AI is not a silver bullet, but it's got a role to play.
&w=64&q=75)
James Harrington
Hello, everyone, and welcome back to The MAJOR Difference, a podcast by MAJOR, an agency specialising in digital engineering, design and brand experience. We're all about making things better than they were before every day. And this season, we're diving deep into the theme of positive impact.
Today, we've got a very special treat, we're joined by Antony Mayfield, who's the CEO of Brilliant Noise, an agency we've long admired and have recently decided to go all in on AI.
Fresh from DLD, Europe's largest AI conference, Anthony's here to hopefully share some cutting edge insights with us. Antony, it's a real privilege to have you with us today. How you doing? And how was the conference?
Antony Mayfield
I'm doing very well, thank you for having me on the podcast, it's lovely to see you both. Can't wait to can't wait to talk all about AI.
James Harrington
It's good to have you here. The both refers to myself and Martin McDermott, our co-founder who's sitting patiently alongside me keeping me on track.
Just briefly, Antony, when I first introduced you, I mentioned that Brilliant Noise, you're going all in on AI?
Specifically, what does that actually mean? How are you leveraging the technology with clients? And how can you help people adopt AI in their way of working?
Antony Mayfield
I'll try and keep it brief, because we've got a lot to talk about. But we are a marketing transformation agency. So what we've done for the last 10 plus years, is help big brands like Adidas, Nike, Asahi, Sanofi with how they're getting to grips with technology.
Essentially, we're big believers in human creativity and innovation, and using tech tools to leverage that, using tech tools to get the most out of there. So we've been looking at AI for a couple of years now, I think, I think it was probably 2020/21, was the first time that we put it in a client briefing, where we were saying this is something that's going to come along and change things a lot.
And now since the launch of ChatGPT has really unleashed a huge amount of energy, and innovation. Yeah, I tend to say that we're all in on it. And what I mean by that is we've got a set of tools, which are about capability around strategy, around running experiments, and things like that to help companies change. And all of that is just as important as it ever was, perhaps more so and more urgently important. Get people using AI, we're going to be running team days for clients, essentially hack days where people can get their hands on and get rid of some of the drudgery in their jobs with it so they can get used to it.
We're going to be running workshops with leadership teams around what it means for their sector, what it means for their business, where to be looking in, in the short amount of time that we can see forward and how to prepare for the future. So yeah, digital transformation, which has been around for last 10-15 years, it also needs to exponentially increase within organisations. It makes everything faster, more urgent.
But then, as well as as creative individuals, as strategists, we love working with creativity, we love working on breakthrough creative programmes with clients. These tools are amazing, just right now for developing ideas for being able to suddenly see 10 different options for anything that you're doing at any one moment. For doing mockups and mood boards with unbelievable quality really, really fast.
I sometimes think that the last iteration or the last generation of artificial intelligence was machine learning. It was all about data. It was all about big datasets and predicting and logic and generative AI is is much more creative. It will affect lots and lots of things other than creative industries. But I feel like when creative minds meet this technology, they can unleash an awful lot of potential. I think the most desirable job during the machine learning age was to be a quant. It was to be some, you know, like in Moneyball, Jones Jones character in Moneyball was that he was the quant made it all possible? I think creators are the new quants when it comes to AI. And I think anybody's creativity can be unleashed to the be the cliched word. If we just get our hands on with it and start playing and understanding where this can take us.
James Harrington
So Anthony, could you tell us a bit more about the conference? We're excited to hear your thoughts on AI and it'd be good to jump in and see what your views are.
Antony Mayfield
Yeah, so DLD stands for Digital Life Design, although the meaning is now lost in time. It's been running for about 15 years in Munich. And I think every January they have a big, big conference where they get together a lot of American VCs technologists think Sam Altman was there last year from Open AI. They've had Mark Zuckerberg, they've had all sorts, and it's brilliant.
But they had a special kind of breakout conference just a couple of weeks ago, focusing exclusively on artificial intelligence and what was going on and it was utterly fascinating. It's fascinating because I think it's the first time that I've been to a conference where people have had maybe 10 months since ChatGPT launched. And I think, you know, universally have changed the game when it comes to artificial intelligence. And because there were a cross section, it wasn't a marketing conference, although they were marketers there, there were political leaders, there were board members from SAP, from IBM, and other companies trying to provide services in this space, as well as NGOs, like the UN World Food Programme, someone from their accelerator was there, there were other people who are supporting entrepreneurs and social businesses.
And there was a strong connection with the theme of circular design and sustainability as well, that actually held a conference on the previous day. And there was quite a strong overlap. So we had actually, one of the opening speeches was from Sandrine from the Club of Rome, which is a big think tank, about just the kind of state we're in as a planet, as societies, and what the lengths that we'll need to go to in order to keep civilization going, keep the species going, as she pointed out, the planet will be fine.
And she was also quite clear about what artificial intelligence could and couldn't do for that. It's not a silver bullet, but it's got a role to play a thing was the the message from her. So yeah, it was it was a really interesting cross section. And it was a great space to reflect and hear lots of different perspectives from people who are having to make decisions right now with their budgets, with their strategy, with their election strategy, with where they're investing, about what's going on in the space. Yeah.
