Rob’s comments below are in italics.Derek’s comments below are in normal font.

Derek and I know each other through an American consultant called Perry Marshall. Perry was originally a marketing guy but has transitioned into an all-around business expert. You attended one of his seminars this week, specifically on AI. We thought we would unpack what you had learned and how we, and anyone listening, might apply that.

Yeah, absolutely. It was a really inspiring, really interesting day, with a strong focus on AI. Frankly, I’ve been very cautious about the whole AI thing and somewhat sceptical of its value.

We use it a little bit, don’t we? We use Claude to clean up the recordings and generate transcripts. I’ve been very impressed by how that goes and how much better it is than anything on offer until a few years ago.

I’ll add some comments on that. When we started the podcast a few years ago, getting the transcript into a readable version took a lot of work. These days, it’s literally 20 minutes before I go to parkrun on a Saturday, courtesy of the tools available.

Yeah. I then give it a second check. Other than where I’ve expressed things poorly, all it’s required from me are minor spelling and punctuation changes and a bit of reformatting. I’ve had less and less to do, more or less week by week.

One of the reasons I was sceptical about AI is that it’s yet another technological innovation which has produced a disproportionate amount of hype. A lot of people who’ve been swept up in that hype are going to be disillusioned, because it’s unclear to many what it is and what it isn’t.

There are two levels of hype. One is “this is going to transform humanity,” but nobody seems to be suggesting how. “This is the future of great prosperity,” but nobody seems to be saying in practical terms what that prosperity will actually look like.

And who for?

Who’s going to get the benefit? Exactly.

There’s also a philosophical dimension to the hype. There’s a lot of speculation about whether AI will become fully conscious, whether it’s going to overtake the human race.

I’m absolutely clear that whatever it is, it’s an algorithmic system running on computers, and that is never going to be conscious. That’s a very big philosophical debate, but I’m crystal clear in the basis I have for being certain of that.

Did you see that Richard Dawkins got ripped to shreds for saying he thought Claude was conscious?

That was so funny. I cannot imagine how he thought he would enhance his credibility by putting that out. Don’t get me started on Dawkins. It’s very much on the side of people who are still trying to say the universe is a great machine, that we’re all bits of machinery, and there’s nothing transcendental. I had a brief word with Perry about the Turing test. Do you know about it?

I’ve heard of it, but I couldn’t tell you what it was.

Alan Turing, the guy who, to all intents and purposes, created the underlying intellectual framework for computing, was also a major player in breaking the Enigma code in the Second World War, a major contribution to defeating the Germans.

The Turing test was Turing’s contribution to the debate about whether a computer could be conscious. The test is basically this: if you could have a dialogue between a person and a computer and the person couldn’t determine whether the entity they were conversing with was human or a machine, then the computer would have passed the threshold of consciousness. According to that criterion, we’ve actually passed that point, and Perry and I agreed on that.

I got the output of Perry’s group, what he calls the Superconductor, a sort of AI business consultant. You do a quite challenging questionnaire and get a report back. The report blew my socks off. I could not tell that it hadn’t been written by a more-than-averagely competent, insightful business consultant.

Yeah, and this is going to completely derail a lot of middle-class jobs for people doing this kind of work.

We’ve seen this progressively, haven’t we? The point is, it still isn’t conscious. Broadly, the main reason is summed up in Roger Penrose’s book, The Emperor’s New Mind. Did you ever read it? It’s absolutely spot on. In summary, what Penrose says is that because of Gödel’s theorem, that’s Kurt Gödel, it’s impossible to create a formal system for mathematics that is both complete and consistent.

What it comes down to is that there will always be mathematical statements which are true but not provable. Gödel showed this by constructing one. At the same time, he showed that the human mind was capable of convincing itself of the truth of that, which a computer could never do. There’s an argument sufficiently obscure that many people who are disinclined to accept that point of view simply don’t get it.

They say, “Who cares whether a computer can prove Gödel’s theorem or not?” Anyway, I digress more than I needed to. But it explains why I’ve been put off by speculations about whether AI could be conscious.

It can certainly be intelligent. If you define intelligence as the ability to extract a simple summary from a complex argument, for instance, you can bring up a whole bunch of definitions. But that’s not the same thing as consciousness. So, my big takeaway from Wednesday’s discussions was that this is yet another tool.

The thing about any tool is that you can use it for good or ill. That brings up another whole aspect of the overhyped discussion. Is AI a force for good or a force for evil? The answer is neither. In the hands of someone motivated to do evil, it can enhance their ability to do evil, which is unfortunate but a fact of life, just like any other technology.

You can use them for good or for evil.

Yeah. That was another very interesting thing, because as anyone who knows Perry will know, he is essentially as strongly grounded in ethical principles as you can be.

There were two special guests at the seminar. One was Scott Chang, the guy who developed the Superconductor programme. It was great to meet him and converse with him personally. I’ll be taking up the offer he gave to seminar attendees, an hour’s consultation. After that, I’ll apply it to the project we’ve got to get this book out, which has been a long time coming.

He was there, and there was also another guy called Finn, who founded a business called Solar Quotes for selecting solar power installations. It’s essentially a lead generation website for contractors. He recently sold the business for multiple millions and is now enjoying his retirement.

What he said was that the biggest factor in his success was running the business not to maximise short-term profit, but to build something of lasting benefit that provided the best possible deal for customers. He structured it so that was genuinely the case.

They had one of those rock-solid guarantees that could bankrupt you if you weren’t doing things the right way. If any installer that a website client had booked gave an unsatisfactory job, they would, at their own expense, rectify it to the customer’s satisfaction. You can’t get a more ballsy guarantee than that.

I asked him what the philosophical basis for that commitment was, and he said, “It’s just the golden rule.” This touches on what we’ve said before: the strongest thing you can do to build a successful business is to operate according to that principle. It’s depressing how few people in business appear to have that as their overriding priority.

Scott Chang, had been a mortgage broker and was one of the few not to go completely out of business overnight when the 2008 financial crisis hit. Mortgage broking is famously another business area operated by complete sharks and charlatans, but Scott operated according to the same principle. It would have been very easy in the early 2000s to make a great deal of money getting people into mortgages that turned out to be immensely profitable for the lender and ruinous to the customer.

The devil will always place short-term opportunities in front of you to tempt you away from the longer-term goals…

Of course, you don’t have to be a Christian to hold that principle. Any ancient wisdom tells you the same thing, whether it’s Taoist, Buddhist, Hindu or whatever.

So it was quite interesting from that point of view. It’s a creativity accelerator, if you like. It’s only as good as what you put into it.

This has been the thing for a long time, hasn’t it? Garbage in, garbage out, or good input in, good output out. The work we do for this show, in preparing the written versions, generally involves good input.

Yeah. I’m looking at that in a whole new way. Scott gave me one or two pointers about how to approach the project we have on constructing this book. It’s more or less what we had been thinking of doing, but I had been struggling because of not being quite clear about how to go about it..

Incidentally, I put one chapter I manually cleaned up into the shared document you created for this purpose, so you can have a look and tell me how close I’m getting. With the insights I’ve now gained, I’m going to take that approach and feed material in, with instructions, to one or the other of the AI tools.

I’ve got ChatGPT on the £20 a month subscription. You’ve got Claude. We can see how one or the other works, and if there’s any way to find synergy between the two, that’ll be an interesting project in itself. I’m quite upbeat about exploring this going forward.

I’m now seeing it in very much the same way that I see the CAD system I use for designing electronic circuitry. I use KiCad. I downloaded and installed i...