Welcome to the first episode of the Expert View. Today we’re hosting Lars Rinnan. Lars is a venture partner at Scale Capital and is the CEO of the Amesto Next bridge, a Norwegian AI and data intelligence consultancy. Lars is also an investor in exponential technologies and holds an MBA from the University of Lancaster.
So Lars to start us off. What is AI?
Yeah, that’s the big question. I think if you ask some people on the streets, you’ll probably get 10 different answers. If you look at the history, it was actually coined there all the way back in 1956, by a guy called John McCarthy, who was a computer scientists, and at a conference at Dartmouth, they coined the term artificial intelligence. At that time, they defined this as being the science and engineering of making intelligent machines.
Today we would perhaps put another definition on it, maybe we’d say that it’s the building of a computer to do tasks that is usually done by humans. And this can be anything from object recognition, it could be pattern recognition across a lot of various data sets. Of course, it could also be very complex numerical, or statistical mathematical equations.
AI is a is a lot of things, but what it definitely is not, it’s not like the science fiction movies like Terminator and all that, where you have conscious machines trying to kill humanity. That’s definitely not what it is.
Lars, now you’ve chosen to spend significant time working with AI. How do you see AI changing the world in say the next five years?
I think that AI is actually going to change the world in fundamental ways. I think it’s going to change the world more than the Internet has changed the world.
If you look back for the last, let’s say 20 years, I think the Internet has changed the world dramatically. I mean, we interact with the internet in everything we do, every day, now, in our personal lives, and in our business lives in if not all industries, at least most industries. I think the same is going to be true for artificial intelligence.
If you look at the really big issues, that the big challenges in the world, like famine, environmental problems or poverty, I think that AI has the power to actually if not eradicate all those, at least, to minimize them to a large degree. That’s probably not going to happen in the next five years, but if you if you look towards the year 2029, I think a lot of those things actually will happen.
The reason why I’m focusing on 2029, is that by 2029 a cheap computer that can buy anywhere, like $1,000 computer will have the same compute power as the human brain. Now that changes everything. So that means that a computer, a cheap computer that basically anyone can own, has the ability to do all the things that you do now, cognitively. And that’s a big game changer. . So that is going to be shaping the world in 2029. I think it’s going to be ubiquitous, it’s going to be everywhere, I think, and I hope that it will change the world for the better.
Now that part, that’s really up to us humans to decide. So we better make wise choices going forward. And we better choose now, the best politicians going forward. That’s also a big question. And we could probably spend two hours talking about that one.
That sounds very interesting. So maybe we can touch a bit about that. What are some of the inherent downsides to AI, now that we’ve spoken about how AI will change the future?
I don’t think there are any inherent downsides. I mean, that’s looking at the semantics, because AI is basically just the technology. It depends on how we humans decide to use that.
You can use nuclear power for energy generation, or you can use it for arms. And of course, those two are totally different. But you know, the nuclear technology is the same, and as the same is true also for AI can obviously use this technology, not for good, not for making the planet, a better place to live. But you know, you can use it to gain control over, bank accounts, people’s opinions, you can obviously have AI-controlled weapon systems, which you know, will be very efficient. And of course, you will have a big difference between, let’s say, the country that has this kind of weapon system, and the country that doesn’t have it. The same way that you always had the differences in weapon technology.
Of course, AI can also be used in there in the wrong ways maybe, unintentionally, you can have bias, because AI has to do with training on historical data. And, of course, you can have an ethical use of AI, you use AI for fraud. I mean, it’s possible, so some people will do that. And then you also use AI for anti-fraud, this is a this is a bit like, spam and non-spam, viruses and anti-viruses, which is it’s kind of a race, who is who is the better at any given point of time.
So I think that’s going back to the question, I don’t think AI has any inherent downsides, but you have all the possible downsides of any technology, perhaps, you know, to a higher degree, because it’s such a powerful technology.
Lars, you’ve helped a large number of companies implement AI into their business model. What are some of the common mistakes that you see companies try to make when they first try to integrate AI?
Yeah. So I think we’ve done you know, just about 30 projects, the last couple of years for the AI projects. I think that the basic mistake is that AI nowadays is such a high technology, so everyone has heard about it, not everyone knows it really well, or understand it really well.
On the customer side, I think that some organizations they’re just intrigued by the technology and everything they’ve read about it. So they come to us and say, “we want to do an AI project, can you help us?” I think that’s starting in the wrong end. Because I mean, often there’s there’s no real use of doing an AI project. Why would you do an AI projects? So our answer then is, yeah, we we might be able to help you. But what are you actually trying to solve? What is the business problem that you’re looking at?
So I think that’s, that should be the starting point, because AI can probably address a lot of different business problems, probably not all. But yeah, I think at least most of the common business problems, and quite a few of the uncommon business problems as well, are possible to address with AI, but then you have to start with that. What is the business problem? You know, what’s the business case? So I think you can make really solid business cases for most AI projects, because it is about you know, improving processes being more efficient, automating finding new answers to previously unanswered questions, and of course, you can make business cases for that.