James Harrington
And I was looking at some of the videos on YouTube, and you've got a massive range of perspectives, as you said, from policymakers through to investors all motivated by different angles, but discussing the same core theme.
Before we get too deep into the specifics. Is it worth discussing the current landscape of AI? From our perspective, AI seems to be shaping industries and experiences faster than we can keep up. Could you maybe help our listeners to understand where we're at with AI right now, you've got any immediate insights you picked up at DLD that you think our listeners would want to know about?
Antony Mayfield
Yeah, that is a good, that's a good place to start a bit of perspective. Things are rushing along very, very quickly in this space at the moment. But at the same time, it's important to get some perspective around what's going on.
So first of all, I think the conference, one of the things that we heard quite clearly was the growth of AI, especially generative AI at the moment is fantastic. Think GPT has had something like a billion users in the last year. But very quickly, we move from just like "wow, scale, hugeness" to this raises difficult questions, all sorts of difficult questions, not just about what you know, we're going to do within business.
I mean, if you're a kind of a tech bro in Silicon Valley, then it's all gravy, right? I mean, it's just oh, this is the next thing. This is an excellent it's going to make another bunch of billionaire. For policymakers, for people running NGOs, for ordinary people working day-in day-out, this feels threatening, it feels like it's happening too fast. They don't really understand it. It's hard. It's hard to understand, in some ways.
So I mean, one of the themes that really we picked up on was the idea that it was scary and exciting at the same time. I think a number of people said this in different ways. But I think someone quoted Tim O'Reilly, who is the commentator who came up with the phrase, Web 2.0, which was social media kids if you weren't there. And so he called it this time around, he said, with artificial intelligence, he's afraid and excited about the same things. So, that's interesting.
James Harrington
You can go back to the 60s and 70s. And Marvin Minsky said, and I quote, "once the computers get control, we might never get it back. We would survive at their sufferance. If we're lucky, they might decide to keep us as pets".
With that in mind, how relevant do you think that is today? As we said, there's a lot of talk around fairly dystopian outcomes around AI. Are we facing an existential threat?
Antony Mayfield
I mean, it's I think it's going to large that perspective. There are there are big important questions to be asked. But those things aren't necessarily going to help you right now. Unless it's kind of I must do something. What you mustn't do is be paralysed by fear and think I need to stay away from this thing because you're becoming the pet of an AI is a long way off, it will happen at all.
Marvin Minsky said a lot of things a lot, and a lot of really useful things as well. But this is not it. This is not the moment where we will become slaves of artificial intelligence.
James Harrington
There is a lot of fear out there, even amongst the senior thought leaders who've developed, I'm not trying to be a fear monger here, but you know, you've got people like Suleyman Mustafa, ex-DeepMind CEO who's written a fairly impactful book, 'The Coming Wave'.
Antony Mayfield
I think you can't really turn on a podcast, except, I'm afraid this one, I mean, you got me instead of Mr. Suleyman, we all aim higher. But you can't really turn on a podcast or a TV programme or open a quality newspaper, if you're still in the analogue world.
James Harrington
Do they still exist?
Antony Mayfield
They do. You read the online version...
James Harrington
The ones written by ChatGPT.
Antony Mayfield
Apparently. Without seeing Mustafa Suleyman both telling you not to worry, and telling you how much you definitely need to worry. And I'm sure none of this is in any way connected with his needing to sell a book, or promote his latest AI company, Pi, which is very interesting. But he has legitimate concerns, obviously.
I think there was a really interesting roundup at the weekend of books about artificial intelligence, leading with Mustafa Suleyman's 'The Coming Wave' in the New Scientist, and a couple of the other books that they included.
One was by David Runciman, who have got a lot of time for a professor of politics at Cambridge. Absolutely beautiful writer, beautiful speaker, deep thinker about politics and power. And what what actually happens.
And Ron Simmons book, I think, just it's called 'The Handover', which is about the fear of a sort of handover of our species to an intelligence or an entity that we've created with all these superpowers that may not be self aware, but we're kind of just giving things up. And what are we doing? Yeah, but his point is that, kind of, as political theorist and a bit of a historian, I suppose, is that this fear has been around for a long time. And we've already created entities that are bigger than us. And that act... we've almost given them superhuman powers, even if they're not intelligent. They're called corporations and states. So you know, if we want to get big and deep, we might say we're already slaves to corporations and states. And I, for one, welcome our new AI overlords. It's the sort of thing that you'll say I'll say in regret forever. But that's what I always do. So I'm just going to say that I think there are several points during the last three or four years in the United Kingdom, where self aware, super intelligent robots might have been better than the government that we had.
James Harrington
I can think of a period this time 12 months ago, where that was very relevant.
Antony Mayfield
Yeah, I mean, a monkey throwing darts at a bunch of decisions on a wall would have been better than than that particular government, which was actually a version... it's not the politics podcast, is it? Sorry, but all of these things do overlap.
One of the things that was like a, you know, because we're in our marketing bubble, it's our domain of expertise, where we're looking at this, we're like, how will AI help us sell more ads, so change people's minds, and then then we get worried about it, changing our minds.
