off the scale, isn't it?

David Brown: 00:06:08

And I, I think two, two points, number one is I, when the very original model, I mean, they've been doing these models in labs in universities for decades, right? Like it's only, you know, chat.

GPT burst onto the scene because they released it to the public because it was far enough along that they thought it might be useful, right?

But they'd been building these models for ages and with, when you think about it within the university context and somebody in a computer science department is doing research, they have access to all of those academic texts. And so they're just doing research and it isn't for money, it's for educational purposes and whatever. So nobody really minded.

And so, you know, when, when the, when the models very first came out, if you remember, they sounded. Everything sounded extremely academic because that's the only thing they'd ever been trained on were Academic texts.

Ben Harvey: 00:07:01

Apparently it's where we get the long dash from.

David Brown: 00:07:03

Exactly. And that's where the EM dash came from, because EM dash is correct grammar and was used correctly from scientific papers and things.

So, you know, there was that bit in the beginning, but the other side of it was, is that it also wasn't being done for profit.

Ben Harvey: 00:07:20

Yeah.

David Brown: 00:07:20

And that was the whole kind of thing as well. Right.

Is that the whole idea behind it was that it was just research and as soon as you sort of explode onto the scene and you start doing that, then, you know, that was the whole idea. The whole thing was a nonprofit.

And I mean, everybody knew as soon as it went out that there was no way that that was gonna, you know, it was gonna, it was gonna make it as a nonprofit. Now, again, we've talked ultimately about what's the ultimate endgame in all of this.

You know, if we reach a point where we have quantum and we have fusion, the processing cost of that goes to almost zero. So at that point does it then become a.

Just a free resource that everyone has access to because it doesn't really cost anything to run it and no one's working. Like, it's going to be a totally different world when all that stuff happens. But, but until then, it's got to be a money making thing.

And if people are making money off of it, then they do need to pay someone.

Alan King: 00:08:23

It's interesting, the money making thing actually at the moment, because we're sort of seeing a bit of a fork, aren't we, between say, OpenAI and Anthropic.

We've got Anthropic on one hand who are sort of openly saying we will never run ads and actually making their own adverts to poke fun at OpenAI who are. Yeah, I think we'll talk about who are, you know, who are certain ads.

So, but you, you have this sort of thing, don't you, where you know, at some point you've got to make money out of it, like you say. And you know, they're spending enormous amounts of money to change these models and that's.

They're going to have to, you know, they're in an arms race as well. So it's not like we can just kind of get this model done and then that's it. It's, you know, it's like Formula one next season. You need a new car.

Right. So you've got to keep this kind of cycle going. And so I think, you know, how.

Then you can see how OpenAI might make some money because Obviously they've got, they're going to go down the Abbott route. People will choose whether that bothers them or not. They'll self elect and that business model will either work or it won't.

They haven't got much else, they have no hardware, they have no other kind of thing. Google, Apple, Amazon have other plays, don't they? They have other incomes, other revenue streams.

So they can to some degree afford to, you know, not a lost leader, but they can, they can, you know, subsidize it to some extent. I mean, anthropic.

I put in the same kind of boat as OpenAI financially at the moment, you know, although they are backed by Amazon, so they do have deep backers.

But, but the reality is that, you know, again, at the moment their only potential source of income will be through the AI systems that they're building. And they've already said out the gate we're not, you know, we're not going to add, run ads. So, so, so what's your choice then?

You end up with a world where, okay, we're gonna have to at some point paying the real cost transfer, which means my subscription is going to be £200amonth or it's going to be quite high or, you know, I'll go to OpenAI and I can still get it for £20amonth and I'm going to have access, but I'm going to have adverts pushed at me and I'm just going to live with that basically. So it's going to, it's going to be interesting to see how this, how this all sort of plays out.

And what we've always seen with other, other platforms like social media and other things, you know, eventually is what they call the isn't it of platforms and processes and I think that's where we will eventually end up. And then when you enter the world where advertising exists, does that break your trust with something like a large language model?

Because if you're asking a question and you're saying, oh, you know, I want to know the best kind of five trainers that are available on the market, you know, and it comes back with Nike, that's the one you need, you know, are you then sitting there scratching your head thinking, I wonder if Nike sponsor OpenAI.

Ben Harvey: 00:10:59

But don't you feel that already with search results quite often?

Alan King: 00:11:02

Yeah, well, but you know, obviously in a search result it's clearly labeled as ads, isn't it? Although I say clearly. Not always so clearly actually, but I don't.

Ben Harvey: 00:11:09

I think that that somewhat that happens.

Alan King: 00:11:12

But do you think like, you could have a situation with OpenAI where, you know, they're clearly labeling what the ads are. You go in and at the bottom there's an advert for something. Okay.

But then is also in the back of your mind thinking, actually if Nike are paying this company some money, would it, unbeknownst to everyone, just be prioritized in their response? You know, could I really?

Ben Harvey: 00:11:33

You can't help having that.

Alan King: 00:11:35

Well, you would think that, wouldn't you? I would be sort, you know, you know, is the recommendation it's making, which is sort of on the surface, independent actually.

David Brown: 00:11:44

So.

So we're going to get back to what my second point was, which I've talked about in almost every time we talk about this, which is it's no different than a human. That's exactly how a human would be. If a human is sponsored by Nike, doctors sell medicines because they get a kickback or whatever.

But it doesn't even have to be that direct. And the thing is we have no, you know, everybody complains like, oh, we don't know what the AI is thinking. We don't know how it came up with this.

You don't know what a human is thinking either. There's no way a human could tell you.

You know, I have no way of explaining to you how I reach a certain point, a decision and a thought that I have, it's based on the sum total of everything that's happened to me in my life so far. And my brain comes up with a funny connection and it spits it out.

Now sometimes that's going to be because I read a book and I'm just regurgitating what I learned in that book. Now, am I stealing someone's copyright because I've regurgitated their idea?

Ben Harvey: 00:12:41

Yeah, sometimes we think because I'm a human idea. Yeah, sometimes like the, the music writing, you sometimes think you have an original idea, but you've absorbed that culturally for decades.

David Brown: 00:12:51

There's three chords. Yeah, right. You write a three chord song.

Ben Harvey: 00:12:54

Yeah, yeah, right.

David Brown: 00:12:56

Like it's like every other three chord song. Yeah, right. And there's so many examples online where the guy's playing the same three chord chord progression and singing 40 different songs.

So you can't go, oh, well, that's my song.

Ben Harvey: 00:13:07

Yeah.

David Brown: 00:13:08

Do you know what I mean? Because there is, there, there are a limited number of options.

Alan King: 00:13:12

So I thought, you know, I thought that I've been thinking about for a little while then is, is that.

So when all this started, you know, the Reaction from a lot of, you know, publishers and artists and individuals is, you know, how dare you scrape my data? How dare you train on my data? You know, that's, that's not acceptable. You should be paying me something if you're going to do that.

But is there a point at which it sort of switches in the sense that actually people want to be scraped because if they don't, if they're not, they cease to become relevant.

And, you know, the example I would give for this would be, I think it was maybe the Beatles and the Rolling Stones and a few other big rock bands over the years. Initially, when things like Spotify and stuff came on, they held back on it.

They said, we're not going on that because, you know, our music is, you know, the highest order and people should have value for that and they should buy our records and CDs and pay money for, or at least pay downloading fees, you know, for an MP3 on Apple Music or whatever, or itunes as it was then. And. But at some point, you know, someone's obviously pointed out to them, do you know what?

If you're not on these platforms, the generation of people coming through simply will never hear your music and you will just become irrelevant. Sell anything, and you'll never sell any either anyway, and you'll just fade away into obscurity.

And I think there's a point where large language models become so prevalent in our society and traditional search goes away. I mean, you think how companies spend all their time doing SEO because they want to be found, right?

