Neale 00:00
Welcome to Another Bright Spark podcast. We are here today with Mark Bernstein who’s had a fantastic and interesting and varied career in lots of technology companies. We’re talking Geocities, we’re talking Virtuality. Mark, welcome to the podcast.
Mark 00:13
Thank you very much for having me.
Mark 00:15
You’re very much welcome. I want to drill into some of these organisations because you’ve had a career which I think is safe to say you’ve kind of been ahead of the curve in a lot of these organisations and I want to sort of drill down into that and talk about it in a little bit. But first and foremost, where does this passion and this excitement for technology come from? Childhoodo.
Mark 00:36
Yeah, when I was in my early teens, I had some elder cousins who were at Cambridge University and they actually took part in the Cambridge phenomenon, the Cambridge computer revolution of the 70s and 80s with Acorn and Sinclair, and they were part of that group of people. And when I first met them, they’re very boffiny sort of guys, and I was only 13, 14, but I talked to them and I thought, you actually need somebody commercial to work with you to help exploit this stuff. And very broadly through my career, I’ve been the commercial guy, helping the founders, the guys with the great ideas, and certainly started off that way, and then I shocked myself as an accountant and I began to have ideas of my own, which, and so I’ve done a number of start-ups of my own as well as helping other people, but I’ve tended to be on the early stages of technologies and indeed businesses, although I’ve done some larger businesses as well. And I had a fascinating time, I had great fun doing it, which is really important.
Neale 01:48
Does it inspire you technology?
Mark 01:51
Yeah, I am. I am interested by, I mean, I’m not a massive gadget person, but I’m very interested in, I mean, for example, autonomous driving is something that I’m finding fascinating at the moment. And I think about how that’s going to change your world. And if I’m changing, maybe we’ll talk about that later on.
Neale 02:10
Well, we’re talking about Diligentia, who uses AI through your current company. So you’re using AI to sort of simplify investments. There’s been a lot of hypothesis around AI. Where do you see it sort of moving yourself? Obviously, you’ve got quite a bit of insight at the minute with where you are.
Mark 02:27
AI is going to change the world more than the internet. And that’s saying something. I’ve got grandchildren, and I’m imagining the world that they’re going to when they enter the world of work, for example, what jobs will be around for them. Now, in the early 80s, when computers came along, I remember actually a big discussion about what computers are going to come. So everybody’s going to lose their jobs. Jobs, yeah. What you found is actually that’s absolutely not what’s happened. Jobs have changed. They’ve shifted. Boring admin jobs have been taken over by now lots and lots of people working in the IT industry and so on. So the world of work with the advent of microcomputers and the computer revolution and Microsoft made things easier submitted. It changed things. It didn’t. It was clearly disruptive, but it didn’t cause mass unemployment. Then we came along with the internet and lots of discussion about what would happen with the internet. And what we found with the internet, I think first and foremost, is that it has democratized information. So now, anybody anywhere in the world, if they’ve got a computer, can have access to information that a generation ago you probably could only find in university libraries.
Neale 03:47
It’s absolutely killed pub conversations, hasn’t it? Because you can no longer debate. It’s like, well, I’m pretty sure she was in that movie in 1972. Well, actually, I’ll just get my phone out and check.
Mark 03:57
But now with AI, and I think AI and robotics are going to combine, we’re going to see all sorts of challenges and the upending of a lot of traditional roles. If you take the example of my grandson in 20 years’ time as he enters the workforce, the jobs that will be available to him will be very different. I think a lot of jobs will have been, they won’t exist, they certainly won’t exist in their current form. And perhaps AI is going to impact 95% of jobs. But again, I think we will find opportunities for people that AI will create and robotics will create.
Neale 04:51
I was sort of hinting towards it just earlier. I think AI can sort of supplement a lot of things. So just a small example, just Christmas just gone a month ago. We sent out a postcard mailer and I had a lot of data and they had data in different fields and the columns were mixed and ordinarily in a previous life he’d get involved in catalog mailings. It’d take me a week to sift and through sort of the columns and of course Common Delimited and Changing Things and Bits I stuck it in chat with GPT and it was done within seconds. There was a little bit of refinement but what would previously take me days going through 25,000 contacts took seconds and just supplemented the
Mark 05:35
Well, I’m involved in a little project at the moment with sending Some information to a government department and I know that the government department is going to put that information into church EPT To help it. Okay. It’s a lot longer. They can they can say they’re going to You know, they could they can use relatively unskilled on people with the help of AI to actually interrogate quite complicated AI stuff so I’ve been using AI to to try to Figure out what they’re going to be thinking about and what what you end up is is a battle of the bots You’ve got them using AI to try to do something and me saying to my AI Anticipate what they’re going to do right and change what I’m doing in my inputs to make sure that it’s going to pass What they want me to do. So suddenly we’ve now got bot battling bots
Neale 06:33
And hosted by Craig Charles.