But one of the things that was really clear and rude awakening to me is what this means in terms of geopolitics, in terms of power, that was discussed that how AI has been an is used in the Ukraine/Russia conflict. How it's been used, and is used to spot terrorist activity. And the fears of the thing, his was a very European audience at DLD in Munich, the fears about the fact that 70% of the foundations of generative artificial intelligence, the large models, the foundation models are American, and 15% of them are Chinese, and everybody else is trying to catch up.
Yeah, not many of them are European. What does that mean for, they started talking about things like sovereignty? What does that mean for us having control of our own destinies? What does that mean for us having control of our own security?
And we know that these things aren't pie in the sky flighty concerns because we've seen what happened with social media. We've seen what happened with surveillance and with surveillance capitalism, which is, which is a very useful concept.
And in fact, if you want to go deep on on how this will play out with the money and the corporations in Silicon Valley, you could do worse than reading Shoshana Zuboff's 'Surveillance Capitalism', which is basically a playbook in retrospect for how technologies take over our lives and how they they already influence us.
This is interesting again, we're worried about artificial intelligence. There's, quite shocked, actually to to see how much this has changed, but there is an epidemic of, of eating disorders now. Which was an issue when we were young. People were worried about young people, especially young women with anorexia, bulimia, disordered eating. It's an epidemic scale now in societies that have been using social media the longest, no one's gone to jail. Nobody's taking Facebook off their kids. I feel guilty. I didn't take it off mine. Same with TikTok. There's, we can worry about robot overlords. But maybe there are, maybe there are other overlords we need to be worried about right now.
James Harrington
I'm conscious that I framed this initial discussion, on a more pessimistic viewpoint, if we switch to the point of excitement around the future of AI. Clearly, it's not just about cool robots and gimmicky automation. It's about being able to enhance human life, and help solve some of the bigger problems facing our planet and the growing population.
With that in mind, what do you think's in the pipeline for AI? In the immediate future, the immediate kind of three to five years?
Antony Mayfield
[laughs]
James Harrington
Are there any surprises we should brace ourselves for? Or should I reframe that as six months or three hours?
Antony Mayfield
No, I'm sorry to laugh.
James Harrington
It's a fair point.
Antony Mayfield
One of the first presentations I saw in Munich was from 468 Capital, some venture capitalists, and they were doing the last year. And really, they meant the last 10 months.
James Harrington
But they were talking about the hockey curve.
Antony Mayfield
Yeah, they were talking about the growth curve of a billion users in a year just on ChatGPT is insane. And that's the most we know.
But they were talking about the the adoption curve, but also the impact is sharper and faster, even than all of the other exponential revolutions that we've lived through in our lives to do with technology. They will like usually, you'd look at these things as a five year block, what's happened in the last five years? What trends can we spot for the next five years? But it's completely legitimate to look at the last year and treat it as like four years. And think through the phases.
And actually, when I cast my mind back, and actually when I look back over my notes, you know, fortunate enough that I take notes in Notion, so I can see when their date stamped, and I can see my newsletters when they're date stamped. So it's really interesting to see the ways in which we were talking about artificial intelligence 12 months ago. And it has changed dramatically.
I had two colleagues who took a sabbatical, this summer of four to six weeks each, I think, you know, long holiday, both of them felt like they were massively out of the loop on the technology when they came back. And they both have postgraduate degrees, they're both deeply interested in computing. They've read deeply on this subject area before. So yeah, things are moving at a fast pace.
James Harrington
Just quickly on that point that the session you referred to is a chap called me, let me try to pronounce his name, Ludwig N. Stahler, from 486 Capital, I listened to this earlier, and it is, there's a few notable points in there. The one that really made me laugh was when he was presenting, he said, I'm a mathematician, by training, I've got a background in mathematics, I wouldn't recommend anyone go into that profession in the future.
And to me, that really resonated, you know, that concept of, in the very near future, the intelligence level, and the sophistication, the ability of the AI that we're generating now is going to far supersede the capabilities of humans. And we're at that tipping point now. But we'll put a link into the show notes to that presentation. That's fascinating.
Antony Mayfield
I'd forgotten that he made that point. But it's an interesting one could form a whole other debate.
James Harrington
He made quite a few really interesting points. But what was interesting was that because he's at the forefront and thinking two or three steps ahead from an investment perspective, he was almost blasé about some of the talking points.
And they were concepts now talking about digital twins, and yeah, personal assistants. And he was just assuming that the technology is there, it's going to happen. I'm thinking about the bigger picture. He's talking about moats around organisations and how they protect their ability to remain competitive. And that shift from it being the companies that who are going to extract the most value out of AI.
And the current thinking is is not going to be the people developing the algorithms and the applications. It's the, at the moment, it's the silicon manufacturers because they hold the compute power. And that, to me was fascinating.
Antony Mayfield
Well, I think what was interesting in where the conversation went from there as well, because he'd set the stage for... set a strong frame for the day, was that both the venture capitalists and other people are not sure at all where the money is going to come from?
Definitely from the silicon in the same way as Cisco benefited hugely from the internet boom, because it was selling routers to everybody left, right and centre. We've definitely got a few years of in Nvidia being more than okay.
