Latest / Merryn Talks Money / What Happens When Quantum Computing Actually Works?

Transcript

What Happens When Quantum Computing Actually Works? · Merryn Talks Money

0:02
Speaker 1: Bloomberg Audio Studios, Podcasts, Radio News.
0:18
Speaker 2: Welcome to Marion Talk's Money, the podcasts in which people
0:20
Speaker 2: who know the markets explain the markets. I'm Maren Sumset Web.
0:23
Speaker 2: This week I was begining with Daniel don Steinberg, co
0:26
Speaker 2: founder and partner at eden Base, an investment firms that
0:28
Speaker 2: invests in frontier deep tech. I invited Daniel on the
0:32
Speaker 2: show because I had a really interesting conversation with him
0:34
Speaker 2: about quantum computing and just how close it might be.
0:38
Speaker 2: During our conversation, we talk about ai future of and
0:41
Speaker 2: what it really is. We also talk obviously about quantum
0:45
Speaker 2: computing and what it might mean for our futures, and
0:49
Speaker 2: we also cover exactly how one might deal with the
0:53
Speaker 2: future that quantum might bring. Spoiler alert read some philosophy. Daniel,
0:59
Speaker 2: Welcome to marind Money Pleasure.
1:01
Speaker 1: Thank you for having me.
1:02
Speaker 2: Will you tell me a little bit about what Edenbased
1:05
Speaker 2: does and then we'll take you from there.
1:07
Speaker 1: Well, in Bases, three partners got together, all with very
1:10
Speaker 1: different backgrounds, but sharing a common theme, which we were
1:14
Speaker 1: all entrepreneurs and we wanted to invest our own money
1:18
Speaker 1: but also create a platform for two things. One to
1:22
Speaker 1: allow us to hire a bigger team to make decisions
1:25
Speaker 1: in a world that we thought was becoming incredibly complicated.
1:29
Speaker 1: And secondly, one of my partners created Technician for David
1:32
Speaker 1: Cameron level thirty nine, which is the fintech ecosystem. And
1:37
Speaker 1: when I really explored what he had done, I realized
1:40
Speaker 1: that ecosystems solved a few of the challenges I saw
1:44
Speaker 1: with what was about to happen. And so Edenbase is
1:47
Speaker 1: a is an investment vehicle and also an ecosystem where
1:50
Speaker 1: we help with adoption of new technologies. We call it
1:55
Speaker 1: investing in a future we can be proud of, which
1:57
Speaker 1: actually came out of AI weirdly in our three years go.
2:00
Speaker 1: It's not sustainability, it's a future we can be proud of. Okay.
2:04
Speaker 2: I like that. That's better than sustainability. Is so vague,
2:07
Speaker 2: isn't it? Okay? So the thing I really wanted to
2:10
Speaker 2: talk to you about it and I think I invited
2:12
Speaker 2: you want to talk about, was quantum computing. But before
2:14
Speaker 2: we get onto that, I wanted to talk to you
2:16
Speaker 2: a little bit about AI because when we were chatting
2:18
Speaker 2: the other day, right you and I and we were
2:19
Speaker 2: talking about various things, and I said to you, is
2:22
Speaker 2: AI the correct name at all for what we think
2:25
Speaker 2: of as AI, because what we're using at the moment
2:27
Speaker 2: is lllm's Large language models. And anyone listening who isn't
2:31
Speaker 2: entirely sure how LMS work, please go back and listen
2:34
Speaker 2: to some of our previous podcasts if you've had lengthy
2:36
Speaker 2: conversations about all these things. So when we talk about AI,
2:40
Speaker 2: we're talking about these large language models, and they're not artificial,
2:44
Speaker 2: they're man made, and they're also not really intelligent, are they.
2:47
Speaker 2: So we we're using the wrong phrase in the beginning.
2:51
Speaker 1: I think we use the wrong phrase because that's how
2:55
Speaker 1: we referred to technologies such as these, which are we
3:00
Speaker 1: building artificial intelligence And the history of compute, certainly for
3:06
Speaker 1: the last twenty thirty years, has been exponential, and m's
3:11
Speaker 1: are a continuation of that exponential and therefore they fit
3:15
Speaker 1: into the bucket of artificial intelligence. But you're right, the
3:18
Speaker 1: word artificial is not correct and the word intelligence is
3:21
Speaker 1: not correct for what we have built. We have built
3:24
Speaker 1: incredibly powerful technology. I mean, all the technologies we've built
3:28
Speaker 1: have been incredibly powerful at the time. This happens to
3:31
Speaker 1: be even more so, and it's an opportunity and threat
3:37
Speaker 1: for what we have as our society and the way
3:39
Speaker 1: we work is enormous. But yes, you're right, they don't
3:42
Speaker 1: fit what we would classify as intelligence, which is a
3:46
Speaker 1: biological characteristic. These are programmed characteristics. The way they mimic
3:53
Speaker 1: intelligence is they have access to an enormous amount of data,
3:57
Speaker 1: most of which was generated by humans, and therefore the
4:02
Speaker 1: ability to write cogent arguments based on human intelligence is
4:08
Speaker 1: what they do.
4:10
Speaker 2: I have thee conversation with one of my nieces the
4:13
Speaker 2: other day about how you can tell what is Ai
4:16
Speaker 2: writing and what isn't Ai writing? And I was explaining
4:19
Speaker 2: to her that, of course, the way that Ai writes,
4:22
Speaker 2: the way that an LM write is actually just the
4:24
Speaker 2: way that old people write because they're using our writing.
4:29
Speaker 2: There used to be a hallmark of my writing style was,
4:31
Speaker 2: you know, short sentences and dashes, This is how I write.
