Sid Gandhi, CEO of Polybase, on ZK Proofs and Decentralized Databases | Ep. 253

In an exclusive interview with cryptonews.com, Sid Gandhi, CEO of Polybase, talks about building products for Cruise, US Navy & Apple, the story of Polybase, and leveraging ZK proofs.

About Sid Gandhi

Sid Gandhi is a technology leader, entrepreneur, and investor, currently CEO of web3 infrastructure company Polybase.

In the past, he led product and technical strategy for simulation at Cruise, co-founded Hmatix Cybersecurity, built AR tools for the US Navy at Moback, and wrote software for iOS at Apple.

Sid Gandhi gave a wide-ranging exclusive interview which you can see below, and we are happy for you to use it for publication, provided there is a credit to www.cryptonews.com. 

Highlights Of The Interview

• Story of Polybase
• Polybase Modular Architecture
• Polybase Private mainnet, including a production-level environment for projects to get onboarded
• How is Polybase different from Ceramic & Tableland
• How does Polybase leverage ZK proofs
• Building products for Cruise, US Navy & Apple

 

 

 

Full Transcript Of The Interview

Matt Zahab 
Ladies and gentlemen, welcome back to the Cryptonews Podcast. It’s your host Matt Zahab. We are buzzing as always and I’m super pumped to have today’s guest on the show. Today we have Sid Gandhi, a tech leader, entrepreneur and investor who is currently serving as CEO of Web3 infrastructure company, Polybase. In the past, Sid led product and technical strategy for simulation at Cruise, Co-Founded Hmatix Cybersecurity, built AR tools for the US Navy at Moback, and wrote software for iOS at Apple, quite the bio, super pumped to get into all this stuff. Sid, welcome to the show, my friend, pumped to have you on. 

Sid Gandhi 
Thank you, pumped to be here. 

Matt Zahab 
Absolutely love that. Let’s show him in your pass here for a second. I’m going to go in order, but you worked from some incredible companies. Obviously, I feel like most of us have heard of Cruise, which is a self -driving startup everywhere in San Fran. I believe I’ve yet to see one myself, but then again, I have been to San Fran and I believe we don’t have those in candy yet. I’ve seen a couple of Google cars ripping around. Obviously, there’s testers everywhere, but no Cruises yet. You also Co-Founded Hmatix, which we’ll get into, built AR tools for the US Navy, so friggin’ cool and worked for Apple. Let’s start with Cruise. What did you and the team build there? How cool was it working at Cruise? Any cool stories, interesting stories, funny stories? What do you got for us from your time at Cruise? 

Sid Gandhi 
Yeah, I think the story is the first time I sat in a self-driving car. Honestly, really good technology. We were building at Cruise, was really advanced. At the time I was working there, during those two years, I think we basically went from performing just below the Waymo cars to surpassing the Waymo cars ability. And this is through our independent studies and stuff, but that was really cool to see the car improve that much. The first ride I went in was pretty smooth, except for a couple of times where we almost crashed. But I think that’s the case of a self-driving car or anyone I’ve talked to that’s been in one. And at the end of it, I summarized it by basically saying it’s like driving with a teenager that just learned how to drive. That’s kind of where it is. I think it’s improving every day, so I’m excited to kind of go back to that next time I’m there and test out some of the cars since they’re generally available now to the public. 

Matt Zahab 
How far away are we if you were to throw the hot take here? We’ll already get into some forward-looking statements. How far away are we as let’s just specify North American, more specifically American society from having self-driving cars. Because I mean, me for one, anytime I’m in a Tesla with a friend or an Uber driver, I think it’s so freaking cool just having your left hand on the bottom of the steering wheel, barely paying attention. Some friends may or not be texting while they’re driving. I think it’s an absolute treat. Obviously, you still got to have your eyes on the road, but so cool. How far away are we until this is like a good chunk of the cars on the roads are self-driving? 

