Interview: Dr. Rand Hindi of Snips, Shares Thoughts on Privacy, Blockchain and More
Cryptovest’s Merav Ozair interviewed thought leader Rand Hindi on the future of privacy, blockchain and our second chance to make a change.
Rand Hindi is a gifted and atypical serial entrepreneur, who founded his first social network company at age 14, started his PhD at 21 and holds two graduate degrees from Singularity University in Silicon Valley and THNK in Amsterdam. It comes as no surprise that he was selected for the “30 under 30” by Forbes and TR35 by MIT Technology Review. These days, he is on a mission to change the way our private data is handled and to protect our privacy. Using artificial intelligence, cryptography and blockchain, his company, Snips, aims to achieve this goal.
At the CryptoBlockCon conference in New York City, we got to sit with Rand and talk. Our discussion spanned from the value of data and its usage to privacy issues to the future of blockchain and the economic system. Rand, a true thought leader, explains why with blockchain we have a second chance, as a collective community, to better our life and calls everyone to join the “blockchain revolution”.
CV: Hello. I'm here with Rand Hindi from Snips and he's going to tell us a little bit about his vision and why he is in the blockchain space. So, tell us a little bit about yourself and how did you get to Snips?
RH: Hey, thank you for having me. So, I've been coding since I was a kid. I started when I was 10 years old. At 14, I started my first company, back in 1999. It was a social network that worked out pretty nicely. One of the things that I realized then was that it felt wrong that I was able to just snoop into people's private messages that they were sending each other. So, this whole idea of privacy started to become a central theme of what I was doing. At 18, when I went to University in London, I started doing Machine Learning. - learning how to teach machines how to do things, or Artificial Intelligence, which eventually led me at 21 to a PhD in Artificial Intelligence applied to biology.
My current company, Snips, I've been running for five years and our objective has really been to create Machine Learning and AI technologies that guarantee privacy by design. So, I’m really trying to merge those two different things, because, fundamentally, one of the most complicated and challenging tasks we have today is privacy. Because, if you don't start thinking about ways to protect people's privacy, then you end up with the kind of abuses we've seen with Facebook and Cambridge Analytica and all the other absolutely terrible, manipulative, anti-freedom things that could happen.
CV: Why do you think blockchain is the solution for privacy?
RH: When you talk about privacy, there are two types of privacy. You have something that is called privacy by Trust, which is when a company tells you: “trust me with your data. I promise, I'm maybe going to delete it or I'm not going to do something bad with it”. But privacy by Trust equals no privacy whatsoever. Because the government can come and say: “give me the data”, a new CEO could take over and not do the same thing and plus why would you trust the company with your data, if you don't know how they're actually doing it.
On the other hand, you've got something called privacy by design. The idea of privacy by design is that you’re mathematically guaranteeing privacy in the way you're building your technology, your business, your product. An example of this could be end-to-end encryption in messaging apps. It's impossible for someone building an end-to-end encrypted messaging app to read the messages because there is no way for them to decrypt the messages. This idea of privacy by design is very powerful. Because it means you don't have to trust a company with anything. Because there is nothing for you to trust them with. At Snips, privacy by design has been the central thinking of everything we've been doing.
From the first day we’d got created in 2013, we decided that everything we were going to build, machine learning wise, was going to guarantee privacy by design. So how does that relate to blockchain? One of the most effective ways to guarantee privacy is by combining three different things. The first one is doing what we call Edge processing. Rather than sending data to the cloud for processing, you can do it locally on people's smartphones or on people's TVs and things like that. The second thing that you want to do is decentralized encrypted machine learning. This is a really cool piece of technology where you're not just spreading the calculation of machine learning on a bunch of different computers, you're actually doing it encrypted, using things like homomorphic encryption, multi-party computation. Meaning that even those who are processing the data on the network don't actually have the data. They have an encrypted version of the data. When you start combining traditional cryptography with machine learning, with Blockchain, then you're able to offer the exact same thing that people do centrally by sucking up your data into their servers, but in a completely decentralized, completely private way.
And of course, one of the very nice things about that is that you can also pay people for contributing to your data.
CV: So that's the incentive.
RH: Exactly. There is a financial incentive for users of your system to send you their encrypted data to help you make your system better. Because you can pay them back in tokens for having done that without them having to trade off privacy. This is not “I'll pay you for your data”. This is “I’ll pay you for contributing your encrypted data to make my product better”, which is very different.
CV: If let's say I have data and I want to use your product, who am I paying exactly?
RH: Just to clarify a little bit. What we do at Snips is we build voice assistants. Things like Alexa or Siri that are hundred percent private by design and decentralized. Using these technologies we just talked about – encrypted machine learning, blockchain, we're able to offer an alternative to Alexa that doesn't require your data to ever be visible to anybody else but yourself. You can put the smart speaker in your living room, nobody is ever going to listen to you or your kids, because your data is never going to be sent unencrypted anywhere on the planet.