But if you don't allow yourself to be scraped and have all your stuff, you're never going to be found in an LLM, Right?

So maybe there's just an arc here that in the beginning everyone's worked up about it, and maybe this is what the long play for the large language model companies. Well, you know what? Eventually everyone's going to want us to scrape them because they want to be in our system because we're going to be the thing.

David Brown: 00:14:58

Well, that's the Google thread, isn't it?

Ben Harvey: 00:15:00

Yeah, I think the big thing, yeah, I think the big thing that will change that is the more we start to use voice and less screen as well, because, you know, you're driving along, you say, when I come home tonight, I want to have pizzas delivered. You're not searching, you're just saying to your large language model of choice, can you send me these pizzas by whatever time?

And I think that's when you, when that voice takes off for a Bigger and wider adoption. That's where you want to be on all the LLMs and your businesses want to be on there and you're sort of interacting on a. More of a voicemail.

Of course green's not going to go away, but it's, you know, for me I'm quite excited about that because one, the speed of it too. I love processing out loud through voice with LLMs while I'm driving or thinking of ideas and I think that will really transform things.

Yeah, I think people want to be found everywhere.

David Brown: 00:15:50

Yeah.

And there's a guy, and I think I talked about this once before, but there's a guy I know who runs, he does plumbing and gas and all that sort of stuff and he's got like eight different companies that he actually runs.

But he, he pays for the like the 200 pound a month version and he talks to it and he basically, he goes on a job and then he gets in his truck and he says, here are the notes from this job.

So he connects those notes to that client and then he's got some behind the scenes stuff that basically will go and then pick that up, it will add it into his notes, into the client, into his like CRM tool that he's got. So all of that basically kicks off the whole workflow that goes in the background. But he can also say, how are the other projects going?

And it will give him a report on all of the other projects that all of his other team are doing because they get in and they do their reports just verbally in the truck while they're driving to the next one. And so it will say, oh, everything's been on schedule on time, except Joe had some trouble with this and he had to order a part.

And it just gives him a summary of everything audibly while he's driving and you know, and he can interact with it and do everything. And it's amazing. He's like, I absolutely could not run my business.

Ben Harvey: 00:17:01

And one of the contractors do voice notes as well?

David Brown: 00:17:03

Yeah, they do.

Ben Harvey: 00:17:04

So it's all, yeah, everybody that works,.

David Brown: 00:17:06

They all do the voice notes. Yeah. And the system runs everything.

Ben Harvey: 00:17:08

One efficient way to sort of process your business. Yeah, amazing.

Alan King: 00:17:12

Well, and I think that's sort of when, you know, the OpenCloud thing came out and obviously that comes with a lot of risks at the moment and we wouldn't encourage anyone to deploy that unless they really know what they're doing. But you know, at this point, you know, it does sort of shine a light though on maybe what the future Looks like.

And I think one of, one of the really interesting things that didn't get nearly as much publicity as Openclaw, but when Opus 4.6 came out, the sort of upgrade to the aging capability there is actually amazing. I mean, okay, yes, it's not running a heartbeat in the same way that openclaw is.

So it doesn't wake up in the middle of the night and go, well, I'll go and do this. But it's still, it has to be triggered by the human.

But actually it's sort of, sort of team swarm agent capability where you know, you give it a task and then the sort of the, the lead agent, if you like, breaks that task down into a number of chunks, gives it to other agents.

Basically they all start working on different chunks of the, of the, of the task or the activity independently and then they're all connected together and they're all sort of cross feeding to each other and sharing information. And so you're doing that. Well, that means I probably need to do this a bit more then or.

And so it's being orchestrated basically, I think through Git in a lot of instances. And then essentially it's sort of all coming back together and it's all happening in parallel.

So it's speeding up this kind of, not this kind of chained process anymore. It's very much a team thing.

Your level of capability and outputs and what it can produce for someone like your mate is just escalating and going through the roof all the time.

This is a new system in the last few weeks, Opus 4.6, to some extent, you see something similar happening in OpenAI's Codex and we're going to see, you know, this thing, I think over the rest of this year just really accelerate.

And by the end of this year, I think, you know, at the moment, agents still require a little bit of understanding and knowledge from the user to really get the most out of them. But again, I think that melts away a little bit. I think in the end, you know, we'll be back to the simple text box input for most people.

So the everyday user, you know, the non technical user would better just jump in, start having a conversation and the agent itself will work with the user to make sure that they get what they need to do the job and off it goes.

And I think when we start to get those, that sort of level and it starts to get embedded into, you know, a bit like then the clawbot thing, you start embedding it into the actual device ecosystem. So maybe it's like Siri. Right. And it won't be Siri as it is now, but.

But a version of Siri in the future that has that real capability, lives, you know, inside your platform, on your computer, it has that connection to all of your tools and, you know, API keys and subscriptions and everything and suddenly has real capability and autonomy. Then you really do have something, you know, very interesting. And I think maybe, you know, a year ago everyone banged on about AGI the whole time.

Is this is the thing that gets us to, you know, this brave new world.

I think actually well deployed agents, you know, that are fully connected will be transferred so transformative that it will almost feel like that, you know.

Ben Harvey: 00:20:15

Yeah. They don't need to be AGI useful.

Alan King: 00:20:17

They probably don't want them to be, to be honest.

Ben Harvey: 00:20:19

Exactly. I've become obsessed by codecs.

Alan King: 00:20:21

Yeah.

Ben Harvey: 00:20:22

I've.

I've done three, three developments in the last week and I tell you, I know Claude on benchmarks is much better, but I just run out of credits all the time, so. Whereas I don't on OpenAI, I just. On the, the lower model I just, I've made three amazing. And it asks you what it can access.

So it's not just got unfettered access to your computer, it doesn't just go away and it's constantly sort of this feedback loop, you know, do you want me to give, you know, do you give me permission to do this, this and this and. Yes. So I've become really. I think it's an absolute.

Alan King: 00:20:57

Well, I'm obsessed with the R1 and for the same reason. And it's. The, the reason is, I mean, R1 has been a bit sort of standing joke, hasn't it, in the AI community since it was launched, really.

It's actually really interesting now because it's running Opus 4.6. It's running Claude 4.6 without credits, right. Without credits. And you plug it into your computer via cable, it becomes a dlan, right.

Direct large action model. It can see everything in your computer, it can go in, you just tell it what you want it to do.

It can be your tech support, it can install software for you.

Ben Harvey: 00:21:26

So who's paying for those?

Alan King: 00:21:27

Right. Software rabbit.

Ben Harvey: 00:21:28

Yeah. Yeah.

David Brown: 00:21:28

Okay.

Alan King: 00:21:29

I mean, good on. They just got another round of funding apparently, so I'm happy. And this thing will just jump around your computer.

When they first launched dlan about a month ago, it had a one hour limit, right. They lifted that. Now it will just run for as long as you need it to. Right. And it will do and build.

And the point here is that you're using Opus 4.6, right? That's what's happening. You've basically connected anthropics Opus 4.6 via this device. This device to your computer doesn't pay.

Ben Harvey: 00:21:59

For itself every year.

Alan King: 00:22:01

It's like 150 quid for a round.

So this thing is connected to your, to your system and it's, it's running, you know, and it's able to do anything that, that you've basically opened up, allowed access to on your computer, you know, and I had it doing tech support for my wife the other day. We had an issue with a computer, we can solve the problem. And so I plugged it in and explained what the problem was.

Within 10 minutes it had it sorted, you know, and it was like, fantastic, you know, amazing.

So it's, and look, you know, the solution in the end probably isn't plugging something in to do that, it's just having the system already on, on, on board. Right? You know, yeah, it's Siri or it's,.

David Brown: 00:22:40

Whatever it is, but that's step one.