Sid Gandhi 
Probably like five years or so. I mean, I think all the other car manufacturers are starting to slowly add in these features. I think Tesla’s kind of the company and Elon Musk is kind of the founder that he just goes max on everything that he does. But I think the other car companies are taking a slower approach. GM with SuperCruise, I think all the other ones have their own moniker for them. But I think in about five years, it’ll be pretty standard to just take your hand off the steering wheel and then kind of drive around. Maybe for legal reasons, they’ll have you keep the handle on the wheel, but technology wise and capability wise, I think that’s where it will be. 

Matt Zahab 
And when you were building the tech itself, you were obviously you led product and technical strategy for simulation. How does one even get a background in that? Like how does one figure out how to help develop and build those algos and all the objects on the road so the AI can quantify what is and what isn’t an object that should be surpassed or moved or not obstructed. Like how do you even get into that field? 

Sid Gandhi 
Yeah. So starting my career, I went into software engineering. So got a really solid foundation of a lot of different technologies across the board, although I’m from kind of low level framework stuff to UI. And then I transitioned into product management and I would call it like technical product management. So kind of really understanding the product on a technical level and being able to understand its capabilities, but also being able to understand its limitations. And that I think that broad technology background plus understanding those two things gave me a really good position to kind of set the direction of where we’re going. Specifically at Cruise, I was working on the simulation team. So we would actually build these like city wide simulations. We’d actually generatively build a 3D city from scratch and then drive our vehicles with the latest simulated software that developers had just released internally in those simulations. And those simulations were crazy. We would model all the lights. We had some some person on the team go out into San Francisco and do light sampling of all the different street lights in San Francisco. And we did that for a lot of different things, sidewalks and things like that to get the most accurate representation. And then we would drive the car there, get data back. And then if there were any big issues, we’d correct those before putting that software on our test cars on the road. And so it was all about safety. It was all about making sure, you know, there weren’t any big regressions that got out onto the road that could hurt people. 

Matt Zahab 
It’s wild. That’s so cool. And what about building AR tools for the US Navy? That must have been wild as well. I’m sure there’s a lot of things you cannot tell me, but of the things you can tell me, what did you work on there? 

Sid Gandhi 
Yeah, I think the coolest story was I built this remote assistance application. So it was kind of like FaceTime, except that you could kind of dive into the other person’s world and like put objects in it or point out specific things or direct them or navigate them in a space. And the purpose of that was to allow technical issues on Navy ships to be diagnosed and fixed without having to have civilian personnel technicians fly out onto carriers or battleships, etc. So it’s obviously a really big cost savings, but it’s also like when you’re in the middle of an operation, there’s civilian like danger as well. And so not having to do that, even 10% of the time, an issue comes up is a big deal. So my kind of the exciting story there is I got to actually go on to a battleship. That was nonactive. Of course, it was all stocked, but I went there. We talked to a lot of the sailors on the ship, demoed this application to them on a specific rack of servers that they had inside of that ship and kind of told them about what we were working on. So that was really cool to hear their thoughts and their perspective on it. You know, there are folks that just need to get things done. So anything that gets in the way of solving the problem, getting things done, you don’t want like a janky UI that you have to like click through and it’s confusing. You don’t want something that’s like, you know, breaking or loading or spinning. There’s a whole different kind of mentality that you have to develop software for applications like that. 

Matt Zahab 
So cool. And was that through like traditional AR goggles? Like was it, you know, thrown your AR vision or goggles and then you get to see and fix the problem? Like what did it look like? 

Sid Gandhi 
Yeah. So we’re starting on smartphones first because those are kind of like easily available. You can really translate all the AR stuff you build for smartphones into like the headsets very easily. So that’s kind of what we’re demoing on. And honestly, there weren’t any really good headsets at that time. And there are any even now that work in those industrial conditions like work. But there’s companies working on those and building those. 

