78. Fry your brain: Quantum computing, AI and marketing

PLEASE NOTE: This transcript has been created using fireflies.ai – a transcription service. It has not been edited by a human and therefore may contain mistakes.  

 
00:00 
Dom Hawes 
I'm taking you into the future. We're going to explore a subject that, at face value, you might think a little bit tangential today's marketers. Well, today that might be true, but tomorrow isn't that far away. Time is accelerating. Think back to the 29 November 2022. That was the day before OpenAI launched chat, GPT and the capability that we now all take for granted. It seemed so far off to most of us on the 28 November 2022. So stick with me today, I'm taking you into the future to explore the implications on marketing as we know it of quantum computing, artificial general intelligence and artificial superintelligence. Now, there's a health warning that comes with this. Some of what we're talking about is scary. If you're of a gentle disposition, you may need a change of underwear. Don't go away.  

 
00:58 
Dom Hawes 
You're listening to unicorny and I'm your host, Dom Hawes. Once upon a time, artificial intelligence was something we saw at the cinema in Sci-Fi or we read about it in the papers when a computer toppled the latest chess master. Now, Skynet, for example, was an artificial neural network and super intelligent AI. In the movie Terminator, Skynet gained self awareness, so humans tried to deactivate it. Skynet responded by launching thermonuclear warfare. Well, who wouldn't do that? In the movie, Connor is our hero, and Arnold Schwarzenegger played the antihero, a machine sent back in time to terminate Connor. It's all very dystopian. And these days, of course, AI, it's on our desktop, it's on our phones, it's all around us, and it's getting smarter day by day. So as marketers, we are leaping into the AI abyss. Many of us seem blind to its imperfections.  

 
01:52 
Dom Hawes 
In the ever changing business landscape of today, it isn't just essential to understand aih. It's a business essential. Well, I'm being a bit naughty here.  

 
02:01 
Dom Hawes 
You either get that or you don't.  

 
02:02 
Dom Hawes 
I hope you did. But if you didn't, answer's on the show notes at unicorny dot co dot Uk. So who's going to be our Connor? Well, I looked towards Mark Ritson to do that. He is the saviour of market orientation and also just so much general common sense when it comes to the guff uttered by many marketers. But he's sold out. He's with the machines. So who the hell is going to save us? Steven Millman, that's who. And while he hasn't exactly flown back in time, more like flown from the US to London. He is here, and he is the best brain I know in the business when it comes to getting to grips with the technology stuff that is accelerating towards us. So today, we are looking at the implications of the next technology revolution.  

 
02:41 
Dom Hawes 
Quantum computing, how it's the key to artificial general intelligence and ultimately artificial superintelligence. And, of course, what all of this means for us marketers. And it does have meaning. Deep meaning. I probably won't live to see it. Thanks, Skynet, but you might. Let's go meet Steven.  

 
03:02 
Dom Hawes 
It's been quite a journey since I last saw you.  

 
03:05 
Dom Hawes 
What's been going on?  

 
03:05 
Dom Hawes 
Your company's been winning awards. You've been winning awards. It's like an award waterfall.  

 
03:10 
Steven Millman 
About five weeks ago, I won the Market Research Council's hall of Fame Changemaker award, which is given to the most impactful market researcher of the last year or two. Personally, I think I can probably name a dozen people who deserve it more, but I'm not turning it back and saying they should give it to anyone else. And then, just last week in Spain, Dyneta won the icom Data Creativity award for our submission and data quality for quality score, which is what Dynata uses to weed out not just bots and survey fraud, but also to identify actual respondents who are just too disengaged to have their data used. It's AI based, it's 175 characteristics. And not only did we win our category, but were first among all 27 finalists.  

 
03:54 
Dom Hawes 
Wow.  

 
03:55 
Steven Millman 
You know, data quality is the poodle. We were best in breed and the overall winner. We were best in show, so we took home the gold. It was exciting.  