4:35
Speaker 2: So is it Ai writing or is it someone in
4:37
Speaker 2: their fifties?
4:38
Speaker 1: Well, a good question. You know, it has access to
4:40
Speaker 1: a lot of information, and one advantage I have I'm
4:44
Speaker 1: writing a second book at the moment is the fact
4:46
Speaker 1: that I've written a first book. It knows my first book,
4:49
Speaker 1: and so when I ask it to write in my
4:52
Speaker 1: and my co author's style, it can copy that. However,
4:55
Speaker 1: I do find the style we wrote the first book
4:57
Speaker 1: in is not the what style we writing the second
4:59
Speaker 1: book in, but it can mimic lots of styles. Whilst
5:03
Speaker 1: it is certainly based on the data that we have,
5:08
Speaker 1: it is able to create new things, and so if
5:11
Speaker 1: you explain exactly what you want from it, it will
5:15
Speaker 1: do pretty much exactly what you do, as long as
5:18
Speaker 1: that's within a framework that it can understand.
5:22
Speaker 2: When you look at where we are at the moment
5:25
Speaker 2: with llms, and we think about, say, for example, the
5:29
Speaker 2: legal profession, and that's the place where everyone says, well,
5:32
Speaker 2: lllms will fully replace lawyers in no time at all.
5:35
Speaker 2: How do you see it playing out?
5:37
Speaker 1: Well, I think that's a very plausible scenario. Going back
5:41
Speaker 1: to my book that are in twenty twenty and twenty one,
5:43
Speaker 1: we use the legal profession as something that llms are
5:47
Speaker 1: about to wreak, havocate, as they are in accounting, as
5:51
Speaker 1: they are in management consultancy and a lot of those
5:54
Speaker 1: rule based roles. The reason that will happen is, I
5:57
Speaker 1: think for two reasons. The first is that even though
6:01
Speaker 1: you may argue or not whether it is better than
6:04
Speaker 1: the best lawyer. It certainly has access to more information
6:07
Speaker 1: than the best lawyer, because however much I read, I've
6:10
Speaker 1: got one hundred and eighty six hours a week, I
6:11
Speaker 1: can't read everything. The second thing, and we talk about this,
6:15
Speaker 1: which is, firstly, if you're a legal firm today, why
6:20
Speaker 1: would you hire junior lawyers to do things that lllms
6:23
Speaker 1: can do? And if you do, it's almost like you're
6:27
Speaker 1: doing it for the future of the industry rather than
6:29
Speaker 1: your personal success, because you might as well replace it.
6:32
Speaker 1: And we say, if that happens, then in fifteen twenty
6:35
Speaker 1: years there are no junior lawyers. The second thing is,
6:38
Speaker 1: as we become more LLLM focused, the only thing that
6:43
Speaker 1: can unpick what lllms are writing eventually becomes another LLM,
6:47
Speaker 1: and therefore you migrate to that. So if the court system,
6:52
Speaker 1: for instance, decides to replace certain parts of it with AI,
6:58
Speaker 1: then people will be writing AI to deal with the AI,
7:03
Speaker 1: and eventually we will have a different legal system which
7:05
Speaker 1: will be based on what computers are able to comprehend,
7:09
Speaker 1: like a code for a computer, rather than what humans
7:12
Speaker 1: are going to be able to understand. I see that
7:15
Speaker 1: as as very likely the other thing that we find,
7:18
Speaker 1: and I've been enforcing this at even Base for a
7:21
Speaker 1: long time, is we do not send contracts to lawyers
7:24
Speaker 1: if we can do it ourselves. I used to do
7:26
Speaker 1: that with my first startup many years ago. So what
7:28
Speaker 1: it meant is our legal bin is probably five percent
7:31
Speaker 1: of what it used to be. And that is also
7:33
Speaker 1: a problem for the legal industry.
7:35
Speaker 2: As long as you are fully confident in your LLM.
7:39
Speaker 1: It depends on the risk factor. So an example is
7:43
Speaker 1: I invested in a Spanish company two and a half
7:45
Speaker 1: years ago and I was putting a small investment in.
7:48
Speaker 1: It wasn't worth sending to a lawyer, but I put
7:51
Speaker 1: it through an LM and it took the Spanish contract.
7:54
Speaker 1: And this is two and a half years ago, so
7:56
Speaker 1: they weren't as good as they out there, and it
7:57
Speaker 1: gave me a plausible scenario not have been all right,
8:01
Speaker 1: and I was able to rather just signing it and saying, well,
8:04
Speaker 1: I don't care. I was able to write to them
8:06
Speaker 1: with a few questions and they answered those, so it
8:09
Speaker 1: was better than doing nothing. And as they get better,
8:12
Speaker 1: your risk appetite becomes greater. Secondly, the risk appetite changes
8:19
Speaker 1: as you are able to deal with things yourself, and
8:23
Speaker 1: I think that's really powerful.
8:25
Speaker 2: So what are the limits to the LM model. And
8:30
Speaker 2: one of the things that we've been looking at, and
8:31
Speaker 2: we've done a few podcasts on this is the idea
8:35
Speaker 2: that you can use an SLM, a small language model
8:39
Speaker 2: instead of an LLM, and a couple of years ago
8:42
Speaker 2: that didn't work so well. But the new study at
8:45
Speaker 2: from Stanford University and someone at pameel Iberium has been
8:49
Speaker 2: through it, pointing out that these small language models that
8:52
Speaker 2: you can run on your laptop, you can even run
8:54
Speaker 2: it on your phone, and you compare the performance of
8:57
Speaker 2: those with an LLM. The LMS, of course what we
9:00
Speaker 2: need the massive data centers for you compare these and
9:03
Speaker 2: in fact the results are pretty much as good, pretty
9:06
Speaker 2: much as good. So if you want a fairly straightforward
9:10
Speaker 2: AI model, and you want it to be private and
9:13
Speaker 2: contained in your own laptop or computer, and you want
9:16
Speaker 2: to use less energy, which is also than crucial, then
9:19
Speaker 2: you can just have an SLM on your laptop. You
9:21
Speaker 2: don't need an LLLM, which rather suggests that in some
9:24
Speaker 2: parts we might be going down the wrong path with
9:28
Speaker 2: full focus on LM's massive data centers, etc.