Matt Zahab 
So cool. And let’s jump over to Polybase here again. You whatever there’s an opportunity to have someone on the show like read your bio. I’m like, wow, this is incredible. Built and worked for so many cool companies built so much cool shit at these companies. And then I always wonder why jump to crypto? Like, you know, you worked for Apple, the biggest company in the world. You worked for the US Navy, you built a company, you worked at Cruise. Like you’ve done almost everything. Like it’s pretty darn cool. And then you go to crypto. So I have to ask, what’s the story of Polybase? Why jump ship and start Polybase? 

Sid Gandhi 
Yeah. I think in general, when you look at the stuff I’ve done, it’s like all about what is the coolest, most interesting, most advanced thing that people are building? Like at the heart of it, I love learning. And so jumping into all these different areas from like cybersecurity to autonomous vehicles to augmented reality. It’s all about coming into a new field, learning as much as I can about it and being able to create some value and push that field forward. And, you know, there’s people that have been working in all these various fields for many years on the research side. What I love is taking that research and productizing it, finding a customer, finding a user and delivering that end value to them. Because that bridge is sometimes really hard to cross from research to product. So 2012 was when I first heard about Bitcoin, first read the white paper, fast forward a couple of years, I read the Ethereum white paper, was always really interested in the technology. Always felt like it was very innovative. But throughout the years as I’ve been following it, I haven’t been particularly excited about any of those end products or use cases. I think stablecoins were super exciting to me. Some parts of DeFi, but other than that, from a technology perspective, I wasn’t really stunned. And so, you know, somewhere of last year, after talking to a bunch of folks in the crypto, in the Web3 industry, I basically found a big gap where folks weren’t able to build the kinds of applications that they wanted because blockchains were fundamentally not scalable, too slow and too expensive. And so to give you a couple of ideas of what those kinds of applications are, it’s decentralized social, it’s advanced oracles, it’s DAO tooling, it’s different kinds of like SaaS applications that people want to build in a Web3 context. And all of those essentially require a database that is crypto native, it is Web3 native. And so that was the original genesis of the idea of Polybase. Can we build a database for developers that is seamless when they want to build Web3 applications? So we started building features around authentication based on wallets, around permissioning based on wallets and tokens, and of decentralizing the concept of a database. Now, over the course of a year, it’s been almost exactly a year since we started, that has kind of changed in many ways. The main product is still the same, but the vision of what we want to build and the people that we want to build it for has changed a little bit. Kind of the main phrase we’re talking about today is a database with the features of a blockchain. So if we look at why people want to use blockchains, there’s a couple of fundamental reasons why. The first one is unbreakable rules. So you have smart contracts, once those are deployed, you can’t break those rules. Another one is immutability of history. You can’t manipulate something to say it’s a, you know, something happened that didn’t happen. And kind of this decentralization permissionless access and use. So we said, can we take these features that people love about blockchains and care about blockchains and build them into regular databases, like Postgres, MongoDB, Oracle databases you’ve heard of before. The reason behind that is if we can do that, then we’re building a system that’s just as scalable and cost efficient as Web2 and has the same kind of UX as Web2, but with all the features that we care about from Web3. And that’s kind of the holy grail to me. You see all this talk about scaling Ethereum, L2s, L3s, you know, cross-chain protocols at the end of the day, that’s kind of incrementally making us a little bit faster and faster. But when we did the math, we’re still about 10,000 times slower in the fastest L2s than we are for just a regular database. You can run an AWS or Google Cloud, right? 10,000 times. That’s huge. And it’s about a million times slower and million times more expensive to run an Ethereum transaction than it is to update a row in a database. So that’s kind of like, you know, million X, right? That’s, that’s like unfathomable. So that gap is never really going to be able to be bridged in an effective way. And so when we look at databases, well understood, solve the problems, and especially in the enterprise use cases that we’re going after, all enterprises use regular databases today. And so that’s kind of the value we’re bringing is like, can we give you all the features of the blockchain? But in this database you used today already, and we’ve made a lot of progress towards that. And I think we’re going to be able to really solve that in a big way. 