Alan King: 00:22:41

But yeah, but you know, this, this is if you want to proof a concept, right, that this little thing right there is doing it, you know.

David Brown: 00:22:47

Well, it's interesting because going back to the. How do you pay for this thing? Sort of.

My idea that I had when you were talking about that earlier was that I think what we're going to end up is we'll end up with two different system types. You'll end up with your enterprise AI, right, which is designed for business. They're not selling it to the, to the general public.

A normal person wouldn't have access to the tool, but it would be a thousand pounds a month, £2,000amonth. But it would also have a lot of other protections in it as well.

So they'd have a lot of stuff about maybe not using your content to build models and all, you know, and I know in some of the tools now that's the way it is, right? If you're using the API, it's not actually using that, etc. But, but there'll be some safeguards built into it.

It will be more expensive, but that's where business will go.

And those companies will support themselves off of those feeds because they will have a significant revenue stream coming in from the people that are using the tools.

And, you know, maybe it's on some sort of, you know, graduated scale that, you know, depending on how many queries you're running or how Much, you know, processing time or whatever. Like aws. Right. It's just going to be another version of AWS but for AI and that's going to be the enterprise version.

And then you're going to have that personal stuff that's for the market. Right. Direct to the consumer.

But I think what, where you're going with what you're talking about now, the actual usefulness of the stuff is that's going to be that enterprise version. Right. That's going to be the one that companies can use.

And I think that's really what's going to drive the adoption because you know, like we, we both run production companies and studios, right. So you know, I end up, somebody comes in and records an audiobook.

It'd be great if I could take 30 hours of audio and go, can you do the rough edit for me?

Which is just go in and remove all the duplication and pick the better one and just give me a rough, give me a rough cut of everything once it's done.

That's not, that's literally just taking out some spaces, breaking up chapters and taking out duplicated sentences where they've reread something, you know, a couple times to get it right. That, that took me 90 hours to do on a 30 hour recording.

And so if an AI could do that in 10 minutes or less and save me all of that time and it's still not fundamentally doing anything, it's just literally finding, oh, this is a duplicate of that. I'm going to take it out.

Ben Harvey: 00:25:21

And there's tools out there, isn't it? You've used OPUS before and things like that. I'll try but that's, that's slightly different.

It's not the pre edit but I've actually tried with Codex to get it to write me a pre editor.

David Brown: 00:25:30

Yeah.

Ben Harvey: 00:25:31

And with using FFMPEG and a few other tools, open source tools and it like I got somewhere but it was, it was a long way from where it was useful. But I just wanted to test it.

David Brown: 00:25:42

Just to be clear, there is a site somewhere that says that they do that particularly for like audiobooks and stuff. But the point is it should be in. Yeah, those sorts of tools are actually.

That's useful and deliver something as opposed to the, the consumer version which is. I don't want to say it's not useful, but it's not useful in an enterprise kind of way.

Ben Harvey: 00:26:02

Yeah.

Alan King: 00:26:03

Well, I just got some notes here. You know, when we're talking about this, the multi agent from Anthropic. Okay. These are Things that sort of been building.

It's built a whole entire C compiler from scratch. It's compiled a Linux kernel, qemu, it's even fmmpeg, right?

Ben Harvey: 00:26:22

Ffmpeg, yeah. It's open source software.

Alan King: 00:26:24

It's debugging independently, it's mixing fixes, no babysitting, it's just getting on with it, you know, which is. Yeah, it's a different world.

Ben Harvey: 00:26:32

Was it Anthropic that also did the COBOL this week? And IBM share price?

Alan King: 00:26:37

Yeah, yeah, let's talk about that. Actually, it's on my list. So Cobalt. Sorry, yeah, no, no, it's a great segue.

Cobalt was really interesting because, you know, here's, here's a very, very old. I mean, so Most cobalt. Yeah, 50s, 50s and 60s. Most cobalt programmers kind of look a bit like us actually.

Ben Harvey: 00:26:53

Right.

Alan King: 00:26:53

Old, old gits in their 50s and 60s, right. They get paid a fortune by the way, these, these, I'm going to say guys, because it really is mostly guys,.

Ben Harvey: 00:27:00

A handful of them still going.

Alan King: 00:27:01

Yeah. So Hamza can get paid enormous amounts of money to go into, you know, old banks and sort out their back end legacy systems.

Ben Harvey: 00:27:08

Right.

Alan King: 00:27:09

And essentially overnight, Anthropic just killed their job basically. You know, produce a system that can do everything you need to do in and around Cobalt. Sort everything else you need to do.

Migrate it from Cobalt into whatever it needs to be in the future from Fortran. Yeah, yeah. So there's a bunch of old crusties like us anyway, scratching heads, going, damn it, like this stuff, you know, Most.

David Brown: 00:27:31

Of the guys I know that do that are desperate to retire.

Ben Harvey: 00:27:34

Yeah.

David Brown: 00:27:34

And they literally don't want to do it anymore. And they're like, they throw so much money at us that we have to do it and they literally don't want to do it. So they'll be happy.

Alan King: 00:27:44

The funny thing about it though, when I heard about the Cobalt thing, I thought, well, okay, and so COBOL is a kind of niche edge case, isn't it? You know, in terms of the program board. But how long is it till we wake up one morning and here basically the same story for Python.

Ben Harvey: 00:27:56

Yeah, right.

Alan King: 00:27:57

You know, everything, everything's changed then, hasn't it? Yeah, that to me, if you've done it for Cowboy, it's really not that far away, is it? Till we hear that.

David Brown: 00:28:06

Well, surely it should be able to do Python already.

Ben Harvey: 00:28:09

Yeah.

Alan King: 00:28:09

But to the extent where you can just say Python's not done, we don't need Python programs anymore, the computers can just do it and it's Better.

Ben Harvey: 00:28:16

Well, I think the reason that the Cobalt one's so interesting is because, you know, I think it can do a lot of the modern programming languages well already, particularly PHP and Python. Particularly. But I think the thing with Cobalt is no one touched it because it was such an edge case.

David Brown: 00:28:29

Yeah.

Ben Harvey: 00:28:30

And you know, IBM still has the market on it, doesn't it? Or the copyright. Was it like their share price went.

Alan King: 00:28:36

Down at least 10%? Yeah, I mean, I didn't, didn't watch it on day two and day three where it continued to go, but yeah, it certainly took a hit.

Ben Harvey: 00:28:43

But I think you mentioned this before, that a lot of SaaS companies are going to be sort of quite edgy now because you're going to be able to. And you say this almost every episode. This is as good, as bad as it's going to be.

So people are going to be able to write solutions that they're paying fortunes for quite easily.

And I know you don't have the security and big companies like to work with big companies, so there is that, but there is a, you know, some really specific software stuff that's going to, you're going to be able to do yourself.

Alan King: 00:29:12

There's always this case, isn't it, where over time it takes for companies to adopt stuff always takes longer than the technology in terms of the way technology develops. I mean, and that's just the nature of business and security and doing things safely and all that.

I mean, an extreme example of this is I was, I once had the pleasure to go on a nuclear submarine in Devonport dockyard in Plymouth. It was a V class and I was very excited about it.

So it's going to be state of the art stuff, you know, and I got down there and it was all just switches and dials and stuff because it is legacy, because it's reliable. Right. There weren't anything electronic on them and also electronics, easier to detect, you know, a signal from.

But, but you know, it was robust and I think for a lot of companies is why they're running Cobalt so. Because they're too afraid to. Yeah, to migrate from it.

So I think in one respect things happen very quickly and in another respect, you know, businesses change a little bit.

Ben Harvey: 00:30:02

Slower than that and the bigger, the slower.

Alan King: 00:30:04

But the point is you can see then things coming and you know that the change is happening and it will eventually get there. Okay, so let's, let's talk a little bit. We mentioned the adverts earlier and they've been sort of amusing me for a couple of weeks now.