Mark 06:37
Exactly. And I’ve got a prediction that the biggest thing that will hit business in connection with AI is actually what are called custom LLMs, custom large language models. So most people are familiar with chat GPT and most people will know from having used chat GPT that you’ve got to give GPT some context before you actually ask it the question. We’ll give it the query and then it gives you a fantastic result but I think most of us would accept that you need to give it the context and the more context you give it the better the response. Well, the GPTs now are now giving you the opportunity to create what’s called a custom GPT where you give it a knowledge base about your business for example, which sits in the background. And so rather than having to explain to the AI the context, it takes the context from the knowledge base that you’ve created and the result is in my experience probably 10 times more powerful because it’s got that context and you don’t have to, you can upload lots and lots of data to give it the context and it becomes incrementally more powerful. And I think there is going to be a divide, one of my predictions for the next five years, there’s going to be a divide between those organizations who utilize the knowledge base and the custom AI versus people who just use the vanilla AI. And what many people probably don’t know is that when you put all your stuff and your IP into the AI, sort of owns it and the AI adds it to its own knowledge base so it becomes more and more knowledgeable and we’re all feeding these knowledge bases. But with custom LLMs, they claim that they don’t then farm your IP and your data and add it to their general models. Now if that’s true, it does mean that people can put in their IP into their custom models and it can remain private if you trust. How long have you been involved with DiliDense here? Started a couple of years ago, we’ve been building it for two years, it started out as a three month MVP and it’s grown from there because we wanted to reduce the amount of time it takes people. We think there’s a need for, if we take a step back, most small companies, if they want to grow rapidly, look to take investment. Anybody who’s trying to get investment as a founder knows it’s incredibly time consuming.
Mark 09:41
frustrating a lot of the clients that we so just a lot of the clients that we when we work with startups say you know they’ll build a prototype that always chased investment always you know it’s the number of times they get to one stage great gets the next stage just need more more this always investment comes through if they make it easy
Mark 09:59
If they’re doing it for the first time, they don’t really know what sort of information they should provide. At the bottom end of the market, the amount of money that they’re looking for doesn’t warrant very expensive diligence from professionals. So I’ve been in that situation looking for money many times, but I’ve also been an investor, both a corporate investor doing larger deals, but also an angel investor. And from the investor’s perspective, it’s not cost effective to hire professionals. It is really a fact that at the bottom end of the market, most investors lose most of their money on most of their deals. Now, that doesn’t mean they can’t come out ahead because one in 10 can do very well and cover all the losses, but it’s a really inefficient process for them. They waste a vast amount of time. Nowadays, one will tend to find a young founder who’s very presentable, knows what he’s talking about with a great deck. And it’s very difficult to get underneath the deck to find out really where the business is. Normally, as a founder, you’re always talking about the transformational deal that’s just around the corner and you just need that little bit of investment. And you’ll get in. You’ll get there. Yeah, yeah, yeah. Promise. As the investor, you go, OK, I’ll take that as gospel. You make your investment. And funnily enough, that transformational deal either never happens or it takes another year or 18 months and the guys run out of money and you’ve got to put more money in. And this is incredibly inefficient allocation of capital. It’s time consuming for everybody. It puts angel investors off and smaller VCs. If we talk about smaller VCs, most small VC funds never raise a second fund. And that’s because the performance on the first one wasn’t good enough for anybody to give them any more money. So we’re hearing from the government right now that they want to see growth. We know that growth comes from small companies, a lot of it. So we all want there to be more growth. We’ve got to try to make the process more efficient so we track more capital in. It’s deployed better. There are better returns because then we get a virtuous circle of more people putting more money in. And that’s how we can achieve growth. So the idea was that if we could put a platform together that would enable companies to some extent, the diligence themselves, we’d give them checklists so that they knew would know what to do and they could upload files and add notes and so on. And if we’re using AI, we can actually risk score, completeness score. We can give you a bronze, silver, gold. Certificate so you can hold up your certificate and go, you know, with gold certified. Now suddenly the investors look at that and go, well, I’ll spend more time on the company that’s gold certified than not prepared itself. So in short diligence here, we think we want to get you organized.
Neale 13:17
Before we move off from Diladentia, a lot of people watching this, a lot of our clients and customers, they’re always chasing the fund and looking for investment. You’ve been on both sides of the fence, you’ve been seeking investment and also as an investor. I mean just as a headline, top tips, what would you say is really important if you’re seeking investment? What do you need to have together really?
