Building a Quantum Computing startup with Chad Rigetti (Rigetti Computing) | Disrupt SF (Day 3)
5:37PM Sep 7, 2018
We're going to go back to the topic of quantum computing. Last we heard from IBM, they're the incumbent. This time, we're going to hear from the up and coming competition. So with that I'm going to bring them on to the stage. Please give a warm welcome to Chad Rigetti from Rigetti computing and your moderator, Frederick Lardinois.
Alright, now that we have a crowd full of quantum physicists here, we don't need to do any big introductions, which helps a lot since everybody was here for the earlier session as well, I think.
Chad, you're building a quantum computing startup. Now just earlier on IBM was telling us how they've been working on this in some form or another for 40, 50 years since the 70s.
What in the hell possessed you to build quantum computers?
Well, look, I've been working on quantum computing for most of my adult life. I got obsessed with this technology when I was a junior in college, and it's just been, I find it to be the most fascinating and alluring technology that is out there in the world today.
And I saw an opportunity really to build a company from the ground up to really solve the central problems that need to be solved to make quantum computing truly useful for the world. And, there's very few things that I could otherwise identify that I thought would have such a positive meaningful impact on the world outside of quantum computing. And I I couldn't resist the opportunity.
So you had no interest in just building a mobile game and become a millionaire?
Well, honestly, I always wanted to synthesize my my aspiration, you know, my I was a physicist and scientist for a long time and I always wanted to synthesize that with an entrepreneurial career, and the opportunity to build Rigetti Computing was was was that that was a resolution of that of that tension that I had held for a long time.
Totally. What did that first step look like? You were at IBM for a while? I think you're working on quantum there. Yes. And then you decided I'm out of here.
So yeah, I spent seven years at Yale working with one one of the best groups in the in the world on quantum quantum computing. I was at IBM Research for about two and half or three years. And I at one point, I just had to take take the plunge. I, you know, I was living in New York City and I moved out to San Francisco. I got an apartment in Berkeley. In fact, I got an apartment in Berkeley.
And I had, you know, I'd saved up enough in my 401k that I could support myself for a year. So basically, as a 34, 35 year old I went back to living on less than I was making as a grad student. And I just started chipping away at the problem of building the company, figuring out how to put one foot in front of the other in terms of building up a team raising some capital.
At one point. In fact, I even had to move home to Saskatchewan. And I moved in with my mom for three or four months.
Saskatchewan is nice and cold. It's good for quantum computers. Right?
Well, it was the summer and Saskatchewan's uninhabitable for the most part of the year. But May, June July is quite beautiful.
I can imagine now, what did those first efforts look like? Because you don't just go and say, I'm just going to buy some quantum computing parts.
Yeah, look, there's a lot of challenge challenges to be solved. And we are a full stack company, which means we need to figure out how we're going to design and manufacturer quantum chips, need to figure out how we're going to integrate those systems into real computing environment, how to build the software, how to build the applications, and eventually how to really engage with customers to deliver those machines to unlock real value for customers.
And the first thing we set out to do was really to prove out that a chip design that could scale to the number of qubits needed to eventually achieve quantum advantage was possible. And that had been one of the things that that I was deeply interested in. And, and it turned out that, actually negotiated with a company called Ansys that makes 3d simulation software to get a discounted license of their of their 3d simulation software. So I could prove out some basic chip designs.
And, and that was something I could do kind of as a sole founder in my, you know, in from my apartment. And with that just kind of license we were able to get a little bit of momentum behind the chip designs show that it was possible and then just start to build up a little bit of momentum on fundraising. And in fact, my brother was the very first investor into the company and so it's it's been a long road it's been one small step after another, and now, five years later, we're you know, we're we're one of the main players in the race globally.
Absolutely. Are you still using a discounted license?
Well, I think the discounts have run out as well as our, yeah, so.
What was it like going out trying to raise funding for something like this because I'm venture capitalist probably had no idea what you were talking about.
So this was 2013, 2014, the quantum computing at that time when I started Rigetti computing was largely a field of physics research. And we needed to basically change what quantum computing was from a field of physics to a field of computing. And that is, so there's a lot of communication challenges. There's a lot of,
There, you know, I think for the founders in the room, or the aspiring founders who are watching I think the key thing to focus on is just find ways to generate momentum, find ways to add momentum into the flywheel that is your company in your, in your venture.
And raising capital at the time was what you know, it always goes in stages. I referenced it kind of to the process of making movies, where you get maybe maybe your first movie, you get $25,000, and you make it with a handheld camera. And if it does, well, the next time you get a few hundred thousand if it you know, and as you as you and that's what raising fundraising rounds is like. So raise a little bit of money. We showed some progress. We raised a little bit more ultimately one of the big things was
moving out here getting introduced to Sam altman at Y Combinator going through the Y Combinator program. It ultimately gave us a lot of momentum and really introduced us into broader networks. And we were able to then I had that I knew otherwise and really made a lot of great connections. And ultimately, you know, some of those investors had been with the company for four or five years now are we have amazing long term relationships and we really have an incredible cohort within the company.