 
04:03 
Dom Hawes 
Well, many congratulations. Best to you and to Dynata. It's always good to know that the guests we have are literally the best in the business. Now, today we're going to talk about a subject that many marketers may not immediately leap to or think is extraordinarily obvious. I'm hoping we're going to dig into it, and I'm hoping I can understand enough about it, but we're going to look at quantum computing and the impact that quantum computing is going to have on artificial intelligence or the development of it. And then we're going to, because it's the industry we both come from, we're going to have a look at what the implications might be for marketing and for market research. So, imagine I just got off a spaceship from another universe. Help me understand what's quantum computing.  

 
04:43 
Steven Millman 
Sure. Of course. If you just got off a spaceship from another universe, you probably would know it better than me, but I'm going to go with answering the question. The spirit who was asked, first off, let me just make sure everybody is aware I am not a quantum physicist. I am just a little bit dangerous with this. And so I would highly encourage any quantum physicists or quantum mechanics who are listening to this now to highlight my mistakes in the comments.  

 
05:04 
Steven Millman 
But at a very high level, quantum computing, which exists in the sense that they have made technology that works but doesn't exist, in the sense that there is no such version of quantum computing, which is better than current computing, is important because the underlying mechanism by which it works is the potential for being many orders of magnitude faster than anything that can be conceived of today with modern computing. So, modern computing works on what's called binary gates, right? Zeros and ones. And you have these tiny little transistors, and they're either on or they're off, and that's how these systems work. Quantum computers use qubits, and qubits don't work like that. Two really important things to know. Qubits, or the way quantum computers work at the atomic and subatomic scale, and so they are naturally tiny.  

 
05:56 
Steven Millman 
So we talk about the improvements, like in GPU's and all the work Nvidia and others are doing that is about making these transistors smaller and more efficient. These are already atomic scale. The second is, and this is always the mind bender, is that they can exist in more than two states. So you'll hear people talk about quantum entanglement and quantum tunnelling, and I'm not going to get into them much, except to say that one of them reasons why this is so much more efficient is because the state of one quantum entangled particle tells you what's happening with another quantum entangled particle that could literally be anywhere in the universe. Distance no longer matters at this scale.  

 
06:34 
Steven Millman 
It's really weird physics, but as a result, you could do potentially millions of computations with one of these units, where you could only make one or two at the same time. It is staggering what is going to happen, but hasn't happened yet. The folks that I talk to, who are experts in the field, think that we are probably three to five years before real quantum computers exist. And in all likelihood, because they're going to be monstrously expensive, it'll be set up as like your cloud services. So there'll be a couple hundred of these computers in the world, and everybody will buy time on it.  

 
07:06 
Dom Hawes 
Okay. I was going to say it's sort of starting to feel a little bit like kind of mainframe world, where you have these machines, but they're not necessarily available or practical for everyday use. But I guess if they're that powerful, as you say, you're just buying, it's like a virtualized server, I guess you're just buying part of the processing power.  

 
07:24 
Steven Millman 
That's right.  

 
07:24 
Dom Hawes 
So what's the relevance? So what's the connection between quantum computing and artificial intelligence? Because artificial intelligence, we all use it sort of daily, but like really big practical applications for desktop users, are still hard to find unless it's packaged in applications we've been talking about before. Where is quantum going to help the development of AI?  

 
07:44 
Steven Millman 
The convergence of where quantum computing and AI hit is that quantum computing allows the AI to compute much faster. And that means, you know, we talk about the large language models, we talked about it on a previous episode where they use an enormous amount of energy, a tremendous amount of carbon dioxide is produced, but also, you know, they take time to run. So if you've ever played with Chachi PT or Bard, you know, it sits and it spins for a little while, but if you want to do something really big, these things will take forever. Quantum computing will make it possible to run these very massive scale computations necessary for artificial intelligence to grow, to happen in timelines that make it rational to actually proceed. And this is where we get to artificial general intelligence.  