9:31
Speaker 1: Well, I haven't read that report, so I can't send
9:33
Speaker 1: it to you comment particularly on that. Now, what I
9:37
Speaker 1: can say is two things. One is straight lines are rare.
9:40
Speaker 1: So when I was starting to code many many years
9:44
Speaker 1: ago in the derivative trading industry, I've built derivative trade technology,
9:48
Speaker 1: and the early aunties we worked on mainframes and the
9:51
Speaker 1: only people that had desktops and the people that wrote letters.
9:54
Speaker 1: And within five to ten years as I finished my career,
9:58
Speaker 1: before I started my first startup, everybody had a desktop
10:02
Speaker 1: and you were doing everything locally. And when cloud came along,
10:06
Speaker 1: Actually I'm not bad at reading the future, I was surprised.
10:08
Speaker 1: I said, why do people need the cloud? You know
10:10
Speaker 1: we can do it all locally. And then what I
10:14
Speaker 1: failed to recognize was the onset of mobile phones, and
10:18
Speaker 1: mobile phones had limited power, and therefore the cloud was
10:21
Speaker 1: very useful. And so there are phases in which SLMs
10:25
Speaker 1: will be very, very useful. But I don't see that
10:27
Speaker 1: necessarily meaning that cloud huge cloud data centers aren't valuable.
10:33
Speaker 1: The other thing, of course, is as we build new technologies,
10:37
Speaker 1: and SLMs are based on the current LLLM technology, which
10:42
Speaker 1: is based on binary compute, and it is based on
10:45
Speaker 1: GPUs and it is based on existing storage. A mobile
10:48
Speaker 1: phone could do that. And the segue into quantum for instance,
10:52
Speaker 1: is a quantum computer is not a binary system. Daniel
10:57
Speaker 1: explain quantum computing to me, quantum computing is a break
11:02
Speaker 1: from the past. That's the first thing, and the sort
11:05
Speaker 1: of the most comprehensive way to describe it is. One
11:08
Speaker 1: hundred years ago mathematicians created quantum mechanics and quantum physics,
11:12
Speaker 1: and it described the way the universe operated extremely well,
11:16
Speaker 1: better than they had before, and they were able to
11:19
Speaker 1: write beautiful mathematics and described how subatomic particles worked. In
11:25
Speaker 1: nineteen seventies, Richard Feynman the physicists suggested that we could
11:29
Speaker 1: actually control them, and then in the eighties IBM started
11:34
Speaker 1: trying to control them. And what that meant is they
11:38
Speaker 1: were able to instead of operate binary bits i e.
11:44
Speaker 1: Ones and zeros, they were able to operate in effect
11:47
Speaker 1: probabilistic bits, so they were all combinations of ones and zeros.
11:53
Speaker 1: And what that meant was we were able to do
11:56
Speaker 1: two types of compute or in theory, because no one's
12:00
Speaker 1: built one that is big enough to do anything really meaningful,
12:04
Speaker 1: or until very recently, they haven't. It does two types
12:07
Speaker 1: of things that binary computer is very bad at. The
12:10
Speaker 1: vers is optimization. So imagine a good example might be
12:14
Speaker 1: a table plan. If you have five people sitting around
12:17
Speaker 1: a table, you have one hundred and twenty five combinations.
12:19
Speaker 1: If you have ten, you have three million. If you
12:22
Speaker 1: have twenty five people, you have one with twenty five
12:26
Speaker 1: zeros on the end of it a combination.
12:28
Speaker 2: This is why it's so hard to make table plans.
12:30
Speaker 1: Exactly, it's impossible because twenty five people is more combinations
12:36
Speaker 1: than the biggest computer in the world could compute in
12:39
Speaker 1: the history of the universe. If we'd started it in
12:41
Speaker 1: the beginning, it still wouldn't have come up with.
12:43
Speaker 2: It gives you a new respect for wedding.
12:45
Speaker 1: Planners, doesn't it exactly? Exactly? That's optimization. The second thing
12:50
Speaker 1: it's good for is simulation. And simulation is where we
12:55
Speaker 1: have the ability to simulate natural mods. So, for instance,
13:01
Speaker 1: the way that a chemical reaction happens, the way that
13:05
Speaker 1: a material actually behaves at the moment. With our data,
13:10
Speaker 1: we can approximate it because it's ones and zero's a
13:13
Speaker 1: bit like a CD and a vinyl A cd is
13:16
Speaker 1: ones and zeros a. Vinyl is the true wave. Quantum
13:19
Speaker 1: is operating at the vinyl level, and it is able
13:22
Speaker 1: to model the way that a new drug will interact
13:25
Speaker 1: with your body. I would use the word perfectly, but
13:27
Speaker 1: we don't know because we haven't We haven't understood the
13:31
Speaker 1: next levels down and the next ovesalguma and a much
13:34
Speaker 1: better resolution than you could with a binary compute. So
13:39
Speaker 1: for things like material discovery, drug discovery, how chemical reactions happen,
13:46
Speaker 1: like the nitrogen bond to make ammonia, we use in
13:50
Speaker 1: seven percent of the world's energy creating ammonia for fertilizer.