Mark 13:42
The most, you’ve got to have a good idea, clearly. Most investors, most VCs would say that they invest in people, not ideas. So you’ve got to have a team, and you’ve got to have a team that makes people believe that you will end up with a successful business. Most startups go in direction A, and then along the way they realize that actually they’ve got to move a little bit to, they’ve got to adjust their expectations, and they end up sort of at B. So you need a management team that is capable of making that change. In my own experience, the number one issue is route to market, which is not one that necessarily everybody would identify with, but having a good idea is great. Having a great product is great. But how are you, and you may well know who you are targeting as your customers, but how are you going to get to them? How are you going to get to the cost effectively? If you’re aiming at large companies, large companies are really slow to make decisions. So you think they’re going to bite your hand off and they’re not, they’re going to take a year or two. Have you got enough funding for the year or two? Is there a route to market that you can actually get to those people more quickly? So in my experience, people tend to fail because they haven’t figured out their route to market. So my tip would be really think through how are you going to reach those customers in a timely way and in a profitable way. But the big, for people listening, particularly people building early stage businesses, I think the biggest lesson from that is get to profitability. Once you’re profitable, you’re in control of your own world and you can do deals or not do deals, but you’re no longer dependent on external funding, changes of conditions. So whilst it’s very fashionable for people to say, oh, let’s spend the years losing money before we get to break even. Get to profitable. One of the things I’ve learned, probably one of the biggest lessons over the last 30, 40 years is don’t do that, actually. Get to profit, be in control of your own destiny.
Neale 16:04
great advice. Is there any other investment opportunities that you’ve had come across that you thought, oh that one would have been a…
Mark 16:13
Now, you and I haven’t rehearsed this. No, no, no. Most people listening will know the story of some A&R guy who passed up on the Beatles. Yeah, yeah. Well, I have a similar story.
Neale 16:24
Oh, great.
Mark 16:28
Fantastic. So in the mid 90s.
Mark 16:32
I was looking, I was thinking about the internet, what opportunities were there. And there was essentially a business that was, it was an online thing today, it was eBay. And eBay at the time was six people in San Francisco and they created a platform for selling and marketing, I think Beanie Babies. Collectibles. And it was a collectibles platform. Very niche. Very, very niche. But they were beginning to think about branching out into the eBay that we know today. And we’d been, it was an auction site, essentially an auction site. And I’d been looking at another auction site. And in the end, turned away from that idea because of various reasons. And so a friend said to me from San Francisco, he rang me up and he said, you should get the European rights to eBay. They’d be really interested in giving you the European rights. I thought Beanie Babies. No, I can’t. You can’t see it. That’s not going to work. No, that’s not going to work, Beanie Babies.
Neale 18:01
So, yeah, I don’t know. I literally don’t know anyone in my life who hasn’t used eBay. Yeah. Thank you for that. Yeah, it’s okay. Yeah.
Mark 18:10
Okay Yeah, so that was They would because they they were known to my my colleague and they would have done a deal. Yeah, so
Mark 18:27
Yeah, did a lot of fantastic eBay. Wow, there we go. Well, I’m kind of glad that you didn’t because you wouldn’t be here today. So at least the silver lining on your eBay dog cloud.
Neale 18:38
A lot of our customers are the same. A lot of people watching this, you know, of a startup. If they’re looking for any of these to bolster the team, what sort of traits would you what would you recommend? What are you looking for? You want to bring different minds to the table? Yeah.
Mark 18:56
I mean, we’re very much in a world of diversity, but actually diversity of thinking is really good. Having people who challenge you, most of us are wrong quite a bit of the time. Even though I would have liked to have had a non-exec perhaps who said to me, actually this eBay thing is quite a good idea.
Neale 19:20
See the bigger picture.
Mark 19:23
So, I think, I was told this probably 20 years ago, and it’s, I think, very broadly true. You spend your 20s and maybe part of your 30s learning. You’re in a learning phase, and you’re doing and making mistakes and learning. Then you come into your 40s and 50s, or early 50s, and now you’re proving, because you’ve learned, now you’re actually executing with all the knowledge that you’ve gained. And so you’re interested in being an executive or building your business. And then you get to your late 50s, 60s, and at that point, there’s something that happens in the chemistry of our brains, where one wants to give back and take some. And you take the experience that you’ve got, and you want to help other people. And you get paid for it, get share options or whatever, shares. And that’s actually a really good fusion. You’ve got people who’ve got less experience, who are prepared to work the 18 hours a day. And then you’ve got people who want to work a little bit less, who’ve got the experience to go, well, based on what I know, that’s not a good idea. And this might be a better idea. And if you can get two or three non-execs, say, non-execs are surprisingly inexpensive compared to what they were charging for their executive roles. And if you can get a variety of people, maybe there’s somebody with a financial background if you’re not particularly financial, or somebody with an industry knowledge who’s got connections who can make introductions, and again, help this route to market, non-execs are a great way to go.
Neale 21:11
One last question to wrap up. We’ve talked about AI in terms of the future. What did you see in the next 10 years? Where do I see? The industry, the market, where did you see that?