 
08:33 
Dom Hawes 
Let's just tie it back now to the world of marketing, because one of the benefits, presumably of speeding things up, particularly language models, is that we can start to model the real world in real time is profound for marketers or for researchers.  

 
08:45 
Steven Millman 
Yeah, it's going to be profound for everyone. We're going to talk about market research, but also just imagine what this is going to do for programmes that look at development of new drugs or protein folding, weather forecasting, all of these things where when the parameter set gets huge, it takes forever to run. But, yeah, let's talk about marketing for a minute. So we generally think about the impact of artificial intelligence on marketing in two ways. We usually think about it in the context of creative development, and we think about it in the context of advertising delivery. We're going to see some really profound changes in both of those. So starting with creative development today, you can use AI to generate ideas for advertising. And that is the predominant use case in agencies today.  

 
09:30 
Steven Millman 
So if you're talking to agencies about how they use AI, usually something like 40% of it is just focused in on creative, it produces something that is in an unusable state, but that is itself a very good prompt for imagination and creativity of the human working with it. And you can play with those and get to a place where it's really valuable. Once you've got the creative, you can then iterate the creative. And these are things that are happening today. Once we have quantum computing and we'll talk about general intelligence later, but the AI's will become profoundly more intelligent and profoundly better able to construct things that might be more usable right out of the box.  

 
10:09 
Steven Millman 
To the extent that it may be possible that the AI's will be able to generate creative on the fly in an actual live environment, a lot of danger to doing that. No one really wants to do it, but today it would be madness. But it's possible in five years that based on a theme directed by humans, that it could be producing most of the content on its own in some circumstances.  

 
10:32 
Dom Hawes 
And if you have that capability to process raw information and data, producing creative is one thing that's fine and hyper personalising is another, but presumably you can also use that same power for analysing. So if you think about, you know, a widely deployed series of sensors in a, like an IoT environment, you could be taking real time information, processing it in real time and feeding back in real time, so that all of your output is hyper localised to what's going on in a particular environment.  

 
11:02 
Steven Millman 
And there are people today who are looking at creating value exchange products. So where I agree to help train the AI on how it needs to talk to me, okay, and I get value back from that in a variety of possible ways. But imagine that I'm having a two way conversation with my device and my device is reading the things that I do. It's potentially listening to the things that I do, God forbid. But let's assume there are lots of things you can do with this.  

 
11:27 
Steven Millman 
And then through that two way conversation, allowing me to interact with brands in ways that today we might think of as quite unsettling, but might in the future really feel like I've got a good relationship with the brand through this process and I'm only being asked to submit myself to the kind of communications that I've established I'm happy with.  

 
11:49 
Dom Hawes 
And that doesn't feel like any extra effort at all.  

 
11:51 
Steven Millman 
Exactly. It's learning from everything we do. And if it's properly constrained, properly trained and behaves properly, these are all things that aren't always the case. Yeah, it could be seen as a very valuable companion. Just bring with you everywhere.  

 
12:07 
Dom Hawes 
How are you getting on? Pretty well, I guess, because you're still here. And I have a feeling that we will have chased away the more lily livered of our community today. It's not the lightest of subjects, but if you're still with us, you're going to find what Steven has to say more and more interesting as we get deeper into this conversation. I'm going to recap quickly, emphasising the relevance to us marketers that first up, quantum computing isn't just about faster calculations. Speed does matter, of course. I'm going to come to that in a minute. But quantum is also revolutionary because it's going to be so much more energy efficient, which means our businesses can become a lot more sustainable. You know, AI consumes up to ten times more energy for a simple prompt than a regular search engine does.  

 
12:50 
Dom Hawes 
Remember that the next time you default to perplexity. But imagine running really complex AI models and being able to do that without a massive carbon footprint. Hey, what's not to like? Next, let's talk about hyper personalization, which is where speed and raw processing power come to the fore. Quantum computing's real time data processing capabilities, they're going to help us deliver infinitely tailored marketing messages in real time to markets of one. But with greater power comes great responsibility. We, us, we marketers must balance the power we have with ethical concerns. We've got to make sure we protect our customers, their privacy, and that we never break trust. Finally, I just want you to think about this continuous market research. Quantum computing allows us to move away from static, periodic research to an ongoing, dynamic process.  