13:55
Speaker 1: If we could understand actually how that bond was structured
13:58
Speaker 1: and how you could break it, potentially you could find
14:02
Speaker 1: out how it works and use a fraction of that energy.
14:06
Speaker 1: And the reason it's so important is optimization and simulation
14:09
Speaker 1: are two of the areas that we are very bad
14:13
Speaker 1: dealing with as humans because we don't have the tools.
14:16
Speaker 1: And we're about to create systems that can do that
14:19
Speaker 1: for the first time. So if you ask the quantum companies,
14:22
Speaker 1: they say they actually have machines that work. Now.
14:25
Speaker 2: Okay, So when I was about to ask you what
14:26
Speaker 2: you meant by about but you say that, the companies
14:30
Speaker 2: would say, this is here already usable stable, it's.
14:34
Speaker 1: A great it's not a black and white. So there
14:37
Speaker 1: is something called a cubit, and a cubit is the
14:39
Speaker 1: equivalent of a bit in computing. To run a quantum computer,
14:43
Speaker 1: you need cubits and you need to connect cubits together.
14:47
Speaker 1: And what confuses the picture more and I think people
14:50
Speaker 1: have to understand this. When people use cubit, you have
14:54
Speaker 1: to ask is it a physical cubit or a logical cubit,
14:57
Speaker 1: and they don't say which is which. The difference is enormous.
15:01
Speaker 1: So a physical cubit is a working cubit. A logical
15:06
Speaker 1: cubit is one that you can control and use effectively.
15:11
Speaker 1: And there is enormous difference, maybe one thousand to one.
15:13
Speaker 1: So certain companies will say they've got a one hundred
15:16
Speaker 1: or one thousand cubit machine, but that's not a logical
15:20
Speaker 1: cubit machine, and therefore it's to a degree not that powerful.
15:26
Speaker 1: And other companies like Quantinium will claim they've got a
15:29
Speaker 1: fifty two cubit machine, but it's a logical cubit machine,
15:34
Speaker 1: which means it's very very powerful.
15:36
Speaker 2: I don't understand the difference between these two cubits. Do
15:39
Speaker 2: I need to understand the difference between these two cubits.
15:42
Speaker 1: You only need to understand it if you're investing in
15:45
Speaker 1: the space, because if you're matching two things that are
15:47
Speaker 1: not equal, then you need to know that there is
15:50
Speaker 1: a difference between them. So if I open the paper
15:54
Speaker 1: and it says company A has a thousand cubits and
15:57
Speaker 1: Company B has ten cubits, which is what they do
16:00
Speaker 1: in the press, those ten cubits may be logical cubits,
16:05
Speaker 1: and the thousand cubit machine is physical cubits, and they
16:09
Speaker 1: have zero logical cubits, And that is the only reason
16:12
Speaker 1: to understand it. You don't really need to understand what
16:14
Speaker 1: that means.
16:15
Speaker 2: Okay, So the logical cubits are the much more powerful ones.
16:19
Speaker 1: They're the ones that are error corrected and work reliably. Okay,
16:24
Speaker 1: that's what you need to know. The point is, up
16:27
Speaker 1: until twenty twenty four, no company had announced a logical cubit.
16:33
Speaker 1: In twenty twenty four, someone announced one, and then in
16:36
Speaker 1: early twenty twenty five they announced ten. Continuum, who are
16:40
Speaker 1: the half British half American company announced fifty at the
16:45
Speaker 1: end of last year, fifty is starting to become meaningful.
16:49
Speaker 1: The one hundred started to become useful. And so if
16:54
Speaker 1: you look at that progression of one to ten to
16:57
Speaker 1: fifty over an eighteen month span, you're looking at one
17:01
Speaker 1: hundred within the next year or so. And then you
17:05
Speaker 1: have a machine that is useful.
17:07
Speaker 2: Useful for what initially? So we've got a hundred logical cubits.
17:11
Speaker 2: See I know or I know what these things are.
17:13
Speaker 2: Now we've got one hundred logical cubits. We've got a
17:16
Speaker 2: machine and it's useful. But what will it do for
17:20
Speaker 2: us initially? What will it change? What will we understand
17:23
Speaker 2: that we didn't understand before.
17:25
Speaker 1: I don't think we will particularly be able to crack
17:28
Speaker 1: anything new because traditional compute, especially traditional compute with AI,
17:35
Speaker 1: is phenomenally powerful. And therefore, even though these things are meaningful,
17:40
Speaker 1: can they do anything more significant in a wide range
17:45
Speaker 1: than where we are with AI today? Probably not. As
17:49
Speaker 1: we start moving up to the hundreds, you know, the
17:52
Speaker 1: three hundred, four hundred, that is the point where we
17:56
Speaker 1: will be able to potentially discover new drugs or discover
18:01
Speaker 1: new materials. Those are the sort of numbers at which
18:05
Speaker 1: it becomes very interesting.
18:07
Speaker 2: Okay, and this is partly down to sheer speed.
18:09
Speaker 1: It's not the speed, it's the number of variables that
18:12
Speaker 1: you can hold. So each cubit can hold a variable.
18:16
Speaker 1: If you can only hold fifty variables, it's limited what
18:19
Speaker 1: the model can do. If you can hold one thousand variables,
18:23
Speaker 1: it's quite powerful because it's factoral, So every cubit doubles
18:28
Speaker 1: the model in effect, like that five, ten, twenty five
18:33
Speaker 1: that I was talking about. It's a factoral number. So
18:36
Speaker 1: adding cubit it sounds like it's linear. It's not. It's nonlinear.
18:41
Speaker 1: The numbers that people bandy about, and let's put it
18:44
Speaker 1: in a bandieback, somewhere in the early thousands you will
18:48
Speaker 1: be able to reach q day, which is breaking encryption.