It's really interesting to see we're at a moment that will, I think we'll see, we'll talk about as maybe - we never really know what we're going to call things - the generative AI moment. You know, this is a moment where something very dramatic happened, where potentially you've got 5 billion people in the world who are able to access a technology for free, that lets you think faster, and more imaginatively, and you don't need to be able to code, and you don't need to be able to understand data, you just need to be able to string a sentence, you just need to be literate.
That's a pretty low bar, and nothing like that's ever happened. So that's, that's amazing. It won't continue like that for long. But what we've got at the moment is everything's kind of free or this slightly free access to these tools. We don't have digital twins, yet. But like, you know, he was talking about this is where you kind of get into speculation and whether people are uncomfortable, kind of tropes or predictions of what's going to happen. But he was talking about the movie, 'Her', starring Joaquin Phoenix, as somebody who downloads a computer programme with Scarlett Johansson, his voice that then basically runs his life for him. And naturally, they fall in love, etc, etc. But the point was, we'll all have this, kind of like, the dream, Cortana, Siri, whatever the hell the Google one's called. Google, is it? I don't know.
James Harrington
No idea.
Antony Mayfield
Yeah, exactly.
[Siri can be heard in the background]
Oh, my life.
James Harrington
Well, there you go. Just let me pause there. That is a classic example of the power of technology. As we were talking, Antony's phone has literally been, his phone and his watch, have been listening to what we're saying and thought they were asked to respond. They've just responded in the middle of the podcast. Perfect example.
Antony Mayfield
Yeah. So imagine that, except it works. I think that's what he's imagining. And at the moment, you can get some of those effects yourself by stringing together workflows.
So this morning, I was putting together a copy for our website. You guys, oh, you guys run an agency. What a pain in the backside that is. You know, 'cobbler's children'.
I once came across an agency, where somebody candidly told me, it was a big, big, big design agency, a million pounds they spent on their website, when they taught it up all the hours that people are dedicated to it, because everybody knows best, and also everybody's filled with self doubt. And you ask anybody in the UX department what the copy should be, and they've got an opinion, and anybody in the copy department what the UX should be, and they've got an opinion.
So writing your own website is a nightmare. And I wanted to write a bit of copy for my website. And I went into a process where I basically outlined it in notes. And then I asked ChatGPT to come back to me with a marketing experts view of what a great proposition sounded like, then I use that as a prompt for it to then create some copy. And then I thought, oh, this is so tedious and boring, and like, what are the things that it's really good at, and what these generative AIs are really good at is repackaging content. So if you give them your content, then it'll repurpose it in all sorts of ways.
So I thought, well, who are all the people that I would like to sound, like to hear what how they would treat this, this writing?
What you used to do in the Google age? Do you remember the Google age? Do you still use Google? Yeah. Anyway, in the Google age, you would, as an agency idiot, trying to write your own copy, you would go out and look at all the competitors websites go, "well, that's quite good", "I love how they've...", "I wish I could sound like...", "oh, they've got nothing at all, how bold".
So what I did was I took this very bland piece of copy, which you always end up writing about yourself, unfortunately. And I asked it to rewrite it in the style of about eight agencies that I liked like Pentagram - the design agency, IDEO, obviously MAJOR Digital - I was fascinated to see when they would come up with.
And then I started doing like writing pairs like our style guide used to be we had the tone of voice of Barack Obama and Grayson Perry so like the incredibly analytical and cool with [makes sound] colour, [makes sound] art, and that's because that's what we wanted to combine.
So I started putting those in and all I had to do was just change one word of the prompt to give me some examples of that. And it was brilliant. It was brilliant. Because I had 10, I had 10 different versions of my website copy, which I then started picking and choosing from started editing down, I put three or four of them together, I sent them out to a bunch of people in suits, or which ones they hated or like, and that was done. Inside two, three hours, that that stuff, which no one will ever admit to that writing the front page, copy for your website, it'd take you a week, wouldn't it?
James Harrington
You're talking about the power of effectively prompt engineering, aren't you?
So there's an element of rubbish-in rubbish-out with anything, I think more so with the use of some of these generative AI tools, where you have to, I think there's a shift now into becoming expert in the ability to write prompts, and actually structure prompts to get the right output.
Antony Mayfield
Well, so, yeah, I mean, that I think that's useful.
There is there's a really brilliant writer, who I would recommend everybody check out is Ethan Mollick, who I think is at the University of Pennsylvania, one of those, probably get that wrong.
James Harrington
We'll put a link in the show notes.
Antony Mayfield
Thank you.
He says, he's a university professor, he writes really eloquently about how you can use generative AI in our workflows. And one of the things that he said that I thought are really agreed with was that having domain expertise, and knowing how to use prompts is the absolute best combination, and will be a great prompt engineer every time.
Because again, yes, rubbish-in rubbish-out, but it's not just the words. I think he was saying, now is the time for grimoires, now is the time for spell books, you need to save up the quotes, the prompts that work and try and understand where they work and look after them and share them with colleagues. But I think someone else put it really well the other day, you can put a link in the show notes, 'Strategy Finishing School', well worth checking out really, really excellent. There's some free webinars as well as a subscription service.
But he was saying that AI will not replace strategists, which he is a strategist, so everyone thinks that's not going to replace them. But he said strategist that know how to use AI will replace strategists that don't. It's such a powerful tool for getting where you need to faster. It's not the answer to all your problems is to learn how to prompt engineer. But this is going to help you a lot to understand how it works? Why it works? And not just the technical side of it.