This is really important. Because if you want people to feel safe using a voice assistant, they need also to be safe knowing that their data is protected, that their kids are not under surveillance, that they're not being manipulated. Because the more data you give away to big companies, the more those big companies can hook you into their services and manipulate you, whichever way they want, and this is really something that's important for us.
As a user of our voice assistant, when you speak to Snips, you can decide to encrypt and then send this data to the blockchain. So that it is used to make the Snips voice assistant better for everybody else. But you're not doing this with your actual data, you are doing it with encrypted version of your data.
CV: So, there will be some kind of a button? Like on-off?
RH: Exactly. There's like an opt setting in your assistant. You can say: “Okay, I'd like to contribute my data”. You can do that safely. There is no privacy implications whatsoever, because of this encrypted decentralized machine learning technology and you can still participate and contribute to making things better.
CV: How am I being paid? I mean, that's what people will ask. It’s nice that it's encrypted, I'm okay with that, but how I am going to be paid? How the money is going to be transferred to my wallet? Who's going to transfer it? Developers? Somebody else?
RH: The way that this voice assistant works is: the voice assistant itself doesn't really do much. What you need are developers building apps on top of your voice assistant that gives it different functionalities or skills.
A developer will build a weather app that you can ask about the weather by talking to your smart speaker. Another one will do like a taxi app or something like that. So really what's happening here is that the users when they speak with one of those apps, they're really helping the developer make this better by improving the understanding of the app on the assistant. So, the developer is paying the users for contributing, and all of that is done completely automatically through a smart contract. Technically, you don't have to do anything. All you need to do as a user is talk to the products and, as long as you've opted in, the developer will be able to use your data to train the machine learning models and then pay you back in exchange.
CV: Is there any incentive for the developers here?
RH: Of course. The way you think about this is: I'm a developer, I've got a very successful app. It's published on some kind of an apps store. If I'm making a lot of money off of my app another developer is probably going to want to compete with me.
So, I've got an incentive to keep improving my products by involving my users and their data and paying them back. Because, that's how I remain competitive with other people. It's a very simple economic principle. I'm making a ton of money with a product. Someone else is going to want to take my place. The best way for me to defend against that is to keep improving my product. So, the driver here is competition between developers for a successful product. This is really important, it’s also a differentiator compared to things like Alexa. Because Alexa doesn't offer you that.
As a user, when you use Alexa, not only do you not have privacy at all, they're not giving you anything in exchange for your data. Actually, you’re paying twice. Because you're paying once for the device itself, as with Amazon Echo, and then you're paying a second time with your data, because Amazon uses it to make money on your back. And as a developer, it’s even worse. Because you're building apps for Alexa and you're feeding all of the data from your users of your apps to Amazon.
We're doing the exact opposite. What we're saying is that this old mindset, that's from the ‘90s, abusing people's privacy, pretending to be platforms. But, it’s really just acting as data aggregators and cheating developers. Cheating users with your data has to stop. Decentralization, privacy by design, all of that has one goal – to destroy Alexa and products that abuse our privacy and replace them by privacy by design, decentralized safe alternatives.
CV: Is it fair to say that data is a commodity?
RH: Data is absolutely necessary for machine learning. What needs to change is the way we handle data. There’s no mathematical limitation to what you can do with machine learning and encrypted data. The only reason a company would not want to guarantee privacy is not because it's not doable, it’s because they’re probably doing something else with your data that you're not aware of. That's also what's so interesting. For an existing company, like Amazon to offer privacy, they will need to shift their business model which revolves around exploiting our data. This is why it's so exciting to be in blockchain when you work in privacy, because you're coming from this side. You've got an angle that nobody else can take, because that will require shifting billions of dollars of revenue.
So, the bet we're making here is that given two equivalent products: one with privacy and one without privacy, same level of convenience, same capabilities, people will choose privacy. Privacy is not something that we should take lightly.
So, to the question of whether data is a commodity. Well, it depends if you consider commodity in the financial sense. It's hard to consider that it’s a commodity, because data can be copied and replicated infinitely. There is no scarcity on a piece of data. It's a cost of electricity, I guess, to copy data. So, I think it's very, very different ...we have to think about it in a very different way.
I think of data more as a fuel. Less as a commodity. It's something that you need to make it work, but it's not something necessarily that you want to hold in a safe, like gold or something.
CV: But, if it's private data, then it is like something that I want to keep in a private place?
RH: Perhaps, or not, or maybe what you don't want is to keep your data. Maybe you want to make sure it’s not that nobody uses it, but rather nobody uses it in a way where they can see it. You want as many people to use your data, because that's useful. But without knowing the data itself. This is the key. I know this is like black magic thinking. How can you offer something: they use my data, but they don't see the data. This is why cryptography is so exciting. Cryptography is not just about blockchain. Cryptography is a huge field which includes homomorphic encryption, differential privacy, all these technologies, you can combine with blockchain to make this magic happen. Blockchain is the incentivization and the security layer, but the actual privacy comes from the combination of blockchain with the secure encrypted stuff.