18:54
Speaker 1: So those numbers are maybe six seven years away at current.
19:00
Speaker 2: The day when everyone's bitcoin becomes useless, right, worthless, because
19:04
Speaker 2: while the encryption is broken.
19:06
Speaker 1: It's very complicated with bitcoin. Actually it's not such an
19:10
Speaker 1: obvious thing. The underlying technology is not necessarily impacted. What
19:16
Speaker 1: is impacted is the wallets. You could potentially break the
19:20
Speaker 1: wallets if you hadn't upgraded them, So it's complex. So
19:24
Speaker 1: the people that have access to their wallets, could upgrade them.
19:26
Speaker 1: The problem with bitcoin is you've got millions and millions
19:29
Speaker 1: of bitcoin out there. For instance, it's a Toshi wallet
19:31
Speaker 1: of a million that cannot, as far as we know,
19:35
Speaker 1: cannot be accessed, so those are vulnerable. The thing I
19:37
Speaker 1: always say to the bitcoin community. What I say to
19:40
Speaker 1: them is if I have a quantum computer that can
19:44
Speaker 1: break bitcoin, Firstly, who is going to sell it to me?
19:49
Speaker 1: Because whoever develops it will be under enormous regulations won't
19:54
Speaker 1: be able to just buy one. And secondly, if I
19:56
Speaker 1: wanted to cause serious damage, I would not be going
19:59
Speaker 1: after bitcoin because it's too small an industry. I would
20:02
Speaker 1: be going after taking down some regional banks and bring it.
20:06
Speaker 1: If you want it to shorter market, which is the
20:08
Speaker 1: only way you make money on it, it's shorting. So
20:11
Speaker 1: so bitcoin is probably in the top thousands of things
20:16
Speaker 1: you would go for, but you will be small to care,
20:19
Speaker 1: too small to care. The bitcoin community cares because all
20:23
Speaker 1: their wealth is tied up in a single asset. They
20:25
Speaker 1: tend to be very hyper focused on that asset, and
20:28
Speaker 1: they're aware of the danger. The problem with most people
20:32
Speaker 1: is they are not aware of the danger because they're
20:34
Speaker 1: not technical, and so most businesses are not considering the
20:38
Speaker 1: quantum threat. Now, my argument would be they probably don't
20:42
Speaker 1: need to because most of them operate on Microsoft or
20:45
Speaker 1: Google or whatever, and they're very aware of the quantum threat,
20:48
Speaker 1: and so most of their technology will be upgraded automatically.
20:52
Speaker 2: Okay, So we get to the point out maybe ten
20:55
Speaker 2: years whenever it is when we get to this QDA,
20:58
Speaker 2: and the first thing that it can have then is
21:00
Speaker 2: we get to the point where that level of encryption
21:03
Speaker 2: becomes pointless. But companies such as Google et cetera already
21:06
Speaker 2: protected against that. So what next? What's the next thing
21:10
Speaker 2: that we see happening that quantum computing will change around us?
21:15
Speaker 1: I think QDA will be scary. I think what next
21:18
Speaker 1: with quantum is it's a very different animal to AI.
21:23
Speaker 1: We have to look at what it does. It's not
21:25
Speaker 1: likely to be a consumer application for the foreseeable future
21:29
Speaker 1: for two reasons. First, it's very complicated to use, and secondly,
21:33
Speaker 1: it's going to be supply restricted for two reasons. One,
21:38
Speaker 1: they're very difficult to build, and secondly, the regulators are
21:41
Speaker 1: looking at quantum already and they're not going to allow
21:44
Speaker 1: you to buy or access a quantum computer. I think
21:47
Speaker 1: what changes is different. What changes is assuming they are
21:51
Speaker 1: supply restricted, and assuming they do what we've discussed, which
21:58
Speaker 1: is optimization and discovery. Both of those are fundamental to
22:05
Speaker 1: business optimization in terms of one percent saving in logistics
22:10
Speaker 1: or in farming, or in weather prediction or any of
22:15
Speaker 1: these things is a huge competitive advantage. The second thing is,
22:20
Speaker 1: in the world that we live where intellectual property is protectable,
22:24
Speaker 1: the second person that discovers something has no value. It's
22:28
Speaker 1: the first person that discovers it. So if I'm a
22:31
Speaker 1: pharmaceutical company and how I have access to the biggest
22:35
Speaker 1: and best quantum computer and no other pharmaceutical company does,
22:39
Speaker 1: my competitive advantage is enormous because I'm out there discovering
22:46
Speaker 1: drugs on a daily basis or an hour however long
22:49
Speaker 1: it takes to run the thing, or an hourly basis,
22:52
Speaker 1: and none of my competitors do. And if they're two
22:54
Speaker 1: years behind, then I've got a two year free run.
22:58
Speaker 1: And because these systems are so difficult to build, you
23:02
Speaker 1: can't just go and find another competitor and say, okay, well,
23:05
Speaker 1: this drugged company has a four hundred or one thousand
23:08
Speaker 1: logical cubit machine. Will you build me one? They won't
23:11
Speaker 1: be able to.
23:11
Speaker 2: So does that lead us into a new age of monopolies?
23:15
Speaker 1: I suspect again the regulators will be on it. But
23:18
Speaker 1: it leads you into a winner takes all scenario. It
23:21
Speaker 1: turns a lot of companies into, in effect, the tech companies.
23:24
Speaker 1: I mean tech companies are renowned for winner takes all.
23:27
Speaker 1: The biggest search engine, the biggest social media way before that,
23:30
Speaker 1: the biggest word processing software, the biggest spreadsheet. They always
23:36
Speaker 1: trend towards standardization, and standardization initially creates monopoly or standards
23:42
Speaker 1: and monopolies, and eventually doesn't because everybody can hook into them.