Matt Zahab 
Can you walk me through some of the perhaps consumer or just B2B and use cases? Because again, after hearing the pitch, it’s like, I understand what you guys are doing, but I’d love for some, you know, applicable use cases just to have a better pulse on the narrative. 

Sid Gandhi 
Yeah, I’ll give you a case study from one of our customers, Credora. So Credora is building a encrypted loan legend. The purpose of that is to allow folks that give out loans to be able to collaborate on loans that they’ve given out in order to reduce fraud. You can think of like a borrower coming in to get a loan. They’re going to show all their assets, of course, so they can get a good rate, their risk profile is lower. But the key challenge in Web2 or Web3 is what are the liabilities of this borrower? Right? Because if they show a bunch of assets, those could actually be based on other loans that they’ve taken out. That’s obviously a really big problem. So if you have a collaborative database of loans that a group of lenders have given out, you can start to understand the liability positions of borrowers and therefore reduce the risk and give better rates to those borrowers. So that’s one example of kind of an enterprise B2B use case where you can’t really build that today because getting multiple entities and organizations to collaborate on one set of database really hard. You know, crew’s hosting it, who has control over it, who can edit it, what kind of data is exposed, what kind of data is kept private. All of those rules can be codified in Polybase. Individual records can be encrypted. Certain columns of those records can be encrypted or decrypted. You can use ZK proofs to prove public attributes of private data and look, we can get into what that means later as well. So that’s one use case. Another one is in the energy side, we’re working with a large energy company in Europe that wants to take all of their operational data like what power plants are generating, how much electricity is a renewable, etc. And be able to expose that in a transparent way without revealing the private data that they don’t want to reveal. And again, ZK proofs can help with that. And so, you know, loading all that data into Polybase or syncing into Polybase and then being able to expose certain attributes of this data in a public way allows them to do kind of the job of a auditor, but in real time. And cryptographically veritable. 

Matt Zahab 
Huge. You keep talking about the security aspect. And again, it seems like you guys aren’t going after companies that have data that’s, you know, rather laissez-faire. You’re going after companies that have very robust, you know, privacy based data. Then it would be nightmare fuel if, you know, a hacker or anyone got in and wanted to do something, you know, malicious with that. How are you guys leveraging ZK proofs to really make your data as secure and versatile and efficient as possible? 

Sid Gandhi 
Yeah. So we actually use ZK proofs for two reasons. One is for the security aspect and the privacy aspect. The other one is for scalability. So I’ll talk about the privacy and then we can go into scalability. Zero knowledge proofs fundamentally allow you to prove that you know something without revealing the actual underlying data. That sounds like magic, but it’s actually just a lot of math, polynomial math specifically. And essentially what that allows you to do is say, OK, let’s say I have, you know, 10 records or rows in a database or rows in a spreadsheet. I am able to prove to you what the average, let’s say I have a particular number in that spreadsheet is for those rows without actually revealing to you the underlying individual data. And when you get that average, without actually having to redo that calculation yourselves, I provide you a cryptographic proof, which is a short string of numbers and letters that represent the fact that I did that calculation correctly. And with those two pieces of information, the average and the proof, you can be cryptographically certain that I did that average correctly. So that has wide-ranging applications, finance, energy I mentioned, in media and advertising as well, when you think about tracking people, but keeping their data private. So that’s kind of fundamentally how we use ZK proofs. We also use them in a lot of different places in our technical stack. So in our consensus mechanism, we have built it in a way that’s ZK -able. So you can think about traditional BFT consensus and Ethereum, for example. Each individual validator is running this consensus on Ethereum, and that each individual validator has to redo all the work to make sure that all the other validators are doing the right thing. That’s why it takes so much computing power. That’s why it requires hundreds of thousands, I think whatever 200,000 validators that Ethereum has to make sure everyone’s doing the right thing. In Polybase, when we use ZK proofs, each validator just sends the next update or the next block and a proof that they assembled that block correctly according to the rules of the system. And no one else needs to redo that work. All they need to do is validate that proof, which is a very quick and easy operation or computation. That’s a scalability aspect of ZK proofs. 

