Daniel Marin is tackling one of the internet’s trickiest challenges: validating a user’s identity without exposing sensitive information to hackers. His startup, Nexus, which has raised $27 million, built a product called Nexus zkVM, which uses a combination of cryptography, computer science, engineering and mathematics to prove you are you—through protective software as opposed to passwords or personal information. The software relies on a cluster of computers called the Nexus Network, which do the heavy lifting behind the scenes. “There has been an explosion of advancements on the scientific front, on the engineering front and most importantly on the market front,” he says. “This allows us to build a for-profit company [around] this technology.” zkVM, which is currently free, was released in June and is used by more than 120,000 people and 15 companies.
Subscribe to FORBES: https://www.youtube.com/user/Forbes?sub_confirmation=1
Fuel your success with Forbes. Gain unlimited access to premium journalism, including breaking news, groundbreaking in-depth reported stories, daily digests and more. Plus, members get a front-row seat at members-only events with leading thinkers and doers, access to premium video that can help you get ahead, an ad-light experience, early access to select products including NFT drops and more:
https://account.forbes.com/membership/?utm_source=youtube&utm_medium=display&utm_campaign=growth_non-sub_paid_subscribe_ytdescript
Stay Connected
Forbes newsletters: https://newsletters.editorial.forbes.com
Forbes on Facebook: http://fb.com/forbes
Forbes Video on Twitter: http://www.twitter.com/forbes
Forbes Video on Instagram: http://instagram.com/forbes
More From Forbes: http://forbes.com
Forbes covers the intersection of entrepreneurship, wealth, technology, business and lifestyle with a focus on people and success.
Subscribe to FORBES: https://www.youtube.com/user/Forbes?sub_confirmation=1
Fuel your success with Forbes. Gain unlimited access to premium journalism, including breaking news, groundbreaking in-depth reported stories, daily digests and more. Plus, members get a front-row seat at members-only events with leading thinkers and doers, access to premium video that can help you get ahead, an ad-light experience, early access to select products including NFT drops and more:
https://account.forbes.com/membership/?utm_source=youtube&utm_medium=display&utm_campaign=growth_non-sub_paid_subscribe_ytdescript
Stay Connected
Forbes newsletters: https://newsletters.editorial.forbes.com
Forbes on Facebook: http://fb.com/forbes
Forbes Video on Twitter: http://www.twitter.com/forbes
Forbes Video on Instagram: http://instagram.com/forbes
More From Forbes: http://forbes.com
Forbes covers the intersection of entrepreneurship, wealth, technology, business and lifestyle with a focus on people and success.
Category
🛠️
LifestyleTranscript
00:00There have been an explosion of advancements in the science, engineering, and economic forces.
00:05So this allows us to combine these three to build a for-profit company that productionizes this technology.
00:16I'm here with Daniel Marin from Nexus. Daniel, thanks for joining us.
00:19Thank you, Steve.
00:20So give me a 30-second pitch, who you are and what you do.
00:24Well, I'm the founder and CEO of Nexus.
00:27Nexus is a company developing bare-file computation and bringing it to life.
00:32So I did my research and I went to the Nexus white paper, and you have to help me here.
00:36So it says, Nexus is a zero-knowledge virtual machine.
00:40It's modular, extensible, open-source, highly parallel... what is it? Parallelized?
00:46I can't even say it.
00:47Parallelized.
00:48What you said, provider-optimized, contributor-friendly, written in Rust.
00:52Exactly.
00:53What are you talking about, man?
00:55Pretend I am a 10-year-old and explain to me what you built and what Nexus does.
01:00Yeah, so Nexus is the company behind the Nexus project,
01:04or at least what we call the Nexus project.
01:06The Nexus project is an attempt to bring this new type of computation to life.
01:12And what you just described is what we call the Nexus zero-knowledge virtual machine.
01:17Okay.
01:17Which is a machine that can prove any computation.
01:21And this is something that scientists have wanted to build for decades.
01:25Okay.
01:25And our whole mission is to bring this to life.
01:28Cool. Well, let's get practical.
01:29What is the problem out there and what does your machine hope to solve?
01:36The problem is, how do we bring truth to the field of computation and the internet?
01:42And this machine uses mathematics to prove arbitrary statements about things.
01:48So you can prove to anyone, anything.
01:53And this helps bring truth, privacy, and transparency to the whole internet.
01:58Give me a real-world practical example of how this is in action.
02:03Yeah.
02:03So for example, if you want to prove some properties about your Google account
02:08or your identity without revealing private data,
02:12you can do it with this machine and with this technology.
02:15And what this machine outputs is quite literally a proof
02:20along with the output of a computation.
02:22So is this like a password killer?