I think I put a lot of store by the idea of having a felt sense of how something works. You know, the experience of using it is so much more important than than just having the theory. I think we will inevitably at some point, say "chess", but chess is one of the first places where artificial intelligence happened, where it had an impact. And the one thing we know from when Garry Kasparov was beaten and then he started the idea of special chess or 'Centaur chess'... I know it's called Centaur chess, that's easy to remember. But was combinations of humans and AI is competing against each other. The perfect combination was two humans and a computer, I think.
Humans plus computers, beat all comers. Basically, they beat humans on their own, machines on their own.
James Harrington
I don't know why, nut my mind immediately went to the Olympics and the humans using steroids and having a super, super Olympics. No idea why my mind went there...
Antony Mayfield
Well, it's not a bad metaphor to try.
I mean, I think one of, the one of the, not quite as a philosophical level, but a level of like, how do we think about this? Because [argh] Skynet is not very useful. "It's going to replace us" is probably wrong. "Ah, it can replace all my staff", is probably also not going to work. We can get into the reasons why.
But what we need to understand is that we don't have, we don't have a way of talking about this yet. We don't have a frame for it that's really, really useful. So playing with analogies, and seeing how they work out is not a bad idea. I have not tried the athletes with steroids and, prosthetics. That's pretty good one.
Other ones that I've heard that have been really useful are like everyone gets 1000 interns, like you know, you've got everybody gets free assistants, who are incredibly bright and enthusiastic, but don't know anything about what you do. That's what working with an AI is like at the moment.
My favourite analogies, quite close to the athlete, one would be that it's like, they're like electric bikes for the mind. Like you can go further and faster, and you can go the big hills and they won't tire you out. All the things that were a slog before you can work out ways to do with ChatGPT.
That's my experience this morning with writing that website copy and feeling really confident about it. It was something that would just drain me for at least a day. It's done. It's off the table now. I'm happy with it. Yeah, all that energy can go into something else.
James Harrington
It throws out the kind of efficiency question doesn't it?
For me part of the problem with framing AI is its application and trying to get your head around it. Its application is so vast, from using it as a tool today to help us improve efficiency or reduce effort spent on doing mundane things, or trying to get to a solution quicker, through to scientific research, medical applications. People talking about the level of intelligence that AI is going to deliver in five years time, people won't necessarily need to see a doctor first, they may potentially get diagnosed by an AI as a first point of triage, and it will be more efficient, more repeatable, more expert.
Antony Mayfield
So, I think that it's, yeah, you're right, it's so big, it fries your brain. So you got to do a couple of things, you got to kind of come up with some frames, or some buckets or whatever, whatever metaphor you choose to think about it.
It's good to think at the macro level, it's good to think at the high level, it's fun to discuss. We don't have a lot of control over that we just have a vote. Right? We have the the actions of a one eight billionth of the species has an effect on on all of the rest. So what are we going to do with our one wild and precious life as the saying goes?
So drawing it back to the micro, then, you've got what does it mean for me? What does it mean for me today? And I think in between, we've got some stuff as well, which is like what happens in organisations what happens in in teams and groups of people?
James Harrington
I'm glad you phrased this, because I've got a list of questions that I was going go through and I've completely departed away from them. But one of them was how should businesses and individuals start thinking differently about AI? And well, how do they avoid a too late scenario?
Antony Mayfield
I think, I'll just like give you a very quick what I think is going to happen.
James Harrington
Yeah.
Antony Mayfield
What's going to happen is artificial intelligence, and I'm specifically talking about generative artificial intelligence here, the kind that we can all lay our hands on and do something incredible with right now.
James Harrington
So just quickly draw out some examples of that.. ChatGPT...
Antony Mayfield
Yeah, so ChatGPT. Perplexity is a great tool. I think it might run off ChatGPT, Anthropic, Po, all of these apps give you access to generative AI engines. Well, the technical term is Large Language Models. They're these very big systems, very powerful systems that can take a request from you and make sense of it and give you some things back that may be useful. And the better you get at asking them things, the more useful the things they give you back.
James Harrington
And at the moment, those things are typically answers, images, video, or audio.
Antony Mayfield
That's right. And you can dig into any area of that and we're beginning to explain it to ourselves really, really well. You should check out the Financial Times, it's visualisation or visual storytelling of how GPT works - like last week, if you're interested in that, it'll blow your mind. But it's a it's a very nice fireworks display to watch as your mind spreads.
So what's going to happen with generative AI, that particular kind that we can use right now, it's going to change a lot of things. But right now it's going to go through three phases. The one that we're just starting in at the moment, even though it's been around for 10 months, is where it will make things that we do every day, it will make them better, we'll be able to do them better.
The next thing that will happen after that is we'll have gained a new set of skills and a new way of working with computers, with technology. And we'll start to come up with better processes, better ways of working, just completely new workflows, new ways of getting those things done. So you thinking tasks, then you're thinking workflows.
And then after a while, and this won't happen at the same time for everyone or every organisation, it will reach the stage where, oh, actually, we just need to redesign the whole organisation around the new ways that we can do things.