Matt Zahab 
ZK proofs are so cool. This is one area of cryptography and blockchain that interests me like crazy, full disclosure and full candor. I do still have a little bit of trouble fully understanding them. Obviously, I don’t have the technical and mathematical background that you have, but on the ZK proof side, not even relating it back to Polybase. I feel like ZK proofs last year, they were incredibly trendy. 2021, I felt like it wasn’t really in front of mind on a lot of people. It wasn’t really two on the radar. Last year, Q3, Q4 really picked up and that momentum is continued in 2023. When will society as a whole really see ZK proofs embedded into our everyday lives? How far out are we? And again, my apologies for the forward-looking question, but this is something that I’m really interested in. 

Sid Gandhi 
Yeah. Well, I can tell you a number in years or months or whatever, right? But I think the key is what will it take? What things do we need to accomplish in order to make that happen? I mentioned this and I touched on this earlier, which is going from research to product. The ZK proofs are still very much in a research phase and they’re solely now transitioning into alpha and beta phase products. A lot of those products today though are in Web3. They’re used for ZK rollups, probably heard of those, ZkVMs, zkEVMs. And the challenge with that is the scope of presenting those products and solving problems for the general population is very small. Now what we are positing is that databases rather than blockchains currently touch almost every single internet connected person in the world. If you’re using the internet, you’re using the database somewhere. So to answer your question, what it takes is to have ZK proofs embedded in a product that everyone already uses today. And that’s kind of like this inception thing, right? You don’t want to have people change their behaviors. That’s really hard to do. Keep doing the same thing and we just swap out things underneath to make things more secure, more private, more scalable, and give people new capabilities eventually. And so maybe in a somewhat self -serving way, like PollyBase is that is our mission is to bring ZK proofs to the general population and allow them to have control over their information, whether that’s through privacy or other controls that they don’t have today. And so I think, you know, that’s going to happen within the next year as we go into production, mainnet, we build out some of these B2B use cases that eventually then touch and end users and consumers. 

Matt Zahab 
So cool. Yeah, I mean, I can’t wait for that. Sid we’re gonna take a quick break. Give a huge shout out to our sponsors of the show and longtime friends of cryptonews.com. And that is PrimeXBT. Like I said, longtime friends of cryptonews.com incredible group of guys and gals who built a robust trading system for both beginners and professional traders. It doesn’t matter for your rookie or a vet, you can easily design and customize your layouts and widgets to best fit your trading style. PrimeXBT is also running an exclusive promo for listeners of the Cryptonews Pod. After making your first deposit, 50% will be deposited into your account that can be used as bonus for open positions. Again, that is CRYPTONEWS50 to take advantage of this offer. CRYPTONEWS50 to receive 50% of your deposit credited to your trading account. And now back to the show with Sid. Sid I want to go a little off topic for a second. On Twitter, you have something this came out on June 13th. It was an interview with Protocol Labs where you talked about your $100,000 burrito. Absolutely loved this story. You got to tell it on the pod. It’s all world. The floor is yours. Tell me about your $100,000 burrito. 

Sid Gandhi 
Yeah, the year was 2012. Bitcoin was, it was like $8 or something. I had just heard about it. I had just kind of been on some forums and read the white paper. And back in that day, you had to, if you wanted to buy Bitcoin, which was the only cryptocurrency available, you had to send a MoneyGram. I don’t know if anyone listening has ever even used a MoneyGram, but it’s the thing where you send money across borders. That was the way you did it before. And you would have to send a MoneyGram to some random company or even person somewhere in the world. And then you would cross your fingers and hope that they transferred some Bitcoin to a little grass-eyed and specified. 