Nice. What, while you are going through those programs that you already have anything physical to show?
In Weiss, when we run through yc, excuse me, only gone through yc, the company was three people. And we were working on building the first prototype of the control system, not even the quantum chip, right? Just the control system. And to prove out that we could build is one of the things we need to show is that we could lower the cost of building a prototype down from something like five or 10 million down to a couple of million. And so we're working on building the first prototype of the control system while working on the designs for the manufacturing plant of the chips themselves.
That must have with it's scary, did you was there a moment where you thought we might not be able to do this?
I've always had a deep confidence and conviction that the company and you know my vision for the company is going to manifest in the world and and we're going to build it as you know, and we're gonna, we're gonna build a great and lasting technology company moment to moment. And day by day, you're figuring things out as you go. And I,
when you're starting to venture like that, and yes, it's scary, but what you need to focus on is kind of shortening the learning loop. So you go through as a company, as a founder, as a small team at the start to because basically, you know, you have some money, you need to demonstrate a certain amount of progress and get a certain amount of money to get to the next round and to get to the next milestone. And the key thing is just like learning as rapidly as possible, because when you start you don't know how to do it in by the end, you will have done it in the key question has to just learn as quickly as you can, as a team.
at this point, you've got you've got something to show you've got working quantum computers.
So yeah, as as I said, when when, when Rigetti when I founded Rgetti Computing in 2013 Quantum was a field contributions a field of research and now the reality of the field is that are of the industry is that we have built real general purpose programmable quantum computers. We have been continuously operating systems with between eight and 19 cubits. Since December of last year.
Users have have used those machines from more than 30 different countries. They've run 90 million different programs. And I believe there's been about 20 different scientific papers published based on use of the platform and exploring the algorithms and the applications to that may run on these machines.
So we have real computers and we end and quantum computing is now has moved out of a field of research into an engineering discipline and engineering enterprise. And we're now really focused on kind of what we think is going to define the industry over the next five years is is the pursuit of quantum advantage and the way we think of advantage
is that this is when you demonstrate something using a quantum computer in generating a solution or a result that is better, faster or cheaper than you can do otherwise. So it's defined relative to a customer requirement in a customer need. And ultimately, that is where quantum joining us today, people are built quantum computers. But no one has ever used a quantum computer to do something better than you can do with the classical machine.
And that's with the systems you have today. That's probably not possible, right.
It's not possible with 19 qubits. It's not possible with 50 we don't believe it's possible 50 there's a pretty strong body of evidence is not going to be possible 50 or even 60 or 70.
Cause a lot of people talk about quantum supremacy and 50 cubits or is about a quantum advantage. People use it interchangeably, but that's the kind of the turning point but if you don't say.
I think it's going to take more I think it is and supremacy is not necessarily the same thing as advantage advantage really focuses on delivering value to a customer and supremacy is a constructed problem meant to kind of it's kind of a drag race for a quantum computer to show specifically constructed to show the power of a computer or the capability of a computer for a specific problem.
And I think advantage is going to take probably 100, 120 cubits. And so we recently announced our plans to deploy 128 qubit system next year as part of our march towards quantum advantage.
When part of that module two involves an announcement you have today absolutely a new platform.
That's right so. Look we what in order to achieve quantum advantage What is needed is more cubits with lower with lower error rates and 128 qubit systems will address this. But there's a couple other things that are needed to demonstrate advantage as well. The first is we need integrated computing systems. We don't need chips on test boards that people can tinker with over the cloud. We need to integrated true computing systems to run the hybrid algorithms that offer the shortest path to quantum advantage.
Talk about this hybrid part there for a moment.
Sure. So modern contemporary quantum algorithms are all based on using both classical and quantum computers as part of a integrated computing system. And until, you know, this is a major part of our announcement today, until we've built and delivered this platform, though, that architecture didn't exist in the world. And we're delivering it now for the first time through our quantum cloud services offering. Now the The third thing that people need, it's not just the chips. It's not just the integrated computing system, but people need a dedicated programming environment in a software development kit delivered alongside those systems to actually build quantum software applications.
And so today, we're announcing for the very first time quantum cloud services which is the first and the only what I think of as kind of a quantum first cloud computing platform. And I believe we have a we have an animation of this.
that describes a little bit of how it works. So this is to demystify this. And to bring a little bit of lightness and simplicity to it. Basically a user logs on from their remote from their own device on their laptop, or their workstation. They use our software development kit called forest to write a quantum application or quantum program that Apple that program is sent to a one time compiler, which generates a set of instructions and kicks off and optimization loop that runs between both of classical and quantum computer. Ultimately, this returns a result to the user that is better than what you the user could obtain using either the classical or the quantum in isolation.
This is the architecture that is needed to achieve quantum advantage. And it's only available on Rigetti Quantum cloud services.
And that loop is really what it's all about, right? Because you offered an API before the give access to your quantum machines.