So what does that mean? That means that right now the smartest thing that anyone can do is start playing with the technology and start just thinking about how they might use it to do things a little more easily. Roll that up to organisations exactly the same response, put it in the hands of as many employees as you can give them some guardrails and some guidance about how to use it. But you're going to get the best results from having as many literate people as possible, coming up with as many ideas as possible.
And I saw two camps really at the DLD and the same ones I've seen in the world. I think the enthusiasts for AI were probably more represented at that conference, as you'd expect. But there were people who were sceptical of its democratisation, I'd say they were like, doesn't, "not everybody needs to understand this", "not everybody needs to use this". Whereas a lot of people were like, "you need to get this in everyone's hands immediately".
And some real surprises there of people who were get this in everyone's hands immediately. So one of the examples which was offered by a speaker was Walmart, in the US. And now, like most of the audience, because they made a bit of a joke of it, my immediate thought was, "oh my gosh, they're just going to replace their employees with robots". Of course, that's what Walmart does, because it's a large, aggressive capitalist organisation. That's what they do.
No, what Walmart has done is, I think, given it to, and I've not found the reference article for this, "I'm going on hearsay m'lord", that they've given it to 1000s of employees, and shopfloor employees, people at all levels. And the way that they're thinking of it is, we need to give these tools to our employees, because it will remove drudgery from their everyday work, which will allow them to think of better ways to serve our customers and better ways to run the company. And I think whoever has been allowed to do this there has put their finger right on the difference of this technology to previous technologies, which is that it's a creative technology can help you be creative. And it makes that creativity a lot more accessible.
So they want as many ideas as possible. They want people to be thinking of new ways that they might use this. And they realise that they're not going to sit around and wait for large, big tech core to come along with the system of this is how you use it to make things better for your customers.
James Harrington
Looking ahead, it is absolutely fascinating, we've talked about what it is, we've talked about the scare mongering, we've talked about practical applications. What do you see the key development milestones upon us now? I know it's crystal ball gazing. But did you come away from the conference last week with any immediate next steps or milestones that you think are going to hit?
Antony Mayfield
Well, I think what I came away with was a sense that there's an urgent job to do of education of people across society. But I think that education and learning and finding out how it works is the number one task for organisations. Certainly for our clients, that's where we're telling them, you need to do this and do this now, get everyone connected, get everyone immersed in it, you will benefit from this.
In terms of the key next steps, well, there's a term that I've not heard before this particular tech revolution, or 'platform shift' as it's known in the trade? A platform shift is when a new technology comes out and it changes how you use all the other technology.
Yeah, so mobile was a platform shift. The web was a platform shift, it changed telephones, it changed TV. So this is a platform shift. We don't know how it's going to change music, telephones, all of the rest of it. Because we're only just starting to imagine it.
There's a term that I hadn't come across before. That's called 'technology overhang', which is Silicon Valley term, God bless them, which means that even if generative AI or artificial intelligence, even development stop right now, even if there were no more advances in it, then there's still five to 10 years of innovation and value about around what we do with it.
An analogy with social media would be someone's invented social media. But we haven't invented Twitter, we haven't invented Facebook, we haven't invented Yelp or Uber, all of these things that come out of that Web 2.0 revolution. So it's in everybody's hands now to, to come up with the new ideas come up with the new invention. So we'll see that we'll see new companies emerge.
I think one of the really interesting, and I'm not a technical person, I've just taken interest, but one of the really fascinating things that will affect all of us and affect big tech, especially, was that Facebook's LLaMA model, which for viewers at home is their version of ChatGPT was leaked earlier this year. And there's a leaked email from Google, which you can find fairly easily, it was widely reported on where the executives there were saying, "we have no competitive advantage". We have no moat to go back to that moats phrase, we've got nothing that something that people can't cross.
Like, Google has a moat with Search Engine Optimization, because its search engine is so good, its brand awareness is so high, and it has enormous technical infrastructure around the world that it can leverage. It's worried that that doesn't help it with AI. Because this model was leaked and immediately effectively became open source.
So there are people innovating with that model and there was a kind of a race to see how small you could make a large language model. I think somebody fitted it on to a couple of gigabytes, which is not much these days. So you could carry around an entire Large Language Model on your phone. That's the that's the technical thing.
So what people do with it next, goodness only knows. Embedding it into all sorts of things. There are reasons to be concerned and worried and fearful, there are consequences of that there are bad consequences as well. But fear is not a useful reaction. Because, I mean, this is the one this comes to the side of AI that we are more expert in and Brilliant Noise we've been working on for about 10 years is what happens when you put humans with machines. And what we know when a human brain is afraid, it goes into fight or flight mode. And in the moment, it does that. When it's generally afraid, it will retreat, it won't want to act, it won't be as imaginative, it won't be as creative. So we have to kind of get over the fear in order to start imagining ways to deal with the opportunities and any threats as well.
James Harrington
With that in mind, when we opened up, we talked about some of the more fearful aspects, there's a lot of conversation around regulation and containment. And there are topics we hear a lot of discourse about when AI is discussed, what are the what are the ethical and regulatory things that need to happen? In your opinion? Is regulation the way to go? Or if we already lost control?
Antony Mayfield
Such a complex question.