23:46
Speaker 1: I remember the early days of word it's the person
23:48
Speaker 1: you send it to who didn't have word. It was impossible.
23:50
Speaker 1: But now you send a word document to anybody and
23:52
Speaker 1: they can open it in anything they like. So there
23:54
Speaker 1: is going to be this change, and I think that's
23:56
Speaker 1: what makes quantum so valuable because what will likely have
24:00
Speaker 1: happen is people will realize over the next two three
24:04
Speaker 1: years that these machines are either available or on the
24:09
Speaker 1: cusp of being available, and if you are not building
24:12
Speaker 1: today to be ready for them, it will take you
24:15
Speaker 1: so long. Therefore, you need to be investing incredibly heavily
24:20
Speaker 1: to make sure that you are not the Kodak of
24:23
Speaker 1: the world and that you are the digital camera of
24:26
Speaker 1: the world. I mean, it's not a perfect analogy, but
24:28
Speaker 1: it makes sense. And so what we foresee happening very
24:32
Speaker 1: much with quantum. It will be the same cycle as
24:35
Speaker 1: the Internet and blockchain and AI, but it will be
24:38
Speaker 1: for a different reason. The amount of money that starts
24:41
Speaker 1: going into the industry will start to be huge, and
24:43
Speaker 1: the chatterpt moment will be the realization of big banks,
24:49
Speaker 1: or of big farmer companies, or of big logistics companies
24:52
Speaker 1: or big farming companies that they are not going to
24:56
Speaker 1: be the winner and they have to start investing in
25:01
Speaker 1: this now. And you look at the banks. Actually, the
25:03
Speaker 1: bank's very interesting. Most banks, big banks have a quantum
25:07
Speaker 1: department and quantum team. And yet I would argue that finance,
25:12
Speaker 1: whilst probably the most valuable industry that quantum can go
25:17
Speaker 1: for in terms of raw cash day one, is a
25:21
Speaker 1: long way from having quantum supremacy because you need so
25:25
Speaker 1: many cubits.
25:26
Speaker 2: But when you say a bank will have a quantum team,
25:29
Speaker 2: a bank is not developing the machine. So a bank's
25:32
Speaker 2: quantum team is doing what just watching IBM all the time,
25:36
Speaker 2: building building systems that can use the quantum machine at
25:42
Speaker 2: a later date.
25:43
Speaker 1: They're doing three things on the whole. The first is
25:46
Speaker 1: they are investing in quantum companies, so JP Morgan is
25:48
Speaker 1: a big investor in Continue, the leading quantum company, because
25:52
Speaker 1: they want access to it. The second thing they are
25:55
Speaker 1: doing is they are building algorithms that they can tell
26:00
Speaker 1: on traditional compute that will be ready for when quantum
26:04
Speaker 1: comes along. And the third thing is they are getting
26:07
Speaker 1: access to basic quantum computing. And if you go onto
26:10
Speaker 1: an Amazon bracket, you can rent a quantum computer. It doesn't
26:15
Speaker 1: do very much, but they are starting to test quantum
26:19
Speaker 1: algorithms on cubits because they know, like everybody, that a
26:24
Speaker 1: machine is made available day two, you're not using it.
26:28
Speaker 1: It'll probably take you two or three years before you're
26:31
Speaker 1: up to speed. So they're doing a lot of that
26:33
Speaker 1: work now in readiness for when that happens.
26:48
Speaker 2: Really, the great power of the next phase of the
26:51
Speaker 2: technological revolution is the interaction between quantum and what we're
26:56
Speaker 2: calling AI.
26:57
Speaker 1: Look. AI is an umbrella umbrella over all the other technologies.
27:02
Speaker 1: One of the things that I strongly believe is the
27:04
Speaker 1: convergence of technologies is far greater than adding them together.
27:09
Speaker 1: AI sits on top. You've got underneath that, you've got
27:12
Speaker 1: probably quantum, You've got the data protocols. I used to
27:16
Speaker 1: be a believer in blockchain, not so much anymore. But
27:19
Speaker 1: how do we handle data? How do we handle better data?
27:24
Speaker 1: Different data? And then under that you have three other things.
27:27
Speaker 1: But the AI, what we call AI, is the control layer,
27:31
Speaker 1: the orchestration layer that's just on top of all of it.
27:33
Speaker 2: Okay, Now, who's good at this? Because one of the
27:37
Speaker 2: things that we've seen over the last decade is we've
27:39
Speaker 2: seen who's good at AI in terms of countries, right,
27:42
Speaker 2: so America, China, Israel, etcetera. Who's good at quantum? Who's
27:46
Speaker 2: going to win here?
27:48
Speaker 1: It sort of breaks down as twenty five to twenty
27:52
Speaker 1: eight percent of the US that we've looked at, twenty
27:55
Speaker 1: five percent UK, twenty five percent Europe as a whole,
27:58
Speaker 1: and twenty five percent to the world. China is probably
28:01
Speaker 1: most of that twenty five percent. So if you look
28:04
Speaker 1: at that, the UK is in a very very strong position.
28:07
Speaker 1: And the reason is we tend to have very good
28:10
Speaker 1: quantum expertise here and very good material science expertise. And
28:14
Speaker 1: we've been looking at quantum for a very long time.
28:17
Speaker 1: And that's why Quantinuum, which is part owned by Cambridge Quantum,
28:22
Speaker 1: it was emergent between Cambridge Quantum and Honeywell Quantum is
28:25
Speaker 1: still located in the UK and a lot of their
28:28
Speaker 1: brains are still still in the UK, and they are
28:32
Speaker 1: likely because we don't know what everybody is doing, certainly
28:36
Speaker 1: in China, they are likely to be the most advanced
28:38
Speaker 1: quantum company in the world.