I mean, obviously a couple weeks ago, we had the OpenAI sort of spat between Anthropic, where Anthropic basically put out an advert, sort of taking the Mickey of OpenAI, saying, you know, you'll be.

You'll be chatting to, you know, your mum or something, and then the next minute it will be sort of, you know, selling a kind of, you know, cougar ads or something. You know, it was sort of. And it was very funny, done very well. And I thought, yeah, good on you for that. That's quite good. I like that.

And then the super bowl happened and I think this. This sort of starts.

I think this starts off as a funny conversation, but actually there's a serious undertone to this which I want to dig into afterwards. But, you know, the ads at the super bowl basically rolled out. I think there was.

Let me get this right, there was one from Google, essentially, which was about a child and his mum, and they've arrived at their new house, you know, and it's all a bit sad because they obviously didn't want to for whatever reason, and all his stuff just piled up in a big pile, you know, and they're trying to.

Rather than sort of working with the child and helping him unpack and sharing that moment, she whips out her phone, jumps onto Gemini, you know, takes a picture of the. The stuff and then asks Gemini to imagine what it would look like in the house and shows it to him.

Okay, so on the surface you go, well, then that's quite nice. Is it? You know, it's sort of helping the child come to terms with his new situation.

So, yeah, so basically, basically it all kicked off online and Google got quite a backlash about this. It was a bit.

There was an ad last year actually, which was similar where I can't remember who it was now, it might be Microsoft, but it was basically a Christmas thing and they got the child to write a letter to Santa or a permacy and they used AI to do it. And again, people were like, this is not. We don't want to replace childhood with AI, basically. So that was one.

Ben Harvey: 00:32:01

Okay, Remember the Pepsi advert as well?

Alan King: 00:32:02

So then there's another advert that runs out in the super bowl as well. But bear in mind how much money these companies are paying for this, right?

David Brown: 00:32:08

And, you know, millions for that one.

Alan King: 00:32:10

Just enormous amounts of money. And it's, you know, so the ring doorbell advert, okay, so basically the scenario in this advert is there's a lost dog in the neighborhood, right?

And through the Ring doorbell network.

Through the surveillance of the ring doorbell to your neighbors watching, track down this lost dog and, you know, return it successfully to its owner. You know, in the marketing meeting when they had this, obviously the ring. This is great. You know, I'm going to find this poor dog wonderful.

Everyone's going to be delighted. The reaction online was kind of like, you, what, you're surveilling everything. You mean you can track everyone in the neighborhood?

My doorbell is part of that tracking system. And you didn't tell me you can just turn it on and do that whenever you feel like it. Backlash.

I think the ring doorbell stock price took a dip and they quickly started rolling it back. They went into full emergency PR mode. They had to put out statements saying, no, okay, we're going to turn off this.

You know, instead of it being automatically on, you'll have to elect to have it switched on yourself. You know, it won't just be on by default, because I think it was on by default, basically.

And of course, the point was that, you know, if you replace a dog with a person or something else or someone that ICE is after or whatever, you know, this just becomes a really malicious tool. So. So that was that. So again, so rings spend a lot.

David Brown: 00:33:31

Of money, but people do realize that they'll be able to do that anyway even if they turn it off. Right. Like law enforcement and the government will be able to do that anyway. So it makes no difference.

Alan King: 00:33:41

People don't realize it. That's. That's part of the problem, is it? You know, and, and, but I think after the Super. Super bowl advert they do realize.

Ben Harvey: 00:33:47

Yeah, that is a known goal, isn't it?

Alan King: 00:33:49

Yeah, yeah. And I'd be interesting to see like, you know, what, in about six months time or a year's time, what the ring doorbell sales profile looks like.

Whether after this it takes a little bit of a dip.

David Brown: 00:34:00

Yeah, I don't think it'll make any difference.

Ben Harvey: 00:34:02

Don't you? I. I'm one. I hate being recorded without my, like, I just don't like it. So the.

I've resisted getting a ring doorbell for that very reason, but that just kind of highlights it to me. That is even worse than I thought in a way.

David Brown: 00:34:15

But okay, so I'm a fan.

Ben Harvey: 00:34:18

Yeah.

David Brown: 00:34:18

I quite like.

Alan King: 00:34:20

Yeah.

David Brown: 00:34:20

Doorbell and the sort of.

Alan King: 00:34:22

I know, I don't know.

Ben Harvey: 00:34:23

Yeah, I know. I'm behind the curve on this. I know my family members have them and they love them.

Alan King: 00:34:27

I saw a really funny sketch a few years ago. Comedian. I mean, back in the Day. You know, back in the 80s, people would just come and knock on your door, wouldn't you?

You know, and these days, the comedian sketch was basically. Basically, you know, that the family is sitting around and there's a ring on the door. Who's that?

Ben Harvey: 00:34:41

Who.

Alan King: 00:34:41

Who's. Who's kind of like this panic like no one brings. People always text first.

Ben Harvey: 00:34:46

Now they're all looking at the phone.

Alan King: 00:34:49

Knocks on your door. Yeah, yeah, yeah. It's just. But look. And so there was one other advert as well that ran, and this was an Amazon advert.

I forget the name of the actor now. I think it might be Chris Hemsworth from memory. Anyway, don't shoot me if I'm wrong, but essentially the advert was AI going bad, right?

And so in this advert, you know, he's going around and, like, the garage doors coming down on him, or, you know, the hot tubs drawing it getting too hot, or the AI's gone a bit rogue, you know. No, this is like a fever dream that he's having. So it's not supposedly actually happening, but again, people just kind of went, yeah, okay.

Even though it's supposed to be tongue in cheek in the sat, you know, satiring people's worries about AI, also, it could happen. You're just highlighting the possibilities. So, again, I think, you know, it just didn't land. So.

But it got me thinking, the sort of bigger picture, which is what I wanted to pick up with you guys on really, was this. That, you know, AI, I think, well, we. We're in a bubble, right?

You know, we talk about AI endlessly, we're in groups, you know, we're quite deep in the weeds on it. We're nerdy about it. You know, we have our views, but out in the general public, you know, people's perception of AI is probably very different.

And I think these adverts kind of serve to show that there maybe isn't that much fondness for this stuff and that, you know, people do see it as a threat and a problem in many instances. And.

And I just wonder, as we go forward, as this stuff does become more pervasive in society and, you know, more ubiquitous everywhere, you are everything integrated into everything, that we end up in a place where you could almost have a bit like the Luddites, you know, sort of back in the turn of the 19th century, where they're very anti technology. So you get this kind of anti AI, definitely starting there and. But really building.

I mean, at the moment, I think it's probably there, but Fairly low test.

Ben Harvey: 00:36:37

Outside the data center in America last.

Alan King: 00:36:39

Week, wasn't there, was there?

Ben Harvey: 00:36:40

Yeah, yeah, yeah. So it's that Luddite thing is been around, not I. You know, there's. People are worried about their jobs basically, aren't they?

Alan King: 00:36:48

Ultimately, I think this starts to build over time and, you know, I think those adverts in a way just kind of highlight that people aren't that excited. I think the tech companies are very excited about this stuff and they have this vision for how people should live their lives.

I'm not sure that humanity entirely agrees with them. And at some point there's going to be quite a pushback.

Ben Harvey: 00:37:07

Yeah, well, I think there's a couple, two or three points there. One of the big points is you've got a handful of people who are more powerful than governments.

You know, eventually, you know, the people like, you know, OpenAI and, and the leaders of these companies are going to be. They're huge, long. Yeah. So that doesn't feel a comfortable place for the world to be.

You know, mass job loss and sort of, you know, I've got experiences of losing out on work already because I haven't responded to someone who's asked that where I got open AI to do it, you know, like a Christmas card is like little things, but they are starting to build.