I think regulation is important, not as an answer, but we need regulation around it, around artificial intelligence, right now. One of the reasons which, I think, was raised by someone at IBM was that people are waiting for legislation to be clear before they can write their company policies.
Large corporations, most of our clients, large corporations, with at least a few billion turnover. Large corporations are usually subject to laws that constrain them in in certain areas. And so when they write a policy, they want to know what the law is, so that they can base it on that and react to that. When they when there's no law there, they don't know what to do and they're hesitant to, for instance, put AI in the hands of staff, because what happens if then the law comes in and makes them liable for the things that they've already done. So regulation is really important, because the lack of it impedes progress holds people back.
James Harrington
Part of the problem there, though, is that the experts in the room, the lawmakers and the policymakers aren't experts in the technology. So there's a piece of education has to happen to actually enable them to comprehend the parameters and the implications of what they're making laws or regulations around.
Antony Mayfield
Yeah, I mean, from the point of view of we need it so that we can have policies and that we can get on with things, it doesn't really matter. You just need something so you can get on with it.
Now there's a big kerfuffle about the EU legislation is likely to be first and it's likely to be not what Silicon Valley wants.
James Harrington
Was that discussed last week?
Antony Mayfield
Yeah, Eva Maydell, who's an MVP, one of the authors of the EU AI act, she was talking about the fact that that some of the people drafting the bill don't know, the technology don't understand the issues, so that trying to do their best to educate them. But she was also one of the most adamant speakers of the conference that everyone, especially leaders should be getting hands on with it now so that they can understand it and can make good decisions and good policy decisions.
I think going to the going to the point the you raise about containment, which I know is part of Mr. Mustapha Suleyman's points. Sounds like a good idea. Good luck. I mean, you know, if you regulate on privacy, then some countries will abide by it, some won't. If you legislate on copyright, some follow up more than others, as long as it's in the interest there economy. As I said, with the Facebook module and the leaking of their equivalent of ChatGPT, it's out there, people are innovating with it, engaging with it, remixing.
It's not to say that we can't do things to make it harder for people to do bad things with it and can't make the incentives for doing things that are evil with it. Evil is a strong word, but probably doing it...
James Harrington
Bad actors.
Antony Mayfield
Bad actors, you know, Richard Gere who do we mean by that 'bad actors', we mean, well, I mean, that's one of the things that was really clear in Munich, when we say bad actors, we mean, either criminals and terrorists within our own borders, or we mean other geopolitical actors who don't agree with us.
And I think the phrase the panda in the room was mentioned and Russia and China and North Korea and Iran. So you know, people we don't agree with. They're the bad actors, according to us, and we're the bad actors, according to them. And if we could just give the world a Coke, and everyone could get along, then it wouldn't be a problem.
James Harrington
That's the answer.
Antony Mayfield
Well, this is a problem not of artificial intelligence, this is a problem of brains. So in his book '1000 Brains', which is about the most amazing, popular science book that I think you can read at the moment. Theory of how intelligence is created in the brain, and what that means for how computers will develop.
Jeff Hawkins says that a lot of our problems are down to the fact that we've got two brains. We've got an old one and a new one. And they know there's more than that. But we've got a neocortex, we've got like the big, foldy, lumpy thing on top everyone thinks of as the brain. And underneath, we've got some animal layers. And the animal layers are the ones that make us jealous, and irrational and angry and, and do things like start wars, or lie or be cruel to small animals. And if we could just think about it intellectually, if we could just find ways to get on and agree and emotions weren't a part of it, then maybe we have a shot at banning AI. But, ironically, AI might be the only way that we that we get that.
James Harrington
And in fairness, I think a lot of the a lot of the dialogue is around taking a pause. So the open letter that was written in and signed by a billion people in May was around...
Antony Mayfield
Really smart people.
James Harrington
Yeah, and I think their argument was...
Antony Mayfield
Who still don't understand it.
James Harrington
Well, I think their argument was, we need to pause the training of some of the more advanced large language and potentially AGI models, we need to pause training, until we've got safeguards in place and an agreement around how we manage the use of AI.
Antony Mayfield
It wasn't really clear on who they were asking. I mean, it was just like, "oh, watch out for the stuff we're doing. Someone stop us. Someone hit the pause button."
James Harrington
I think their theory is there's an arms race going on.
Antony Mayfield
Well, there is an arms race going on. It is a weapon. It is a it is a very, very powerful tool that can be turned into all sorts. I mean, it's not a powerful tool. It's a foundry. It's a forge. It's a new alloy, that we've suddenly got that we can do a million things with. And everyone's got it at once.
So it can be a weapon, it can be a teacher, it can be a war, it can be a banker, it can solve the world's problems, or it can exacerbate all of them. It's just there. And I'm sure some of them are talking with with, I mean, I suppose they're talking to America really, because if 70% of models are coming from there, then that's sort of what they're talking to.
But it is also if you read, Shoshana Zuboff, again, 'Surveillance Capitalism' it is also a move straight out of the surveillance capitalism playbook, which is that you agree emphatically with the idea that you should be regulated, even as you know that states are too slow and won't do anything. And you sit there going, "please regulate".