28:40
Speaker 2: So how does someone who's stayed with us so far
28:44
Speaker 2: thank you listeners for staying with us. I know this
28:46
Speaker 2: stuff is not immediately easier than non titans among us.
28:50
Speaker 2: How can we be quantum ready, both not as an investor,
28:53
Speaker 2: but also as an investor well as an investor.
28:56
Speaker 1: The one of things about quantum is there are quantum
28:59
Speaker 1: companies listing and secondly, most of the big tech players
29:04
Speaker 1: are starting to explore significantly quantum. So if you look
29:09
Speaker 1: at Microsoft, Google, Amazon, they have big quantum teams. If
29:16
Speaker 1: you look at some of the other companies like Cisco
29:19
Speaker 1: and in Nvidia, they are investing significantly in quantum already.
29:24
Speaker 1: So if you want exposure on that, there are those
29:27
Speaker 1: You've got companies like IBM and Honeywell who actually have
29:31
Speaker 1: significant quantum machines already. IBM had definitely have very powerful
29:36
Speaker 1: quantum Honeywell is part owner of Quantinuum, which is, as
29:40
Speaker 1: I said, considered the leader. And then we have seen
29:44
Speaker 1: over the last four or five years a number of
29:46
Speaker 1: pure play quantum companies listing, particularly on Nasdaq. And so
29:51
Speaker 1: if you want to invest in the base layer, which
29:54
Speaker 1: is how you build the actual quantum compute, there are
30:00
Speaker 1: probably five or six other companies you can invest in
30:04
Speaker 1: as a retail investor, But the more interesting area is
30:08
Speaker 1: clearly the private markets, because the private markets are in
30:12
Speaker 1: the layer above the actual quantum cubit. They're the people
30:18
Speaker 1: that are building the connectivity, the networking. And if you
30:22
Speaker 1: consider quantum a little like electricity, the people building the
30:26
Speaker 1: machines are building the generation, the wind, the solar, the gas,
30:31
Speaker 1: the oil, et cetera, the ways of generating. But you've
30:35
Speaker 1: got to connect them together. It's no point just having
30:37
Speaker 1: a generation station in the middle of thewer So what
30:39
Speaker 1: we're going to see is as those companies start building
30:43
Speaker 1: powerful quantum computers. We're going to need to put them
30:46
Speaker 1: into cloud data centers. We're going to need to network
30:50
Speaker 1: them together. We're going to need to network them to
30:52
Speaker 1: the existing infrastructure we're going to need to store quantum data,
30:56
Speaker 1: and we can already see those starting to be built.
31:00
Speaker 1: That area, to me is the most investible at the moment,
31:05
Speaker 1: as as if you wanted to invest in space.
31:08
Speaker 2: And there are a variety of ETFs already, both in
31:11
Speaker 2: Europe and in the US that list themselves as quantum ETFs.
31:15
Speaker 2: Was that a reasonable way for an ordinary investor to
31:18
Speaker 2: get a little exposure to this space or is that
31:20
Speaker 2: just filled with junk?
31:22
Speaker 1: When I looked at them last six months ago, there
31:24
Speaker 1: were very few quantum companies in there. Now that will
31:27
Speaker 1: have changed a little bit because a portfolio approach, I
31:30
Speaker 1: think investing in the big tech companies and the companies
31:35
Speaker 1: that are pure play quantum companies as a portfolio approach
31:38
Speaker 1: is not a bad approach. You can probably build up
31:40
Speaker 1: ten fifteen positions. If there was one or two, I
31:42
Speaker 1: would say not. And I think you will see over
31:44
Speaker 1: the next year there will be much more interest in
31:49
Speaker 1: floating listing quantum companies, so that will be in the space,
31:53
Speaker 1: so that would be my preference. That's what I do.
31:56
Speaker 1: Some companies get acquired by the bigger companies. Ion Q,
31:59
Speaker 1: for instance, is acquiring lots of companies because they have
32:02
Speaker 1: something in there that they really need. So even the
32:06
Speaker 1: ones that are less sophisticated in the number of cubits
32:09
Speaker 1: or their customer base in the photonic space, for instance,
32:13
Speaker 1: some of those are going to be incredibly valuable because
32:16
Speaker 1: of their networking capabilities that these bigger companies are going
32:20
Speaker 1: to need, and because it's going to be this asymmetric advantage,
32:24
Speaker 1: a small time lead is incredibly valuable. So these companies,
32:29
Speaker 1: many of them will have enormous value and many of
32:31
Speaker 1: them won't. But it's quite difficult to determine analytically which
32:36
Speaker 1: ones those are going to be.
32:38
Speaker 2: What could go wrong with all this? It all says
32:40
Speaker 2: absolutely marvelous. You make me very excited about the future.
32:43
Speaker 2: What could go wrong here?
32:45
Speaker 1: The big challenge will be that there are scientific problems
32:49
Speaker 1: that come up that we cannot scale as fast as
32:53
Speaker 1: we want to, and therefore we have a massive delay
32:58
Speaker 1: in delivering these technology. I don't see that happening, but
33:02
Speaker 1: it's not unlikely. The second thing weirdly is that these
33:06
Speaker 1: companies become too big too quickly. I e. They don't
33:09
Speaker 1: go from fifty to two hundred, they go from fifty
33:13
Speaker 1: to two thousand or five thousand, and suddenly the regulators say,
33:18
Speaker 1: these systems are not going to be available to people.