And I think just that taps into that sort of whether some of it is, you know, conspiracy theory, not, but it taps into that global sort of, you know, I'm losing control to, you know, to something that's going to not only take my job, but is also going to be able to dictate how the world is run. You know, we talked about it in cars or we talked about it with lots of policies, you know, 15 minute cities, all these things.

There's a feeling that that's getting away from us and AI is going to have a big part in that.

David Brown: 00:38:12

Yeah.

And I think, you know, you touched on globalization and I think what bothers most people about it, I think now is the fact that I think it used to be maybe 100 years ago, your local government and your local, even your country's government had some control.

Ben Harvey: 00:38:31

Yeah.

David Brown: 00:38:31

Right. So when you elected somebody, those people actually had control to make decisions that could impact your life.

Ben Harvey: 00:38:39

Yep.

David Brown: 00:38:40

Now, that's not the case. You can't. And this was the whole argument behind Brexit, and let's not get into Brexit.

But the idea behind it is that here, you know, you're in the mumbles, right. You've got a local council Here. Right. The local council can pass any regulation they want.

Ben Harvey: 00:39:00

Yeah.

David Brown: 00:39:01

But EU law is going to supersede that on a lot of. Yeah, on a lot of areas. So let's say you decide to manufacture some.

You know, we come up with some new camera design or some new tripod design or whatever it is.

Ben Harvey: 00:39:13

Yeah.

David Brown: 00:39:14

And it's like, fine, you can design it to the UK requirements, which are the only UK requirements. That's fine. But the problem is, is that you can't sell it anywhere else because all the regulations in other places are different.

Ben Harvey: 00:39:25

Yeah.

David Brown: 00:39:26

So if you want to sell it in Europe, you have to build it to the EU regulations, whether you want to or not.

Ben Harvey: 00:39:31

Yeah.

David Brown: 00:39:32

And so it's probably not the best example, but it sort of is what I mean. So we've now lost control of our local area because the global. Because we now have a global market.

Ben Harvey: 00:39:46

Yeah.

David Brown: 00:39:47

And that's for jobs, that's competition for manufacturing, that's everything. And so I think that's what it's getting to, is people are. It's that feeling of lost control and that's. It's this whole globalization thing.

And the AI is only making. And technology is only making that work.

Alan King: 00:40:04

I think I can very quickly as well become the kind of the whipping boy.

Ben Harvey: 00:40:08

Right.

Alan King: 00:40:08

Because when people start losing their jobs, I think that's the moment, really. And I mean, it's happening on a small scale at the moment to some degree.

But if we see mass job losses, let's say we see 20% of the workforce go, right, suddenly you've got 20% unemployment in a country that's unheralded, you know, in sort of modern. In the modern era, if we suddenly end up in a place like that, AI very quickly becomes, you know, the devil, basically.

This thing that's basically destroying everyone's lives, it's causing social unrest. You know, you'd be having riots, you'd have all sorts of things going on. And at that point, you know, then.

Then there's a bit of a reckoning, isn't there? Because as a society, we have to decide what we do with this thing. Because it's one thing in the Internet which gives everyone.

There's a very big difference that people often don't fully understand. You know, the Internet, sure, that distributed information to everyone, but this puts intelligence potentially in everything.

And like you say, decision making now is being taken by a machine, not a human. And you could be on the wrong end of that decision. That starts to feel very uncomfortable. Right.

You know, at least with a human on the other end of the wrong decision, you can think, well, I'm going to go back and complain about this and there's some course of action, but suddenly if an algorithm is, is just deciding that no, you don't get that, or you don't get this or you don't get that opportunity, or you feel like you've been discriminated against for some reason because of an algorithm, I think people will start to fight back against that much harder than say fighting back against humans in some respects. So it's. I. Because it's very easy to demonize as a thing. Right. Because it's. Because it's not human. So it becomes easier to do that.

So it's going to be very interesting to see, I think, how, you know, this, this plays out and be interested, look at the ads next year at the super bowl. Whether the AI companies are kind of all in on it and try again or whether they just kind of go, we're just going to leave it, I.

Ben Harvey: 00:41:56

Think, I think they'll change.

I think you start to see, I think with Chevrolet did a, you know, old American value style advert recently and it, you know, the car that lasted for 30 years and they went back to their childhood home where they raised their kids and it was a really emotional, well done advert. All not single bit of motion graphics in it, no AI. And you know, the praise it got by just across the board.

David Brown: 00:42:20

Yeah.

Ben Harvey: 00:42:21

Because it, that, that is going to be the narrative, you know, that a lot of people are going to lean into the authentic, the non digital that. But you know.

Alan King: 00:42:30

Well, there's a show I've just watched on the BBC for viewers in the uk definitely check this one out. It's called Small Profits.

It's a wonderful little show and there's some animation in it for some of the characters and they use stop motion and they could have used AI very easily. Right. And it would probably been okay. But actually the stop motion was delightful. Yeah, absolutely. So.

So, you know, there's definitely that sort of craft element to this thing.

David Brown: 00:42:53

Yeah.

Alan King: 00:42:53

You know, I still can't quite believe the Coke advert that came out this Christmas. You know, to me that's just unbelievable. But also when their strap line is, you know, always the real thing or something, isn't it?

It's just like using AI, like that's just.

David Brown: 00:43:08

So can I drop a bomb in here?

Alan King: 00:43:10

Yeah.

David Brown: 00:43:11

So I. The one thing that we haven't really talked about, but that did just come to mind because we're talking about 20% job losses. Right.

That's what made me think about this. The knock on effect of that though is you've got 20% reduction in tax income for the government.

Alan King: 00:43:29

Yeah.

David Brown: 00:43:29

And those job losses are going to have massive impacts on the amount of money that governments have to operate.

Ben Harvey: 00:43:36

Yeah.

David Brown: 00:43:37

And I know that, I know that this is a huge topic of discussion within government agencies and within governments around the world is trying to understand how do you tax that or how do you get that money back from the employees that are lost. So, so I think, and I think we should dedicate an entire show to this.

Ben Harvey: 00:44:00

Yeah.

David Brown: 00:44:01

Maybe later and maybe get somebody else on with us who's like an economist who's maybe at the front edge of this. And I have some contacts in Cambridge and Oxford so maybe we can get somebody who, who has thought seriously about this.

But, but the only thing I can see happening is that there is going to have to become some sort of attacks that goes along with the usage of AI.

Alan King: 00:44:27

Yeah.

David Brown: 00:44:28

That would equate to an employee in a business.

Ben Harvey: 00:44:32

Yeah.

David Brown: 00:44:32

And so I don't know how you do that calculation because AI is so much more capable than a single employee. And I don't understand the metrics there to say, well, if you're using AI, that's equivalent to 10 employees.

So therefore the average wage for these 10 employees would be X. And therefore you have to pay things like national insurance or whatever. Right.

Like and you've got to pay the employment taxes that you would pay for those people because just because you're using that instead of people.

And I don't know if that situation will actually push back on AI and people will end up using humans instead because actually it will cost less at that point to use a human than it would for AI. And then you really only use AI for very specific, very high end purposes to do very limited tasks.

Alan King: 00:45:27

I don't know, do you think, I mean one, one thing that, I don't know how you would do it actually, but be hard to get them to pay the money.

But do you think rather than going to the employers, you just go straight to the AI company to go to, go to Elon and say, look, you know, you're responsible for 20% job losses, there's a drop in revenue, you need to pay the government X maybe and you can actually find them or put a levy on them for if you're going to roll these tools out, they're going to wipe out all these jobs, then you know, you need to pick up the lost government revenue that we were going to get maybe because you've.

Ben Harvey: 00:45:57

Caused that and then you get to a socialist utopia of universal credit.