I mean, Mark Zuckerberg in Congress, "please regulate us." Well, "you know, I just don't know how all of these these terrible effects of radicalization and disordered eating and division and political polarity, which being exacerbated by our algorithms, we need to do something about this. I mean, I'm just getting richer by the second. But could you do something about this?"
And meanwhile, the inventor of DeepMind, the inventor of xAI, and these, say, "yeah, hit the pause button, guys. Anyway, I'm still turning up to work on Monday and I'm not turning down the next VC round, and I'm still going to the White House to talk about it. I really want to be a part of this conversation, while I define what this conversation is. And while I de facto take over another public domain".
I'm sounding a lot like I'm from Navarro media. You know, I actually, I'm a capitalist and run a capitalist enterprise. Yeah. So I'd like things not to be terrible.
James Harrington
Final question, then. Is it possible if this is a borderless issue? Is it possible for the world to come together?
Antony Mayfield
Such a good question. Because is it a borderless? I mean, it's a borderless issue in the effects the whole world. It's a bordered issue, in that those technologies and those servers sit in certain countries.
James Harrington
But they were always well, they, you know, if quantum computing is really around the corner...
Antony Mayfield
Oh, what? No. You can't ambush me with quantum computers halfway through the podcast.
James Harrington
If it is..
Antony Mayfield
Yeah, well...
James Harrington
That opens, I mean, the convergence of that, and that is the... maybe that's the subject for another podcast, but the convergence of AI, quantum computing...
Antony Mayfield
It took me a year to get my head around AI, you invite me in a room, and then you start wanting me to understand quantum computing. I mean, you know,
James Harrington
The outcome of that is potentially that the processing power could grow exponentially. What you fit on a silicon wafer today, could be a billion times more powerful. Once that's in play, and we've got these models, applications, it takes the CPU the compute layer out of the equation, and it might not be developed in the Western world.
Back to the point around it being a borderless issue, for me personally, I think it is and that's why I think the argument for regulation exists we have there has to be a way to coordinate that. Otherwise, we're all doomed.
Antony Mayfield
Well, we're all doomed then because there isn't a way to coordinate that. And there won't be. You know, we've got the United Nations that's as good as it gets. And that's, you know, has America started paying its subscription to the United Nations recently, I can't remember exactly.
James Harrington
I'll tell you what we'll do. It's September, let's come back in a year, let's put a diary note in.
Antony Mayfield
If we get quantum computing breakthrough, I will come back to this moment in time and tell you exactly what happens.
James Harrington
Via a wormhole.
Antony Mayfield
You know, now, this would be the point where it cuts to in the film where I'm actually standing in it. "No, don't do it". I'm twitching the thing. I'm trying to get the dust to fall in patterns.
James Harrington
This has been really interesting.
Antony Mayfield
[laughs] Remove him from the room...
James Harrington
No, I've actually got a completely left field question that we're asking everyone we're interviewing as part of this series. And that is if I gave you control of all public media, so ad networks, billboards, TV ads, radio ads, if I gave you complete control now in the UK, for the next day, what would you say and why? Now I've ambushed you.
Antony Mayfield
Right? Actually, yeah, quantum computing, let's go back to quantum computing.
Now, what would I say? Oh my life. Ban advertising? Put that on every billboard. And I'd be interested if anyone noticed.
James Harrington
There you go. And before we say goodbye, are there any resources that you'd recommend our listeners dive into? I'm aware of your Substack. Could you tell us more about that and any other resources you think are a good to listen to or read?
Antony Mayfield
Sure. My Substack is 'antonym', antonym.substack.com. And we've got the Brilliant Noise Substack as well, which is a bit more of, a for marketing leaders. One says bnedition.substack.com.
In terms of resources, I cannot recommend Ethan Mollick's newsletter enough to anybody who's thinking about this. There are a lot of great writers on this. I would say that the Financial Times is worth the very high subscription price for their AI coverage alone and it's grounded in reality as well.
Those are two resources I would mention. And I think that there are a lot of free resources coming out. Google's doing quite a lot of work on this as well to help people get to grips with with generative AI.
But the best tutor in in the world actually is probably going to be a generative AI. And if you learn a little bit about how to use them, they can start to teach you things as well. It's really incredible. So I think the best resource I would say is get yourself on ChatGPT. If you can afford to pay the subscription for a month or two to to learn it on the on the pro level.
And I'll throw out a plug for my two favourite apps. If you could only have one app, it would be Po, which is an app brought out by Quora - the the question and answer company - and that gives you access, for the same price as ChatGPT, to not just ChatGPT but whole load of other models, and lets you build bots as it calls them, which are basically little versions of the AI that you've primed to talk in a certain way or address a topic in a certain way. There's an example of that I've got one on mine called 'Orwellbot', which I put in George Orwell's six rules of good writing and I'll run copy that I've written through it, and just see where I'm using cliches or where it could be a bit tighter and all those sorts of things. Great thing to have fun with.
James Harrington
Anthony, it's been fantastic talking with us today. Thank you for joining us and shedding some light on the subject of AI. Clearly, we've got a lot of catching up to do. But I think you may well have swayed my opinion in favour of further development of AI and how we use it responsibly.
So with that in mind, and to all our listeners. Thanks for tuning into The MAJOR Difference. It's been great having you with us. And until next time, stay curious.