 
13:46 
Dom Hawes 
This means we can all stay agile, responding to customer trends and market changes in real time. So whether it's sustainability, personalization or agility, quantum computing will transform marketing as we know it. Stay tuned as we explore more about these terrifying advances and what they mean for you.  

 
14:09 
Dom Hawes 
It's interesting, when we think about artificial intelligence now, where we are right now, people often say, oh, it reminds me of the early days of the Internet. But of course it does, when there was a lot of money spent because everyone believed this thing, the Internet was going in a particular direction, and there was this whole thing about new economics and all that stuff.  

 
14:28 
Dom Hawes 
It strikes me there may be a.  

 
14:29 
Dom Hawes 
Parallel now, and I don't have any grounds or knowledge upon which to say this, but I see the huge amounts the big networks are spending on AI at the moment. Of course, it's dwarfed by the work that's going on in the really big models. Given what we've just been talking about that the emergence of a new type of AI facilitated by more powerful compute. Does it feel to you like the investment now is defensive and actually at some stage is just going to have to be switched off? Or do you think that networks, what they're investing at the moment, can be built on in future?  

 
15:03 
Steven Millman 
So I think what we're looking at today is that people who've invested nothing right now are well behind their competitors, and they are becoming less and less competitive. And it flips two ways. There's a general sense that if I'm not investing heavily now, wherever we are in three years, I will be so far behind, I might not be even able to catch up. You've got people like Publicis group who just invested. I want to say it was $326 million.  

 
15:31 
Dom Hawes 
But look at their results over the last three years. I mean, they've gone gangbusters.  

 
15:35 
Steven Millman 
Yeah, they've been doing an incredible job. They've got some really interesting tech. Not able to talk about it, but they're doing some really amazing things. Ipsos has really interesting walled garden now that they're doing some really cool stuff with. It's also got the best name of any walled garden anywhere. I think I mentioned this before, it's ipsos facto. But everybody's working on this, and there is a risk to being the next blockbuster and watching Netflix run away with it. The other piece to this is that companies that are not creating sandboxes to play with these are going to be wholly unable to take advantage of the new technologies as they deploy. So part of it is defensive.  

 
16:13 
Steven Millman 
Part of it is also you want to be the first person to be able to take advantage of the new technologies to be hyper competitive in the future. Lastly, in the, and it goes without saying, in this marketplace where most of the companies are either public or are, equity valuation increases dramatically if you show you're at the forefront with the AI, and that's probably going to be true for many years.  

 
16:34 
Dom Hawes 
Do you think in this case, that has to be proprietary AI, because the things are developing so quickly? It seems to me that as long as businesses understand how it's going to affect their process, even if they're not developing something that's proprietary, as long as they're keeping up to speed, at some stage, they'll be able to buy a lot off the shelf.  

 
16:53 
Steven Millman 
Most of this will be bought off the shelf, to be clear, but my grandfather, who was an electrical engineer, he used to say that the guys who get rich aren't the guys who build the brand new thing that no one's ever seen before. They're the guys who figure out a novel way to use it. So the guy who invented the clock radio didn't invent the clock and he didn't invent the radio, but he did figure it out that if you put a wire in between them, you've got something you can sell. And so I think this is really where we're at. So publicis is spending all this money on AI?  

 
17:19 
Steven Millman 
It is, and I don't know the internal workings, to be clear, but it's unlikely that they're going to use that money to build a brand new language model of their own or some brand new fast compute technology that's not their business. And they don't want to do that. But it is going to be them figuring out how to take novel advantage of these tools in ways that other people haven't thought of yet. And that's where I think they're really going to see multiples.  

 
17:40 
Dom Hawes 
Yeah, I mean, it's really interesting to see how the major networks are faring based on their attitude to this technology.  