CV: What is your vision for blockchain? I'm guessing that you do believe in the doctrine that it represents, and this is why you're doing it. Can you tell us a bit about that?
RH: Well, from a technology perspective, for having been thinking about privacy for over 10 years now, machine learning for 15 years, all of those technologies exist for a long time. Decentralized encrypted machine learnings is nothing new. But we were missing the incentive. Why would someone give their encrypted data? Why would someone process this encrypted data? Like what would be the incentive? So, when you plug it to blockchain, all of a sudden, wow. It just works. I think it's very important to understand that theoretically you can do a lot of things. But if you want to make it work in the real world, you have to play by the rules of the real world, which means incentivizing people to participate and blockchain is a very effective way to do that.
So, the more I was thinking about this, the more I realized, a token is really fundamentally about creating incentive for communities to self-organize and do something beneficial for themselves. In the case of machine learning that's improving the machine learning algorithm that they can use to whatever. So, that typically is something I think is going to be an enormous outcome of blockchain: decentralized autonomous organizations.
I think this is going to be something major, major in the thinking around building products, around handling users, around governance, all these different things. So, if I was to think about what's the biggest, like what's my vision for blockchain in the future? I think currencies is just a small part of it. It's really this idea of creating a glue between people to help themselves organized rather than having someone telling them: “that's what you're gonna do”. I think that's the kind of a big, big thing that's going to happen.
CV: And how do you see Snips participating in this vision? What will be your part in contributing to this vision?
RH: Well, we're offering the decentralized private by design voice assistant. We're offering you this AI that you can use every day safely with your kids that you never have to worry about what's going to happen. But we're just one piece of a bigger puzzle. We, like many other companies are decentralizing and securing existing products, by taking them on a blockchain, doing things differently. But you need the same thing for search. You need the same thing for messaging. You need the same thing for finance. I think you're going to have pretty much everything that sensitive in the centralized world is going to be decentralized and made more secure, more private.
Maybe not even more efficient. Maybe there is a trade-off of efficiency that people are willing to accept for privacy. But I think we, with other companies, need to work together to offer an alternative to all these old economy products and services that don't necessarily have the same alignment of values as we do. It's really, really important that more product managers and more designers and more UX people and more entrepreneurs start looking at blockchain as a new infrastructure, as a new way of building products that you can use. I think we have to start to stop thinking purely in terms of infrastructure and we need to start building stuff on top of it.
CV: From what I hear from you, it's like you want to change the entire economy, in a sense.
CV: Everything that drives the economy now, most of it, is centralized, if you think about it. From enterprise to government even private.
RH: I'm not solely the one who's going to do that. But I think we should do that. As a collective community. We should definitely do that. Because, all those companies, all those governance models, maybe not all of them, a lot of them, they were built to serve a world where humans were machines. Industrial era type of thinking. But we kind of passed that. We've got the machines to do machine stuff now and people are trying to do things which are a lot more human. So, the entire infrastructure to support people has to change accordingly.
It's just very hard, because, well nobody wants to give away power. So, I think it's going to take a lot of time and it's going to take a lot of crazy thinking. It's going to take a lot of closing your eyes to what people tell you in the real world and just going for what you believe is right.
And I think this is really what's getting me so excited. It's the first time that we have a community since the early days of the internet where so many people are so passionate about making things better for a wider community.
CV: So, where do you see blockchain in the next 10 years, and Snips in the next 10 years?
RH: So that depends on what we do today. I think as a community we can decide to make blockchain the revolution it’s supposed to be, or we can be greedy and try to fix blockchain into the old way of thinking, in which case it's going to be a moot point. And that's a real risk. The internet started in the beginning with this idea of decentralization, of information sharing, of freely accessible information. But it ended up being controlled by the same powers with the same mindset than before. We've got a second chance – that’s changing things with blockchain. We've got a second chance …
CV: A better chance, you think?
RH: I think it's a better chance, because it's harder to actually control and it's also happening in a political and economic environment that is very unique. We have a second chance at fulfilling the whole thing, and this is something I definitely want to be a part of. I want to put my dent into this whole universe in a way. I want to do it for AI and I'm hoping that other people are going to do it for other things. If an established company, like myself, having been around for five years, we are 70 people, raised a ton of money (traditionally from equity funding). If I can make the leap to blockchain then everybody else can.
CV: Is there anything else that you would like to share with us?
RH: I'm always very easy to find online. I love chatting about that. All these things are debatable. So, if anybody is interested in continuing the conversation, on Twitter, I'm Randhindi, on Telegram I'm Randhindi. Basically, Randhindi on every social media. Despite how much I believe in privacy, I’ve made myself extremely easily accessible. Because I like talking about those things and I like engaging with people who believe in the same things I believe in, so that together we can start spreading the word and making a difference.
CV: Thank you very much, Rand. And I hope that your vision will come to fruition.
RH: Thank you very much.