I don’t think you should regard an AI project as a technology projects, because it’s more about processes and ways of working, you know, it’s it will affect the processes. And of course, processes affects people and employees and the organization. So I think you need to have a more holistic view of such a project as probably true of most technology products, they should not be seen as just a technology project. But I think that AI is perhaps even more so. Because again, it’s such a powerful technology. So it will change the way that your organization is doing their work.
So you would say it’s more an organizational project that also needs a strong underlying business case or a to be optimally implemented in companies.
I think a lot of people in the startup world grappled with this question, when many of the large tech giants and enterprises have entire AI divisions. Is it possible for startups to compete in this AI space?
It’s a good question. I mean, how do you compete with Google in terms of AI? I think that, I think it definitely is possible for startups to compete with the large tech giants, and you see that all around, because there are so many startups with at least some kind of AI component.
Of course you need s some really good data scientists. it’s also quite hot to be an entrepreneur these days, and that’s a pro for the startups. So a lot of data scientists wants also to be an entrepreneur and that’s a perfect match. You probably can’t pay them as much as the big organizations, but you have equity, so you can give the data scientists some of the equity in stock compensation plans, and I think they’ll probably be, at least, the less risk averse data scientists will be quite happy with that.
I do think you to need to be really careful about who you employ. I think that’s even more so for startups than for the big enterprises. Because I mean, if you’re a big bank, and you have a divisional of, say, 20 data scientists, you can actually kind of afford to have, one of them who’s not so good, but if you’re a startup, and maybe you have only one data scientists or hopefully, a small team, so that the one data scientist has someone to discuss with, then you need to be more careful about who you employ.
There are so many people now that take a Coursera course, or Udemy course, or two or three, and then think they are data scientists now. It’s not that easy, you need to understand the math, you need to be really good at programming, and then you also need to be really good at understanding business. Because as I said, it is about improving business, or government institutions, this understanding is critical.
You need to also be able to explain what you have done, and what’s happening to non data scientists. And for some of them, this is really hard, because they have no problems discussing with their peers, but if you are going to explain to our chief marketing officer, what have you actually done with this algorithm? Why is this is the answer for him or her, then you better be really good at understanding how the marketing officer is actually thinking.
So in continuation of this sort of figuring out who is actually a good data scientist, I think another question the industry grapples with is the talk of AI snake oil, how would you approach identifying well made or badly made AI when you’re looking at a company?
This is a bit of a problem these days, primarily for two reasons.
I think one reason is that just about any startup has some kind of, you will see AI or machine learning or something like that somewhere in that pitch, it has become almost a requirement.
Obviously, not all startups have an AI component, maybe they say it has, maybe it has like a bad AI component or a wrong use of AI. But here comes the other part. Now, most investors, angel investors or VCs , or large companies, they don’t have the competence to tell right from wrong when it comes to use of AI. I mean, how could they? And you probably need to be a data scientist yourself, and most investors are not.
But I think that the best answer is that the first question will be: ‘why?’. Why is a fantastic question, really, why are you using AI? How does AI improve your solution? And now, if they can’t give a really good answer to those two questions, why and how, now that is probably something fishy for them use some kind of outside resource, and they haven’t understood what they have done, and that’s also a red flag. So just those two basic questions, why and how will take you a long way.
That’s really good advice Lars. On a more broad level, a trendy question to be asked in the AI world, will Humans ever be able to develop AI with full consciousness?
Now we’re going back to do the the sci.fi movies, Matrix and Terminator and all that. But even taking into account that I’m fully aware of the exponential rate of development, I think that’s actually quite far into the future, because it’s so complex.
Now, it’s extremely complex, and there is a lot of research on the subject, but so far, it’s it’s really, really basic. Do I think that we will ever be able to develop that? Yeah, sure. I never say that this will never happen when it comes to technology. Because I know that most things will actually be able to, to happen now, when it comes to technology, could take three years, could take five years could take 10 years.
When it comes to full consciousness on AI, I think that it’s going to take a lot more than 10 years. And I think that for now, we should regard AI as a fantastically useful tool. It’s a technology and it’s a useful tool with a potential, you know, just to solve the big global problems, as well as improve just about any business. And you know, I think that’s, I think that’s good enough for now.
Well, I think that’s a perfect closing remark, Lars. So thank you very much for your time and your insight. It’s been great having you here on the Expert View.
Thanks for having me.
Lars Rinnan is a visionary CEO, angel investor, board member, public speaker and futurist from Oslo, Norway.
Lars has more than 25 years management experience, and over the course of 7 years has started 6 companies within business intelligence and artificial intelligence, all joined together in Amesto NextBridge.
His entrepreneurial and business experience makes him a valued investor, advisor and board member for startups within AI and exponential technology.
Amesto NextBridge is a consultancy specialized in Business Intelligence, analytics, ArtificiaI Intelligence and Data Science.
In addition to working with the largest enterprise customers, the company also has an AI Lab which is dedicated to helping start-ups succeed by implementing AI-capabilities from an early stage.