Matt Zahab 
So you’re just praying that it’s you. 

Sid Gandhi 
Yeah, that was on ramping circa 2012. And so you had to like print out a form, put it in an envelope and then take it to like a Walgreens or a CDS or some department store. And so I did that. I had out all the paperwork, everything ready to go. I was going to buy $20 worth of Bitcoin. So I would have had like two and a half Bitcoins or something like that. But, you know, I had that all set up and ready to go. And then my roommates were like, Hey, we’re hungry. Like you want to grab some Taco Bell? And naturally, like Taco Bell comes up. I’m like, I’m there, not in college. And so I take that $20 and we go to Taco Bell and we buy burritos and tacos and whatnot. And that’s the story of the $100,000 burrito. 

Matt Zahab 
Geez Louise. Oh, I have one similar one like that. It was first year uni 2013. It was on Bodog, a gambling website. I think I was of age. Maybe I wasn’t. But yeah, you got an opportunity to get your bonus paid out in Bitcoin. And it was, I forget the exact number, but it was in the low 20s of Bitcoin. And, you know, I read up on it back then, but I was like, Damn, there’s no way this is ever going to pick up. And so I took the Bitcoin and then they also let you sell the Bitcoin for dollars on the platform as well. Just like a quick little transfer, transferred it out. So I could be sitting on 20 BTC, you know, life would be good. Life’s still good. I can’t complain. But I feel like all of us have a weird little story like that. So you got to love that. Let’s go back into Polybase for a sec. Two things I’d love to discuss is the private mainnet. And you guys are doing some really environmentally focused stuff in that regard. And I’d love to get into your competitors. This is something that even when we did a little outreach on socials, people are asking how is Polybase different from ceramic and table and so we’ll start with mainnet and if you could segue into your competitors and how you guys differ that be a treat.

Sid Gandhi 
Cool. So private mainnet is our first production level service. And we in the last couple months, we’ve onboarded our first paying customers onto that private mainnet. The reason we phased it out this way and we kind of have a roadmap laid out on our website and our blog. The reason we phased it out this way is a lot of our customers actually want what we’re building before we even have the incentivization layer. And then by incentivization layer, I basically mean like token launch with the utility token that if you contribute to the network in terms of providing storage and indexing, you get paid for that. So we don’t have that yet. That’s scheduled for March 2024, the token launch, but a lot of folks wanted to use it, use Polybase in production today. So we spun up that service. We have, you know, service contracts, SLAs, kind of all the payment thing set up as more of a infrastructure as a service. So something like Google Cloud or AWS, how you would use that and pay for it. You can use Polybase today. I mean, that’s actually really attractive to some of our larger enterprise B2B customers. They don’t want to deal with crypto. You know, again, it’s like this thing of like, they want all the features of blockchain, but they don’t want to deal with all the downsides of that. So that’s scalability costs, but also like the way you pay for it, right? And then gas fees and how it’s like a customer or a consumer driven fee market. And so that’s what private mainnet represents. It’s up and running today. You have to fill out a form. We’ll chat with you about your use case. And if everything aligns, then we’ll onboard you and whitelist you for that. 

Matt Zahab 
Gotcha. Interesting. Walk me through the competitor side. 