Alan King: 00:46:01

Well, I think that's most definitely an episode we should have.

David Brown: 00:46:04

But that's how you, that's the only way you could pay for it because otherwise there's no money to pay for it.

And knowing how governments work and how they think, I suspect there would be an ulterior motive to that because as well the flip side of that is if you do go to the AI companies that's much easier. Right? And it becomes a. We're just going to tax the shit out of them. Right, but, and that makes it easier.

But then that also now means that you don't need such a big tax department in your government which then puts tens of thousands of people out of jobs because you don't need them to do it.

You go to something like if you were going to reorganize the system today, you would probably say look, there's a flat tax, it's X percent, whatever that is, 10%, 15%, whatever on your income. That's it. Everybody pays the same tax. There are no deductions, you don't get any write offs, there's no complications.

You take 70,000 pages of tax law down to five. Yeah, it becomes very simple.

But the knock on effect of that again is these massive, massive tax organizations no longer needed the cost of the government no longer needed, the entire businesses that run around tax, no longer needed. And no government is actually going to make that decision because they don't want to be the ones to put all those people out of work.

So you've got this weird thing because when you're in government you, you have to make decisions around things like that. So even though something seems like the right thing to do and it's much more efficient and it's easier, that wipes out yeah. Careers.

Ben Harvey: 00:47:45

Well and also, you know, governments can be popular doing that.

David Brown: 00:47:47

And yeah, they're not going to do it because they're politicians and they're like I'm not going to be the one to do that.

Alan King: 00:47:52

Big tech are going to have to help find the answer I think because they need society to continue and to be stable. Right. For that, for them to continue to be successful as a business. Right.

Microsoft, you know, Play I X all these companies, it's no use to them if there's complete and utter societal meltdown that, that, that damages their business enormously. So you know, and they want to obviously continue where they are at the top of the pile presumably.

So, so they should Be ought to be incentivized to try and find and help solutions. But it's going to, I think, you know, it'd be interesting to come back in 200 years, wouldn't it?

Because you sort of think that the way society functions in the model probably has to look fundamentally different than what it does now.

Ben Harvey: 00:48:30

You know, and don't forget this is, you know, a lot of these job losses are going to be in the knowledge industry.

They're not, you know, in previous revolutions and big leaps it was often the working class or the, you know, that lost out because of improvements in equipment or technology. Whereas this is actually, you know, the knowledge people are going to go, you know, junior lawyers and junior doctors, junior, you know, creatives.

It's, it's so, yeah, it's gonna have a huge impact. It's gonna be interesting to see how.

David Brown: 00:48:59

That it's already happening in, in PR and stuff.

And again, I've talked about my friend who runs a PR company and, and you know, she had said in the beginning when I talked to her very early in my own podcast when I talked to her, she was of the opinion that AI couldn't replace it. It's very human driven business and you know, it's a lot driven on relationships with journalists and all sorts of stuff.

And then I talked to her recently and her opinion on that had changed and she was like, yeah, you know, we just got rid of 14 interns because we didn't need them. And you know, she's like, it was actually more trouble dealing with the interns than it was, than it was worth for the business.

And my people who were on the team could just do the work using.

Ben Harvey: 00:49:40

AI and the difficulty, and again we've said this almost every episode is for some industries, like going through being a junior lawyer, for example, you know, it can take 10, 15 years before you get there, but if you're not doing that process anymore, you're never going to get the, the seniors, the seniors coming through in 10.

Alan King: 00:49:57

Well, I suppose the end, you don't need them. The end game is you don't need them in turning is because the systems are so good that you don't need them either.

Ben Harvey: 00:50:03

And then.

David Brown: 00:50:03

Well, that's exactly. But this is the thing. So we're on a rolling clock.

Ben Harvey: 00:50:07

20% Of job losses seems kind of quite conservative then you know, it might.

Alan King: 00:50:11

Be there's always this assumption, isn't there, that you know, but you'll still need the senior really, you know, clever, smart, important people actually know. You know, that goes eventually as well,.

David Brown: 00:50:21

You know, and, and the other one. And, and we may have even talked about this last time.

The whole thing with radiology, right, is that because AI, the scans, the analysis by machine learning or AI of the scans is more accurate now than it is having a human do it. And it's faster.

Ben Harvey: 00:50:37

Yeah.

David Brown: 00:50:38

So everybody was like, oh my God, there's going to be no more radiologists.

Actually they need more radiologists than ever because they need to deal with the volume of people coming through because now every doctor is sending people to radiology because the scans are more accurate and they're faster. So it's like, oh, well, we can just do as many scans as we want now.

So they now have to physically deal with the people coming through the pipeline, which only humans can do that part of the process. And so they need more radiologists than ever. So they've now got more radiologists training than they've ever had before. Yeah, because they need them.

So I kind of.

Ben Harvey: 00:51:16

There's new opportunities, you can see where.

David Brown: 00:51:18

Some of the new opportunities to the robots are.

Alan King: 00:51:21

Right. But I saw Elon actually talking about the Optimus robot the other day and he was sort of saying, and this is Elon, of course, right.

And his timings are always wild, but you know, he's basically sitting there saying, look, you know, in the, in his not too distant future, which is probably next week, but in the not too distant future, you know, you'll have this robot in your house, the Optimus robot that, you know, people will be able to afford and will be available and it will be the best doctor you've ever met, it will be the best lawyer you've ever met, it will be the best surgeon you've ever met.

It will be to do everything that you need that system and it will have agency and it will be to physically, you know, move around and do things as well.

So, you know, and I, I don't think that's happening anytime soon, but it probably is in 100 years, you know, and 100 years is not, is only one, one life, human life generation, really.

So it's, I think that, yeah, when we get to that stage, you know, all the interns, all the extra people that we might say, or we've got more people coming through now, so we need, that does go away eventually, you know.

And so as a society starting to work out what that world looks like and how you start building towards it, I think the countries that get that right will be okay. I think the problem is really going to come in that the powerhouses here are in America and China.

So if you're the uk, if you're the American government and you're saying to Elon Musk, hey, you need to give us X billion pounds a year to meet up the kind of, you know, the government shortfalls because you've wiped out all our income, they might do it. But if the UK government say that to Elon, nothing, you know, you're not America. So, you know, that's your problem. Yeah. You know, consume me.

Do you know what I mean? Yeah. How are you going to get the money?

So, so basically any country that's, you know, non America is probably going to have a real problem trying to access that money because they're not going to start giving, Elon's going to start giving out money to Iceland and, you know, Brazil. I mean, it's just not going to happen, is it? You know, it's not going to cover the whole world.

David Brown: 00:53:18

Well, then again, but this, this gets back to the whole reorganizing how the tax system works and everything else globally because now you're in a global market that you've never been in and so the laws aren't really designed or set up.

The cynical side of us would say that they're designed that way on purpose so that only people with a certain very high level of knowledge understand how they work and know where the loopholes are. But those loopholes are designed for the people with money to be able to continue to have money.

And, you know, this is why, like, if you want to be rich, if you're from a poor background and you want to be rich, there's kind of levels to the game, right? The first level is you got to get a job. The next level is you then have to be able to make enough money that you can afford an accountant. Right?

And then that accountant opens up a whole new level of managing your money and understanding how to do things. But you just need an accountant.

So you get to that level as soon as possible, but then you work with an accountant for a few years, you've got that accountant. But then you need a wealth manager because that's a very different skill.

So the accountant's dealing with cash flow and how you're managing your credit and all that sort of stuff, whereas then you're building up a little bit of pot of money and then you go to a wealth manager and, you know, there's levels to this game.

And I think all of the loopholes and all of the, you know, the company director stuff and all of that for, for ages was all built around that wealth manager, accountant sort of knowledge level. And again, that's all coming out in the open. And so now you're getting tools that can give you that advice.