 
17:46 
Steven Millman 
There's definitely going to be some winners and losers based on the how seriously you take this stuff.  

 
17:51 
Dom Hawes 
What about then for small and smaller, or independent agency networks, whether that's an ad agency or whether that be an independent market research business, where do they stand? Are we looking at massive consolidation in the market, do you think? Or do they have to develop boutique skills? How will they survive if they can't invest?  

 
18:10 
Steven Millman 
I actually think this is going to be a lot more like the.com boom than some sort of giant sweep up of smaller companies. What these tools are going to do is they're going to enable small or niche firms to do things that big firms can do at very low cost and with very little advanced skills and very little investment. And I think as smart, driven, very clever people start playing with these tools, you are going to see some of these boutiques not only take advantage of all the labour efficiencies. So, you know, the number one thing is how can we do more for less everywhere, but especially for small boutique companies? But they're also going to start figuring out interesting new ways to do things.  

 
18:48 
Steven Millman 
And I think what we're going to more likely see than accumulation and building of conglomerates, I think it's more likely that we're going to start to see stuff that happened with the.com boom, where direct to consumer started to appear in these big brick monsters, which were never really worried about competition from an upstart suddenly having to say, wow, that is an agency that could grow rapidly, that's a creative that could grow rapidly, that could really expand in ways that disrupt my business if I stay tried and true to the things I've always done.  

 
19:21 
Dom Hawes 
And in house, I'm seeing quite a lot of really cool, it's kind of low level innovation, but it's everywhere. So I mean within the organisations that I'm looking at. I met with a super smart CMO of a PE backed tech business the other day and she's got a team in house. She gathered her team together and said, right, you've all got 5 hours of time to go out, research and find a piece of artificial intelligence that can improve your daily work. And we're going to get together in a week and we're going to review them all. Anyone that brings one back within budget, we will adopt. And 17 different bits of technology got adopted into her team and one of those bits saved her 40,000 pounds off her annual marketing budget within that first week.  

 
20:05 
Dom Hawes 
That kind of grassroots innovation I think is really interesting.  

 
20:08 
Steven Millman 
It really is. And I think we're going to see more and more of that, especially as people start getting clever with combining some of these AI applications or startups into much more interesting clock radios. Yeah, yeah.  

 
20:21 
Dom Hawes 
Well I'm not going to give anyone the name of that particular CMO because I'm trying to get her to come.  

 
20:25 
Steven Millman 
And talk to us on the show.  

 
20:27 
Dom Hawes 
And I want to dig into that story in particular.  

 
20:28 
Steven Millman 
Well if I can help you, let me tell you really should come. Dom is great, this is a wonderful experience and you should share what you're learning.  

 
20:41 
Dom Hawes 
Well there's some food for thought.  

 
20:42 
Dom Hawes 
And as we wrap up todays episode, its clear the future of marketing. Well its not yet on the brink of a revolutionary shift, but the storm is gathering. Quantum computing and AI will redefine the how, the what, the when, the where and the who of our engagement with our customers and non customers. At least the why will still remain true. But look at it another way. And hyper personalization, real time data processing and the power of Internet of Things, which today it's been a little bit disappointing unless smart light bulbs get you off. Well, the opportunities of a combination of these things, it's immense. As I've mentioned already, and I will mention again, with great power comes great responsibility as marketers.  

 
21:27 
Dom Hawes 
This technology will accrue power to us, so we must manage these things, these advances ethically prioritising customer trust and privacy, the integration of these kinds of technology is going to enable us to deliver more accurate, more efficient and more personalised customer experience.  

 
21:46 
Dom Hawes 
Right, I'm off to take a cold shower.  

 
21:48 
Dom Hawes 
I need snapping back to reality. But do join us on part two. You have been listening to unicorny. I'm your host, Dom Hawes. Nichola Fairley is the series producer, Laura Taylor McAllister is the production assistant, Pete Allen is the editor and Peter Powell is our scriptwriter.