Sid Gandhi 
Yeah. So I think we always get this question. We get it a lot. And the answer is really simple. It’s that’s for Epoch and TableLand are very, very different products from a Polybase is building. Polybase has the same interface as a traditional database. It stores data in traditional databases. And it provides these database level attributes using zero knowledge proofs. You can have public data, you can have private data, you can have data somewhere in between, you know, this fine tuning of this knob. The cost for Polybase is the same as running, you know, like a traditional database in AWS. The performance is also the same. Now, when we go to a product like Ceramic, first of all, it only allows public data. So it’s really useful for things where you want to have maybe a very public social network and the graph you want the social graph to be fully public. It’s also very limited in terms of its querying capabilities. And so it’s actually made as a stream of events rather than a database. And you have to do some stuff to make it work like a database. And so it’s more of a decentralized event streaming platform. And also, I think some of the usability aspects and the SDKs and APIs aren’t as friendly for Ceramic as for Polybase. And Polybase, it’s almost the same API as Firebase. Most, I would say, like 90% of developers I’ve talked to have used Firebase already. So it’s almost the exact same API. Nothing really to learn or no new concepts to pick up, it just works. So that’s not true about some of the other ones. Now, on the table and side table, it is very different. It’s about representing database tables as NFTs. So that’s kind of cool to be completely honest. I don’t really know of any concrete use cases that that applies to. It’s like one of those things in Crypto, it’s like, wow, that sounds really cool. And I can maybe think of a million things that you might be able to do, like buy and sell databases. That’s cool. But what does that really mean when you can actually just copy data once you have it? It’s just kind of like the JPEG question, right? And so it’s like, well, in the JPEG question for NFTs, the idea is that NFTs represent cultural value in art, and just regular has value, picture NFTs at value. I don’t think the same thing really applies to databases, to be honest. And so that’s always been kind of confusing. Also, the interface, anytime you write data to the database, requires an Ethereum transaction, which means if you’re running a social network on there, you’ll be paying trillions of dollars in cost. If you wanna run a social network like Twitter. And so obviously that’s a no go for 99% of the use case that we look at that requires scalable fast data writing and reading. 

Matt Zahab 
Yeah. So I wish we had more time here, Sid I have so many points and questions that I’d love to bring up. We’re gonna have to save this for round two. We are getting a little tight here, and it’s been an absolute treat having you on. I would love to get some Sid hot takes. I know you got a couple. It doesn’t have to be crypto related. Can be health, wealth, happiness, space, AI, you name it. But anything that’s really tickling your fancy at the moment or you think a non obvious outcome could arise. Any Sid Gandhi hot takes before we let you go. 

Sid Gandhi 
Yeah. I mean, I could talk about some topics that are popular right now. I guess like everyone’s like AGI and LLMs and AI. I’m super excited. We use AI like a lot in Polybase to give you an idea. We have a Discord bot that answers 90% of developer questions automatically provides links to documentation, sample code, weird edge cases that developers ask about almost completely saw by this ChatGPT-4 based bot that one of our developers wrote in a weekend. That’s real value. 

Matt Zahab 
In a weekend? 

Sid Gandhi 
In a weekend. We’ve been improving it over time, but the initial one was like, yeah. You can think about, you know, for a startup like us, that’s driving tremendous value, right? Like tens of thousands of dollars worth of value or we don’t have to hire a separate developer to monitor our Discord and answer developer questions. I can imagine for larger enterprises, that value could be in the billions. And so this is real, right? This is not hype. This is real business value. So I’m super excited about that. We don’t need AGI to provide billions of dollars of real business value, right? It’s here today. It’s about how you use it and what the interface for it is. 

Matt Zahab 
Hold up, Sid. On AGI, before you go to your next point, is that a big old boost or are we, you know, maybe a year, a couple of months away, like a lot of people are saying? 

Sid Gandhi 
I think first of all, no one has a clear definition of it. And so it’s really hard to say. I’ve seen some demos of using ChatGPT in a loop where you keep asking itself, like, hey, what should I do next? Or lay out an outline of like accomplishing this task and to kind of like iteratively solve that. That seems like the pathway towards AGI, which is like basically giving ChatGPT a conscience, like a goal and then running it in a loop to just ask itself. That’s kind of what our brains do as well. Like when you’re stuck on a problem, you just ask yourself, like, how can I solve this or whatever? I would say like a very basic, maybe toddler level of AGI is already here and in that kind of loop. That’d say, yeah, in the next couple of years, we’ll just keep making that better and better. Like I mentioned with the self-driving cars, we’re at a teenage level now. But getting some teenage to a fully formed adult into PhD level. 