33:21
Speaker 1: They're going to be military machines, and companies that really
33:26
Speaker 1: need access to them for specific things like material discovery
33:30
Speaker 1: or pharmaceutical discovery have to go through a firewall. They
33:35
Speaker 1: have to give us their software, and it grows slowly
33:38
Speaker 1: because of that.
33:39
Speaker 2: Isn't that why you might run into geopolitical differences in
33:42
Speaker 2: that you might have a regulatory system like that, for example,
33:46
Speaker 2: in the US, but you might not have one in China.
33:49
Speaker 1: That's true, but I suspect if China develop quantum computers
33:53
Speaker 1: of that scale, they're not going to release them in
33:55
Speaker 1: the commercial market anyway. It's not a great example, but
33:58
Speaker 1: it's an example that I find usefulness. If I've developed
34:02
Speaker 1: a nuclear bomb, I could sell it to lots of people,
34:04
Speaker 1: but I'm not going to.
34:06
Speaker 2: Okay, brilliant, right, Okay, my brain's about to explode.
34:10
Speaker 1: Daniel.
34:11
Speaker 2: What haven't we covered in this area that we should.
34:13
Speaker 1: Have anyone that has heard about quantum and doesn't look
34:17
Speaker 1: into how it impacts their operations or what they do,
34:21
Speaker 1: is doing themselves an injustice. And I think if you
34:24
Speaker 1: demystify quantum by saying it doesn't matter how my iPhone works,
34:29
Speaker 1: it doesn't matter how TCPIP works. No one says that.
34:33
Speaker 1: They say, how does it change what I'm doing? I
34:35
Speaker 1: think quantum's actually very approachable. There was a quote I
34:39
Speaker 1: heard in twenty twenty two from the deputy head of
34:45
Speaker 1: it for the US Air Force who said QDA is
34:49
Speaker 1: almost closer than it will take us the time to
34:53
Speaker 1: fix the problems that we have. And that is really
34:57
Speaker 1: how you need to look at it, which is there's
34:59
Speaker 1: an opportunity in a threat, the time to get ready
35:04
Speaker 1: for the opportunity and deal with the threat is significant,
35:09
Speaker 1: and therefore you need to look at it now and
35:12
Speaker 1: how it's going to impact you.
35:14
Speaker 2: Brilliant Daniel, I'm going to ask you Krito that I
35:16
Speaker 2: haven't after for a while. That's got a bit boring
35:18
Speaker 2: further the old fashioned, But I think we might be
35:20
Speaker 2: in the correct zone here, which is that given the
35:24
Speaker 2: choice over a ten year period, bitcoin or gold.
35:28
Speaker 1: Wow, I would probably say gold. Now the bitcoin industry
35:32
Speaker 1: has been built on hype. It went up to one
35:34
Speaker 1: hundred and twenty five thousand, it's now sitting around sixty.
35:36
Speaker 1: I'm still a holder of bitcoin. I sold most of
35:38
Speaker 1: my holdings in twenty twenty two, but it's going to
35:43
Speaker 1: require so much money into the system to start pushing
35:47
Speaker 1: it up to those numbers again. However, to be honest,
35:51
Speaker 1: I don't own any gold, and I do own bitcoin still,
35:54
Speaker 1: but it's it's a fraction of what I owned.
35:57
Speaker 2: We'll come back to any year or cera and see
35:58
Speaker 2: if that's changed, shall I Indeed? Final question, what are
36:02
Speaker 2: you reading at the moment?
36:04
Speaker 1: Actually, what I want to read is a lot more
36:05
Speaker 1: philosophy because of what I think quantum is going to do.
36:09
Speaker 2: Sorry now I was about to finish, but I can't
36:11
Speaker 2: leave on that. What is quantum going to do? That
36:13
Speaker 2: means we need to read more philosophy. Are we going
36:15
Speaker 2: to have one great existential crisis all of us in
36:18
Speaker 2: a honer from what we find out?
36:20
Speaker 1: This is such a great question. It's quantum unlocks because
36:25
Speaker 1: of the way operate unlocks in effect a new data
36:28
Speaker 1: set and a new understanding, which is quantum physics. As
36:32
Speaker 1: we do that some of the assumptions that we hold
36:38
Speaker 1: very dear, that's probably the one we're but without question
36:42
Speaker 1: suddenly get challenged. For instance, what is reality? How does
36:48
Speaker 1: the universe operate? Something that I heard Demis Hassebis talking
36:52
Speaker 1: about in December. I was surprised I heard it from
36:54
Speaker 1: him is maybe energy and matter are not the foundation
36:58
Speaker 1: of the universe. Maybe information is. These are the type
37:01
Speaker 1: of things that this technology is going to start allowing
37:05
Speaker 1: us to explore. Many of those are philosophical questions, and
37:10
Speaker 1: so I want to be much more knowledgeable about about
37:15
Speaker 1: what that means. I'm absolutely fascinated by the understanding that
37:21
Speaker 1: we may be unlocking again another layer of what it
37:26
Speaker 1: means to be human and what it means to be
37:28
Speaker 1: sitting in this universe.
37:30
Speaker 2: Daniel, thank you so much for joining us today. Thanks
37:39
Speaker 2: for listening to this week's Marin Talks Money. If you
37:41
Speaker 2: like us show, rate, review, and subscribe wherever you listen
37:44
Speaker 2: to podcasts. Also gives any questions or comments to Marry
37:47
Speaker 2: Money at Bloomberg dot net. It's an also following me
37:49
Speaker 2: and John on Twitter or exam at marys w and
37:51
Speaker 2: John is John underscorss depic. This episode was hosted by
37:55
Speaker 2: me Marren undset Web. It was produced by Some Society,
37:57
Speaker 2: Moses and Im and Jennifer ciliun by Mabel's and special
38:01
Speaker 2: thanks of course to Daniel don Steinberg m HM