02:25Well, it's more powerful than that
02:27because it works for any type of computation that you want to do.
02:30So how does it prove?
02:31Like, I want to prove that it's me, right?
02:34Logging into my Forbes subscription account that everyone else has.
02:39Yeah.
02:39I forget my password or my pet's name or my first grade teacher's middle name, whatever it is.
02:45What does Nexus do?
02:46What does it help me log in to Forbes?
02:49So Forbes knows I'm a real user, I know it's Forbes, and I get this subscription.
02:54Yeah, yeah.
02:55For example, you can prove to me anything about your Forbes account
02:59without having to reveal to me your password, right?
03:02Which is something that you would not want to do.
03:04How do I prove it?
03:05So you run a program inside the machine that does some computations on private data.
03:11And you can do this for something like login technology,
03:14but also like AI, for example, verifiable AI.
03:18If you want to evaluate a machine learning model on private inputs,
03:21you can do it with the machine.
03:23Or in general, anything that requires privacy or verifiability.
03:26So I have this program on my machine and I log into that thing once
03:30and then it's like a skeleton key that opens up all my accounts
03:34and all my access and that sort of thing?
03:36Well, you do that on your end, but the machine generates a proof.
03:40This proof is a tiny string of bytes that allows you to prove to anyone
03:46that what you claim to be true is indeed true.
03:49How'd you come up with this wild idea?
03:52Well, we didn't come up with this.
03:54This is the field of verifiable computation,
03:57which is a field of computer science and more broadly, cryptography.
04:01The problem is this technology has never been able to be productionized.
04:06And what we have done is to bring together a lot of research from academia,
04:10advancements in engineering,
04:12combined with other advancements on the economics part of things,
04:16combine the three of those and actually productionize this technology.
04:21How'd you do it?
04:22This sounds like a problem everyone's like a holy grail in cryptography.
04:26It is.
04:27And you're how old are you?
04:2823.
04:2923, young buck coming out of Stanford.
04:33And what is your solve?
04:37How are you cracking this code where many others have tried in the past?
04:42That's an excellent question.
04:43So many others have tried in the past, as you said, but the timing has not been right.
04:48Only until now, decades later into research,
04:52there have been an explosion of advancements on the scientific front,
04:57on the engineering front, and most importantly,
05:00on the market front, where there exists a large amount of demand for producing these proofs.
05:05So this allows us to combine these three,
05:08science, engineering, and economic forces to build a for-profit company
05:13that productionizes this technology and brings the best of the three together.
05:17Yeah, is Nexus a company?
05:19Is it a collective?
05:20Is it a commune?
05:21What do we got here?
05:21Nexus is a startup, a full-on company.
05:25And our approach is to be as fast as possible in developing this technology
05:30and pushing it out to market.
05:32How many employees do you have?
05:33We have 21.
05:3421. And how much have you raised?
05:3525 million dollars.
05:37Great. And what was the pitch? How did you get that money?
05:40Well, you know, this was our two rounds,
05:42the seed round and the series A that recently happened.
05:45The seed round happened a year and a half ago, almost two years.
05:48The series A was actually fairly easy to do.
05:51There was a lot of demand from investors to invest in the company.
05:55Reason being, it is clear in the industry that the demand for these proofs
06:02is growing up exponentially.
06:04And so the question is,
06:05which will be the team that will be able to capture this market?
06:08OK.
06:09And thankfully, in our team, we have pretty much some of the people
06:13that invented serial knowledge cryptography itself in the team
06:16and many other forms of cryptography and mathematics.
06:19So we just have a really good team that is tackling a really, really big problem.
06:24Yeah. This sounds very advanced, very technical.
06:28How did you recruit these experts in a very, obviously, specialized field?
06:33What was your tips and tricks to recruit the best of this mind, brain trust here?
06:38Yes. Well, great scientists are not super excited about funding or investors.
06:45They're excited about working with other great scientists.
06:48OK.
06:48Right. So the number one strategy for our team has always bring class A players.
06:55So that means it's not settled down for like B or C players,
06:58but always having the best of the best.
07:00And that's exactly been our approach is compounding effects.
07:03Once you bring someone that is really, really good,
07:05the next person that you can bring can be really, really good as well.
07:09Cool. How did you come up with the idea for Nexus?
07:11And how did you get it started?
07:13Well, I actually studied cryptography at Stanford.
07:16OK.
07:17And so verifiable computation is the holy grail, as you said, of cryptography.
07:22So the timing was right and the industry was ripe for disruption, as they say.
07:28And so it was obvious to me that someone was going to do it.
07:31Did you start it right out of college?
07:33Actually, yeah, exactly. During college. I think it was junior year.
07:37So you've never had a job except owning and running a company?