Matt Zahab 
Takes time. You gotta get your reps in. 

Sid Gandhi 
That takes like exponentially more time, I think. So hard to predict, but we’re a ways away from a PhD level, I would say. 

Matt Zahab 
That’s gonna be, I mean, the only scary part, and again, this is so cliche. This obviously isn’t an outlier take by any means, but the only scary part is the amount of jobs that will be nullified from that. Like I just, you know, I look at myself having even an econ background, working for startup. We had a team of like 10 full time VAs who were just, you know, just answering product questions nonstop. You don’t need that anymore. And actually the company I used to work for, I saw one of the founders on LinkedIn, like go the whole team. And it’s now like building a couple in house. I don’t even know if it’s in house. To be honest, it’s probably just, you know, it’s probably just ChatGPT that, and he’s calling it in house, but it’s crazy. A lot of jobs are gonna be gone unless you specialize in some area. What’s the, I think it’s the Naval quote, always be a specialist and not a generalist. Like if you’re a generalist and you’re not really good at things now, I think you might be getting into spooky town. It could be a little scary. So I’m super excited, but I’ve been enjoying, you know, the ChatGPTs and the Midjourneys and the, you know, the, you know, all the applications out right now and I can’t wait to see what’s next. 

Sid Gandhi 
My answer for the jobs question is honestly, we’re probably gonna have to do you universal basic income at some point. You’re just like, isn’t enough stuff to do. People are just on their phones nowadays. Like think about that number of hours people are either on their phones on TikTok or watching Netflix. And subtract that from a productive society’s work hours. Like we’re already kind of there. We’re like, no one’s really has anything to do. There’s a lot of like physical work. Of course, people are still doing, but I think in the next like 10 years or so, we’re really gonna have to just tell people like, hey, hang out, stay home, be an artist, make music, do whatever you want and just like enjoy life. And here’s, you know, 5K to 10K a month to live on depending on where you are. 

Matt Zahab 
You think UBI will work? 

Sid Gandhi 
I think regardless of whether it’ll work, we’re gonna have to do it. Like we’re gonna have to make it work because I truly believe there’s like, isn’t really that much to do anymore. We solve everything, everything’s automated. 

Matt Zahab 
Oh, some crazy times I had said what a treat man. I had an absolute blast talking with you and can’t wait for round two. Before we let you go, can you please let our listeners know where they can find you and Polybase online and on socials? 

Sid Gandhi 
Yeah, Twitter is the best. I’m @sidgandhi_xyz on Twitter and Polybase is @polybase_xyz on Twitter. Really excited to hear about use cases if you’re a builder and you want to build in the space. We have meetings with everyone that wants to talk to us. So hit us up and we’ll talk about it. 

Matt Zahab 
Love that. Sid, thank you so much for coming on. Appreciate the time and can’t wait for round two, wishing you and the team all the best until then and we’ll keep in touch, my friend. 

Sid Gandhi 
Thanks, Matt. Appreciate it. 

Matt Zahab 
Folks, what an episode with Sid Gandhi, CEO of Polybase. Go check them out at polybase.xyz. As always, I will include all of the links, everything in the show notes under all of the major podcast links and guys, do go check out their website. It’s one of the cleanest websites I’ve seen as well. And again, this is a free pod. No one pays us to come on. We just, when we see something we like, we like to show some love back. If you guys love this one, I hope you did. Please do subscribe. It would mean the world to my team and I. Speaking of the team, love you guys so much. Thank you for everything. Justas my amazing sound editor. Appreciate you. Back to the listeners, love you guys. Keep on growing those bags and keep on staying healthy, wealthy and happy. Bye for now. We’ll see you on the next episode.