And you know, the AI tools are going to be able to. We're going to have a finance and an AI, you know, an accounting trained AI that will come out that will be able to do all that stuff for you.

It'll do. I, I don't understand why AI doesn't do bookkeeping for everybody already. Anyway, it's pretty simple.

Alan King: 00:55:11

Go back to your comment earlier about sas. You know, this, it's AI becomes the software. Right. You know, I mean, basically. And you know that a lot of software that we use today will go away.

We will be replaced with AI systems. It will just do, you know, do the thing you need basically and give.

David Brown: 00:55:26

You the information on the fly.

Alan King: 00:55:27

On the fly. This is what I need. You provide it, you know.

Ben Harvey: 00:55:30

Yeah, I mean, my eyes open this little bit.

I know I've been playing around with some codec stuff, but, you know, I've got a family member that lives in Portugal and he often invites people to do this. He sends around a spreadsheet of booking days and stuff. So I spent a couple hours and just made him an online tool with a calendar.

And it pulls in all his favorite football teams games into the calendar, all the opening times of the local attractions. It's, you know, it has a mini golf leaderboard for whoever goes and visits and the reviews. I mean, Codex did it all like it found the.

Even asked me, do you want to put in affiliate deals? Like it's like what it can do. I mean, so if you take that to your own personal finances.

And I've seen people online saying they've downloaded a few years worth of their bank statements and got some of their agents to do cancellations on subscriptions they don't need it tells them where they're wasting money, tells them where they need to be spending more, spending less. And you know, you spend 14% on eating out at Greggs this month, you know, all that kind of thing. But giving them really nice reports and suggestions.

Alan King: 00:56:37

Only, only, only 14%. That's right. We can do better than that.

David Brown: 00:56:41

My wife tells, My wife tells me every day. Did you go to Costa? You went to Costa again.

Ben Harvey: 00:56:47

We should do a Greg's leaderboard, shouldn't we? Number one in Swansea.

David Brown: 00:56:51

But, but you're right. I mean, I can see. And to take your point that you mentioned earlier, Alan, like building specialized Tools like I've never been able to find a tool.

So I, for those of you who don't know, I run a small studio, video studio. We have an audio room and we have a video room. So we have two spaces that we need to manage and can put time in.

And I've not been able to find anything online that accommodates the two room model in the way I want to do it. That's tied into my CRM tool so I can keep track of everybody.

So when somebody books for three hours of a session and maybe they want to split it into two, like just keeping track of all that stuff in a nice way. And I could very much see where I could probably, if I were so inclined, I could probably spend a weekend getting it.

To build a simple CRM tool with a webpage interface where people could go in and interface with my Google Calendar or create a calendar where people could go, yeah, they could book what they want and it connects to our Stripe account so people can just go and book and pay online.

And it, I'm sure it would do that and it would build all the specialized stuff around video studios and the sorts of work that we do around editing and you know, planning and production and all the other stuff that we need. I'm sure I could do that. And I don't need a notion or a hootsuite or any of those because they don't do.

Ben Harvey: 00:58:18

You're making exactly what I need.

David Brown: 00:58:20

But I could probably just make something and if I said I want it to be sort of like this.

Ben Harvey: 00:58:24

Yeah.

David Brown: 00:58:25

But here's what I specifically needed to do. I'm sure I could just build my own.

Alan King: 00:58:28

Oh, you would be able to. Yeah, yeah, yeah. I've had a lot of fun or.

David Brown: 00:58:33

I can pay someone to do it,.

Alan King: 00:58:34

But a lot of. Well, don't pay them, do you? Yeah, for sure. I mean I've had a lot of fun building, you know, very bespoke apps I suppose for very specific tasks.

That. Yeah, you know, you wouldn't, you would never. No company's ever going to make that app. Do you know what I mean?

But you know, think actually there's a very sort of unique workflow or thing that I do and this can just do it and, and then you can set up like a shortcut, like a trigger word. You can just say the word to the thing and then it just gets on and does it, you know exactly.

Ben Harvey: 00:59:00

What it is I use every day. Yeah. Just. Yeah. Amazing.

David Brown: 00:59:05

Do you think about. For us, right. Like you think you get a New client. And it's like, here's the new client.

And it can go and create all of your directory structure that you need for your filming and your recording. It can go and do that automatically. It creates a customer record. It then can go in and say, yes, I've created all the directory structure.

It takes that off the list of. To dos, you know, it creates all of your release forms that you need. Right.

Like you start putting people's names in, it creates all the forms for you. Like, you don't have to go and do that. There's the cost sheet, like everything. Right. It can send your email, hey, welcome to the studio.

Glad you booked a session. Like all that stuff it can just do automatically. And again, this gets back around to agents, right. So that's more of an agent kind of thing.

But it's like having someone on your team who can just go and. But I just go, thanks very much. Or they put through the website and it just all happens.

Ben Harvey: 00:59:53

Yeah.

David Brown: 00:59:53

And then when it comes to the data filming, it's like, right, okay, we filmed the thing and all I do is upload the video files and then it picks it up.

Ben Harvey: 01:00:01

Yeah.

David Brown: 01:00:02

And I do know a guy who has a very complicated AI system that goes and actually processes all the videos and everything, cuts it into social clips, gets everything together, like sends all the emails and everything. And he's already built that on the back end.

Ben Harvey: 01:00:15

He did himself to be.

David Brown: 01:00:16

Yeah, he did it himself. So, I mean, he can do. He can produce an entire show and get the social clips out in like four minutes after the recording's finished.

Alan King: 01:00:24

Yeah, I was using. I was using a tool the other day. I did a podcast and a different podcast the other day. And not. I was the guest. I was a guest on it.

It wasn't my podcast.

Ben Harvey: 01:00:36

I feel dirty.

Alan King: 01:00:38

It's called the Chief Uncensored. I did post it in the group, actually, but the chap who set that up for me to do it, he. He sent me the full video at the end.

And I can't remember the name of the tool now.

I'll have to look it up, but I popped it into this AI tool and it just chopped the whole thing up into, you know, sort of two 30 seconds, 60 second clips. Yeah, that's it. Yeah, yeah, Opus.

David Brown: 01:00:59

Yeah, Opus.

Alan King: 01:00:59

Yeah, yeah. You know, it's free and I just kind of. There you go. There is a paid version. I'm sure it does something else, but.

But I mean, you know, I was impressed by it. You know, it was some interesting. And it found Some really good.

David Brown: 01:01:11

So what we do is, is I use opus, so I get a rough cut of a video. I. I send it to the client and I upload it to Opus and I have OPUS run its algorithm on it. It comes back, I say, tell me the ones that you want.

They choose the ones they want and then I go cut them manually from the actual video because it looks a lot better.

But it's that decision making process of just identifying what are the interesting or potentially interesting bits of the conversation that OPUS has actually got really good at. And it's got even better lately because it now will take different segments and pull them together.

Ben Harvey: 01:01:47

Yeah.

David Brown: 01:01:47

Which it didn't used to do. It used to be a continuous bit, but now it's actually starting to edit, which is also quite interesting.

So it's, you know, it's starting to pull stuff because I did that the other day and I was like, oh, wait a minute, that's not.

Alan King: 01:01:59

If anyone from OPUS is listening and they want to sponsor this show.

David Brown: 01:02:01

Yeah, if they want to sponsor.

Alan King: 01:02:02

We're here for you.

Ben Harvey: 01:02:03

And this episode is not sponsored by opus.

Alan King: 01:02:06

There we are. So I've got one more topic I want to throw out there today as well. If you've got any others, Ben or.

Ben Harvey: 01:02:14

No, I. I mean, I know we talk about sometimes at the beginning about what's impressed us with AI recently, and mine was just Codex and one GPT that someone else has done for learning. I'm redoing some of my drone exams and the way that helped me learn was just off the scale.