07:40Well, I had an internship at Google.
07:42OK. Very cool. What did you know?
07:44So this is obviously very scientific.
07:46How have you learned about fundraising?
07:48How have you learned about running a company when you've really only had an internship?
07:53How do you be a boss when you've never had a boss, more or less?
07:55Well, you have to learn it on the go.
07:57And there's many things that you can't learn unless you do it.
08:02For example, recruiting, pitching, managing problems inside and outside,
08:07communicating effectively, being able to see who is a team player and who isn't,
08:13and many other things.
08:14Gotcha. Do you have a product yet?
08:16Oh, yes. And it's open source. We're on our second iteration,
08:19and the third version of the machine is coming out soon.
08:22How do you make money?
08:23Well, so we actually pay people for idle compute,
08:29which is something quite interesting about the project,
08:31because we can use any compute to help parallelize proof generation.
08:37And so we're building a marketplace for compute.
08:40Essentially, you can lend us your compute, and we pay you,
08:44and you do some work for us, your computer, your machine does some work for us.
08:48And we compute these proofs and sell it to the people that want these proofs to be generated,
08:53which are highly specialized players.
08:55So give me an example. What sort of clients do you have?
08:58Could you name some customers?
08:59Yeah. So these are usually experimental projects.
09:03You can see, for example, on our Telegram channel and in our GitHub repo.
09:08These are projects that fall into two categories.
09:11Essentially, people that are trying to experiment with real-world computations,
09:15like AI and login and such things, and blockchain clients,
09:21because verifiable computation is well known to be
09:24the only long-term sustainable solution for blockchain scalability.
09:28So verifiable computation is really general. You can do a lot of things for it.
09:33We essentially have a general-purpose platform that can serve any of these clients,
09:38and they can use it for free.
09:40So you can use it right now.
09:42We plan to monetize it later on through many other things.
09:45So you're not making really money yet.
09:47It's really you're kind of proving it out and stuff.
09:49That's not our focus.
09:50We have some really clear plans to how to do that, but it will come eventually.
09:57Well, you're on the cutting edge of tech, and obviously everyone's talking about AI.
10:01There's various levels of hype along the hype spectrum.
10:05What is your take on AI right now,
10:07and what do you expect to happen in the next few months, few years?
10:10Yeah, well, AI, our project is actually closely related to AI on two fronts.
10:17One is it brings a new property to AI systems, which is verifiability and privacy.
10:24So without a CKBM like ours, it's virtually impossible to have trustworthy AI agents.
10:31That's number one.
10:31That's really exciting.
10:33This is a new property that we have not seen in this space.
10:36Yeah, big growing, big growing segment.
10:37Exactly, exactly.
10:39The other one is just the advancements on compute, on GPUs, FPGAs,
10:44and just the rate at which we can do high-performance compute.
10:48It happens that producing these proofs consumes a lot of computing power.
10:53Okay, a lot of energy too.
10:54And a lot of energy.
10:55So all of the advancements that we're seeing on GPU architectures
10:59actually accelerate the production of proofs, which benefits us as well.
11:04What are you seeing in the world of tech, in the world of security that you live?
11:07What's kind of your big bet here?
11:10Well, verifiable computation, as we are introducing it to the world,
11:14is pretty expensive in terms of money, so to say, that you have to pay.
11:20But we expect it to be reduced, the price or the costs, by five orders of magnitude.
11:27That is like a million x.
11:29And so the goal of the project is how can we combine advancements in science and engineering
11:36to reduce the cost of producing proofs by 10x, then by 10x, then by 10x, so on and so forth.
11:43And answering your question, what we predict is that
11:46we will get to a point where it's so cheap within the next couple of years,
11:50well, no one knows what will happen there.
11:52We call this the proof singularity.
11:55Gotcha. Well, I'm still, this whole proof thing is over my head still.
11:57I'm going to admit, I'm still having trouble grasping it.
12:00If this all comes to fruition, like your machines are working, the cost is down,
12:04like this is happening in a couple of years, what does life look like?
12:08What is like, how is, for the everyday user like me, how does my digital life change
12:14if this is all like, you have a magic wand and it's exactly how you want it to be?
12:18Yes. Well, number one, the internet is more secure.
12:21Okay.
12:22And there exists private computations.
12:25And so people are able to have agency over their data, over their identity,
12:31and people are able to trust each other.
12:33And AI models will be able to trust each other.
12:35Why?
12:36Because they exchange proofs, mathematical proofs that prove mathematically
12:41that what we tell each other is true.
12:44So my hope is that this will allow humans and machines to collaborate better
12:50by trusting in math and not, you know, on other things.
12:54Daniel Merritt, thanks for joining us.
12:56Thank you, Steve.