That's right and you know as I said since December and and and other other new IBM has, has allowed people to tinker with and play with their systems remotely over the cloud. This has been a, this has been kind of like offering a quantum chip on a chessboard, you know, and along with someone remotely to access that and play with it. But those chips haven't yet been integrated until today into a true computing system. Ultimately,
What we're building are the most powerful. Our mission is to build the most powerful computers in the world and to help solve humanity's most important and pressing problems. That requires real computers, not just quantum chips. And that's what we're offering with quantum cloud services.
Is that open to every developer today?
So you can sign up for the waitlist today to get a spot on quantum cloud services. Over the next few weeks, we're going to be opening up the platform to early partners to early developers who will be helping us test and test the platform. So we're entering the beta, you know, beta testing of the platform now, and we expect to open it up more broadly for public use near the end of the year, early next year.
Okay. What does it look like trying to program for one of these, these platform for your platform?
Yeah, so our environment is a Python based environment. It's actually very easy to learn, we have the aspiration, that the easiest way to learn programming is by learning quantum programming through forest or software development kit. And the tools themselves are quite easy to learn. They allow you develop an intuition for how to harness quantum computing, but it is truly a developer environment for building real quantum software and not just for tinkering with the physics
How hard is do I still need to develop some kind of intuitive I mean, that's quantum computer is not intuitive to begin with. It does need to get to that point. If I want to build a quantum program.
I think that we've seen pretty broad adoption and broad accessibility of the tools if you will, people who don't have a background in quantum physics are building interesting applications on forest and on UCS. In fact, as part of this, another thing we're announcing today is our very first cohort of developer partners. So we've partnered with 12 of the leading startups in the world, some of the best technical teams that are building applications for quantum computing. And their applications are going to be available and distributed over our quantum cloud services platform. So ultimately, what this allows us to do is to engage with a broader ecosystem of incredible technical teams
who are bringing domain specific focus from, for example, genomics or protein full or you know, or protein based therapeutics or computational chemistry and bringing that domain that domain expertise, combining it with algorithms, expertise, and building applications on on quantum cloud services, which then allows the end users to both accesses, libraries and applications and shortens a path to advantage for everyone within the ecosystem.
You're putting some money on the line as well.
We are so I'm super excited to to announce the the rigetti Quantum advantage prize we are giving $1 million to the first team or the first individual that is able to conclusively demonstrate quantum advantage on quantum cloud services.
Has to be on your platform.
It has to be on it has to be on quantum cloud services, we we strongly suspect it will be the very first platform to get there and we're going to announce more details of this on October 30, including the panelists who will be adjudicating the prize the details of the qualification criteria and we're super excited that quantum computing is now entering this phase where quantum advantage is is the focus of the industry rather than just more cubits or better cubits and and different chips.
If you had to guess today not who was going to win the prize but what is going to win the prize? What What will that look like?
So what I love about the prize is that we're not betting on our own opinion on the matter of what will be the first demonstration of quantum advantage. we're empowering the community but you must have a guest I have a guest and in fact my guest is different than some of our application researchers who are the best people in the world at what they do. I believe that quantum machine learning could be one of the very first application areas but here's what I'll say there's there's there's three different there's three different areas where it may come one is quantum machine learning and this is a wild card because it's very hard to benchmark things through a machine learning approach and also know we're talking about beating classical computing and classical traditional machine learning is evolving so fast it's really hard to know what machine learning is a wild card the most likely culprits are the most likely candidates if you will, is really the field of computational chemistry.
In in computational chemistry we're really doing is simulating the behavior of molecules that are implicitly quantum mechanical device or mall implicitly quantum mechanical themselves in those problems natively map onto a quantum processor. So quantum, so computational chemistry is probably the most likely.
And there's a third category of problems, which are called optimization problems. And this is something where, again, we know a little bit less about it a little bit less mature. But there is a real chance that could be one of the first demonstrations. Now, I think the things I'm most excited about in terms of the pursuit of quantum advantage are also the things that people haven't thought of yet. Because,
Look, quantum computing eventually is going to redefine almost every industry it's going to touch every market much like AI is today, if you fast forward or machine learning and AI today, if you fast forward 10 or 15 years, that's where quantum is going to be where every major company has an effort. There's going to be a small number of providers of the actual compute capability and the platform but ultimately, every major company will be involved in some capacity in that space.
Well, let's fast forward 10, 15 years we're probably, well, are we going to have little quantum computers at home?
That's a good question. In a cloud first era, I'm not sure that the economic forces are going to be there that will drive us towards developing that you kind of miniaturized and the room temperature technology needed to deliver quantum computing literally in your laptop when you can access it remotely over the cloud.
Is there still a chance that this this whole industry will fail?
I don't think so. There's always there's always a chance I expect the quantum may may find itself through a through a brief winter, if you will, we don't know how long it's going to take to get to quantum advantage. We don't know if it's going to take 128 kilobits or maybe 200 but ultimately, that's a matter of a year or two. I think inevitably quantum computing is going to change the world and it's all going to come in our lifetime whether advantages two years or five years. Look that that that's the exciting an open question.
All right. Well, let's meet somewhere in between two and five years and talk about that.
Awesome. I'd love to. Thank you.