I can't remember what it was called, but now I'll bring that up next time, but let's go on to the.

Alan King: 01:02:34

Yeah, so the last one for me was just Amazon put out a statement this week saying that they'd had over 600 firewall breaches by AI, cyber security breaches by AIs in like the past week or something, you know. Yeah.

It's got me thinking that, you know, we're probably just now going to enter this new world, aren't we, where AI is both the problem and probably the solution as well. You're going to have this kind of, you know, AI's fighting AI wars, basically. One's trying to get any other ones trying to stop it.

And I just working out, you know, who wins actually.

Is it, is it the, the kids sitting in their bedrooms with the, with the AI tools creating stuff, or is it the big tech companies trying to build the, you know, build the, the barra.

David Brown: 01:03:15

But I think we as consumers of Those platforms win ultimately because of that battle.

Alan King: 01:03:22

Yeah.

David Brown: 01:03:23

So I, and I saw another article where somebody was talking about, they, they set, I don't know if it was openclaw or something but they set the agent off against their, their system and it found 500 vulnerabilities.

Ben Harvey: 01:03:36

It was on a massive SAS, wasn't it?

David Brown: 01:03:38

Yeah, it was a. Yeah.

Ben Harvey: 01:03:39

Which had a whole department for looking for bugs.

David Brown: 01:03:42

It was the, it was actually the cyber security guy who posted it. Right. He's like, I'm out of a job. Because he's like, I have a whole team of people. Maybe it was on X or something. I think I saw it as. Well.

Ben Harvey: 01:03:51

Yeah. But they were saying like that there's a whole industry around this that hadn't found these 20 year old problems on this major software.

David Brown: 01:03:57

Yeah. And it's like the system just went and got it.

Ben Harvey: 01:03:59

Yeah.

David Brown: 01:04:00

So I think it's ultimately, it's good for us.

Ben Harvey: 01:04:04

Gotta trust your AI as well. That's the trouble.

David Brown: 01:04:06

And it's. Well, but I think so.

But I also think that there's a lot of stuff that probably hack, you know, proper hackers have known about for years and years and years and so they've been able to get into any system because a lot of this stuff is so low level that nobody even knew about it.

But now the AI is surfacing stuff that I think only a very high level that the, the very highest level security person would know and be able to find. It's now bringing that level of knowledge to everybody.

Ben Harvey: 01:04:35

Yeah.

David Brown: 01:04:36

And I, and it's. But this, this is the risk. It's doing it across every industry.

And I think I mentioned yesterday when I was, when we were talking while I was waiting for the train or I left a message. But you know, knowledge used to be expensive. It's not expensive anymore. Yeah. It has no value anymore. And so everybody can have access to it.

Good and bad.

Alan King: 01:05:00

Anyone whose business is built on selling knowledge is in trouble. I think. And in a way that's been true for a while. But AI is just absolutely amplified that and made it much more accessible and easier to.

Because you can, you know, you can have accessibility to say a load of law books, but you yourself need a certain capability in order to understand and, and actually then use that. How to apply it and apply it. But the AI can do that now. So it takes away that friction. It just, it basically becomes the answer engine that can.

You've got the problem. You describe that. It gives you the answer you need. You know, so it's interesting Isn't it?

David Brown: 01:05:35

And even my sister in law, who's a lawyer and my brother in law, who's a lawyer as well. It's like, you know, they're losing the industry.

Not them specifically, but the industry is losing a lot of the low level jobs that used to fund what they did. Same with creative businesses. Right. Like you don't need an, you don't need a lawyer. We don't need a lawyer to help us write a release form.

No, right, Like AI can do that and we can read that and go, yes, that covers, right, that covers everything we want to cover. Or you can give it a contract and say, you know, what's in this contract? Is there anything unusual in this contract? Am I missing something?

And it will just come back and tell you. And it's, it's. Yeah, it's pretty simple to do those sorts of things.

But that was the bread and butter of a lot of corporate lawyers and stuff like that. Yes, there's still a need for the high end. You know, you're talking about civil cases and all sorts of complex things.

But those people whose bread and butter was just generating those sorts of documents and working on agreements and standard business contracts and all that sort of stuff, that's all going away.

Ben Harvey: 01:06:34

Yeah, I mean, take what we do, video production. You know, 10 years ago we used to make quite significant amount of revenue doing captions, you know, for.

We used to do a lot of long form courses, you know, six, seven month projects that then would have 15, 20, 30 hours of captioning to do. And DVD production, you think There wasn't work I enjoyed doing, but it was, but it was a source of revenue.

David Brown: 01:07:02

Yeah.

Ben Harvey: 01:07:03

And so you would never dream of, you know, still now maybe take me a couple of hours to do some of that, to check it, to realign it, but a huge amount of that work's gone away.

David Brown: 01:07:13

We were complaining, my business partner and I the other day were complaining about the fact that somebody wanted captions on their video.

And it was going to take us like a half an hour to go through an hour and a half long video and check the captions, not just put them in, just check and make sure that all the words were correct and that they were spelled in the right way. And it was like, I mean captions.

Alan King: 01:07:34

Still, particularly live ones, are often wrong, aren't they? It misunderstands the word. And said something slightly different.

David Brown: 01:07:39

But multiple days it was like we were just like, oh, do we even want to do this?

Ben Harvey: 01:07:44

We actually had to listen to it and type it out. You know, there was no tool to.

David Brown: 01:07:48

Translation and all that stuff. I mean, you know, now you. You throw it in a. In a tool like all the.

All the major editing tools have it now, of course, but you do your video, you put it in, you get your final edit done, and then you literally click a button and it adds all your subtitles and everything to all of it, will generate a subtitle file if you want to, upload it to YouTube or whatever. And then, you know, off you go. And it's. Yeah, okay, it needs some checking, but that's a tenth of the time and it's ridiculous.

It's great, though, at the same time.

Alan King: 01:08:17

Yeah.

David Brown: 01:08:17

But yeah, it's this. I mean, like, this is why we have this show, right. Because we talk about this all the time. And it's why I started my show is, you know.

Ben Harvey: 01:08:24

Yeah.

David Brown: 01:08:25

We are going to have to figure out. We as a society, society and a world are going to have to figure out how we can use this and how we're going to get along afterwards and.

And what that means. And, you know, this is. We. We can talk about this forever.

Ben Harvey: 01:08:39

Yeah.

David Brown: 01:08:40

Which is why it's a great topic.

Alan King: 01:08:42

Well, I think we are probably, you know, however, coming to the end this particular episode talked about it forever, but there we are. So just a couple of things just to update listeners on the website for AI Org behind this podcast has been updated.

So please do go and check that aiororg.com we would love again to say, we would love to hear from any listeners, far flung places especially. That would be fantastic.

And the email for that is alanirororg.org so hopefully people can remember that and we look forward to hearing from you, but I think that's probably a wrap, guys.

Ben Harvey: 01:09:21

Yep.

David Brown: 01:09:21

Sounds good.

Alan King: 01:09:22

Until the next time. I'm sad to say next time we'll probably be back on our screens, won't we? But it's been.

David Brown: 01:09:26

And a special thanks to Ben.

Alan King: 01:09:28

Yeah.

David Brown: 01:09:28

For actually coming and bringing all of his kit and lighting and everything else so that we could do this in person. So thank you very much. Anybody in the west of England needs help with video stuff, go to Ben and if you're in the east, come to me.

Ben Harvey: 01:09:45

We'll cut it in half.

David Brown: 01:09:45

Yeah, yeah, we'll both do it.

Alan King: 01:09:47

Definitely don't come to me because I don't know what any of these cameras are doing, but they look fantastic. So.

David Brown: 01:09:52

All right, thanks very much.

Alan King: 01:09:53

Thank you.

David Brown: 01:11:57

Sa.