Mark 21:27
One of my things I’m just enjoying at the moment is the idea of autonomous vehicles. We’ll come back to your traffic lights then, Mark. I think autonomous vehicles is fascinating. It’s not just individuals owning cars that will drive themselves, but the opportunities to have swarms of driverless vehicles because the utilization of the driverless, if you think of robot taxis, and I was lucky enough in San Francisco a couple of months to actually sit in a Waymo and actually go on a journey within a driverless taxi. My wife, who’s not particularly techie, didn’t want to get in.
Neale 22:15
Right. I thought you were going to say she was something about screaming.
Mark 22:19
No, she didn’t. She didn’t know I video and you could hear her moaning in the background.
Neale 22:27
I bet she loved that. She didn’t want to get in this car. We’ll attach that as a download on the bottom of the podcast.
Mark 22:33
And then, by the time we got to the end of our journey, she gave it a 10 out of 10. Okay, great. It felt really safe, but you begin to think about the implications of that. Now, it’s not just it’s a driverless experience, but you think about the main cost of an Uber trip is the cost of the Uber driver. You take that off, you take that out, you then say that either the car, you don’t need the steering wheel, so you don’t need a driver, so you’ve got an extra passenger seat. But also, you can build these things. The Chinese are talking about building these things for $20,000, $30,000.
Neale 23:12
Well, they already undercut a lot of EVs because of tariffs that were inundated with Chinese cars. Yeah, we can come.
Mark 23:18
come to that the so if you’ve got and they can be networked so they’re swarming around a city so they’re always a minute away they’re actually quicker to hail one than to get your car out of the garage because they’re being used all the time their their utilization is really high so their marginal cost per minute or their marginal cost per mile is going to be less than it is in your personal car now if that’s the case and it’s cheaper for you in living in a city to hail one of these things you can actually like my son lives in San Francisco he lives in a three-story house the bottom story is given over to a garage now in in concept in the future if he doesn’t need a car he can have a third more living space at no cost you get these things swarming around the cities will give licenses permits to the companies and perhaps the deal is you can operate an uber style environment but for the elderly for the for the disadvantaged the less well-off you’ve got to do it for free or near free now you take the example of the maybe an elderly lady who’s gotten at the hospital appointment at the moment she has to come out of her house she walks to a bus stop which may not be that near she waits for the bus gets on the bus gets off the bus walks to the hospital now if she can actually hail a taxi that picks her up at a door takes her to the taxi to the hospital and it’s free or near free that’s a pretty good thing so I actually think some of this autonomous driving will democratize transport in a way we’ve seen the internet democratize information now so I think I think that’s pretty exciting
Neale 25:23
Just what you’re saying about your son on the bottom floor, if you think about a shopping centre or a supermarket, I mean, their main use of land is for car parking. If all of a sudden you take the need for that and people are just getting dropped off the supermarket, you could use that land for housing, you could use that for…
Mark 25:40
There’ll be massive changes of land use. Of course, yeah, because of the car parks I’ve stayed in. You won’t need all these NCP cars, sorry, NCP, but you won’t need all these NCPs. You won’t need all these city historians. Small city car parks. No, no, you can’t park.
Neale 25:53
All sorts of stuff. Like you say, it’s democratise and information democratise and transport, freedom of land space is just one possibility.
Mark 26:03
I mean personally, I’ve got to an age where I don’t want to be driven, I want the car to drive me, I don’t want to sit there being driven and the sooner this comes in, the better. So I think we’re going to see massive changes, but also big issues. If the Chinese have got economies of scale, they’ve got the raw materials. How are we going to compete in some of these industries like that, and should we be competing? So I think, and it’s also to do with, this ties back to our very first comment, AI. Actually, what are these things? They’re actually, they’re not really cars, they’re laptops and wheels.
Neale 26:51
You want to have a robot driver. We’re on the conversation about automated cars just as a day and we’re saying there’ll be a driver you can talk to. I’m pretty sure very much it would be you’ll have your phone almost like an Uber you you hail something you’ll go in and then either connect your phone bluetooth it or you’ll control the car via an app. It’s not going to have a robot driver or you don’t have to do the controls yourself it’ll be very much a plug and play.
Mark 27:15
Absolutely. And so I think that’s fascinating. I’m slightly involved in one or two projects in that area. But I think AI and in a way that I talked about robotics and how AI and that, you know, autonomous cars in a way, a good example of those two things coming together. So I mean, I think I will see these things. I guess I will see these things as opportunities. They’re not threats. They’re fantastic. When’s a great time to start a business? Well, it’s throughout my career. It’s always been now.
Neale 27:57
Yeah, don’t be afraid. Yeah, absolutely. Yeah. Fantastic. Well, on that note, thank you very much. Thank you for your time. It’s been great. I really enjoyed it. Okay.
Mark 28:05
Excellent.