FLOSS Weekly 727 Transcript

Please be advised this transcript is AI-generated and may not be word for word. Time codes refer to the approximate times in the ad-supported version of the show.

Doc Searls (00:00:00):
This is FLOSS Weekly. I'm Doc Searls, and this week Dan Lynch and I talk with Rand Hindi of Zama and fhc.org. But here's the thing, it's all about homomorphic encryption. This is a kind of encryption that's more private than anything we've had before. The pre the encryption is done in such a way that you're only, you're private at the ends. And all kinds of processing can be done on the encrypted data in the middles, as it were. Here's the thing. He thinks it's gonna be really big in the next several years, and he's working on this. It's really interesting stuff. We go off on a lot of tangents, but it's a new thing. It's big, and it's coming up Next.

Announcer (00:00:42):
Podcasts you love from people you trust. This is TWiT.

Doc Searls (00:00:50):
This is FLOSS Weekly, episode 727, recorded Wednesday, April 12th, 2023. Homomorphic encryption. The next big thing, this episode of FLOSS Weekly is brought to you by Kolide. That's Kolide with a k Kolide is a device trust solution that ensures that if a device isn't secure, it can't access your apps, it's zero trust for Okta. Visit Kolide.com/floss and book a demo today. Hello again, everybody everywhere. I am Doc Searls, and this is FLOSS Weekly brought to you this week with Dan Lynch himself. There he, there he is with red very strong red theme there. You've got the red, the red, the red spit filter on your mic. You got the a head.

Dan Lynch (00:01:45):
I know. Yeah.

Doc Searls (00:01:46):
Not a red hat hat, but a one with a red chicken on it

Dan Lynch (00:01:49):
Or something. Yeah, this is, yeah. So, I mean, I suppose a little cultural note for everybody. Hello. Nice to be back by the way this little bird on here is called the Liva Bird, and it's the symbol of Liverpool.

Doc Searls (00:02:00):
I thought so it's, it's l l it says there also,

Dan Lynch (00:02:03):
Yeah. Liverpool football club. But the actual bird is is part of the, like the city. There's two statues of them on top of Liva building, which is which is like one of our cultural landmarks, if you like. So yeah, but red's my favorite color. That's, that's the only explanation for that.

Doc Searls (00:02:19):
Yeah, because you, you got a lead red lampshade in the background, and you also have a red line I did circling your both of your ear cans

Dan Lynch (00:02:27):
And a red guitar that you can't see. That's

Doc Searls (00:02:29):
Just, yeah, I can see it. I can see the guitar. Yeah,

Dan Lynch (00:02:31):
The red, red guitar. Yeah, I, I do I

Doc Searls (00:02:34):
Red thin glasses that or Rose do. Yeah, glasses. <Laugh>. That'd be good.

Dan Lynch (00:02:38):
Tinted ones. Yeah. I'm wearing my amber glasses today.

Doc Searls (00:02:42):
And you're also wearing your Firefox T-shirt, which I have one just like that, but much more ratty looking than yours. 

Dan Lynch (00:02:47):
This, this is getting pretty old now. I I need to get a new

Doc Searls (00:02:50):
One. Yeah. It, it used to be blue and now it's kind of gray. Yeah, they, they do that

Dan Lynch (00:02:54):
<Laugh>. Yeah. It was kindly given to me by somebody from Mozilla. Yeah. So I have to kind kindly ask them, see if they'll give me another one. I dunno if you can buy them. Yeah,

Doc Searls (00:03:02):
I, I, I, I worked for them for about a year and a half, something like that back in the mid teens. So yeah, we can make an earth sandwich. You're in Liverpool. I'm now in Los Angeles, or near mm-hmm. <Affirmative> and mm-hmm. <Affirmative>. so we're eight hours behind each other, or I'm behind you. You're in the future for me. So, so have you done your homework on this one with Rand Hindi today?

Dan Lynch (00:03:25):
I have, yeah. I've been reading all about all the, all the f h e encryption stuff, and I can't say I'm an expert, but that's why we have an expert guest, you see <laugh>. But yeah, I do, I I have been looking into this today. It is very interesting stuff. It's quite an exciting time for obviously privacy. Very very, it's always been important to most of us, especially I would imagine in the community around this show. But becoming more kind of publicly unknown at the moment. The arguments about privacy and stuff. A lot of stuff going on in the UK right now about encryption and the government wanting to enforce companies to open up their encryption. So we've got WhatsApp talking about possibly being banned in the uk signal have said they'll leave the UK absolutely 100% if, if they're asked to do anything to compromise their encryption. So it's an interesting time. I

Doc Searls (00:04:14):
Did not know about that one. That's interesting. Mm-Hmm. <affirmative>. Wow. Well I wanted, it's all new stuff to me, and I think it's really interesting. I think it'll be really interesting for our listeners and viewers. So I'm going to introduce him. Our, our guest is Dr. Rand Hindi. He's an entrepreneur. He's a deep tech investor. He's a C e O at Zama and open source homomorphic encryption company and an investor in more than 30 companies across privacy, ai, blockchain, MedTech, and psychedelics. Lots to talk about there if we want to. And and, and he is with f h e dot org, which is another interesting thing. So welcome to the show, Rand. How you doing? And where are you doing?

Dr Rand Hindi (00:04:59):
Thank you. Nice to be here. I'm calling from Paris, actually.

Doc Searls (00:05:03):
Paris. Oh, so you're even farther advanced. We can make a bigger earth sandwich.

Dr Rand Hindi (00:05:06):
I'm an hour men are

Dan Lynch (00:05:08):
Far ahead of me,

Doc Searls (00:05:10):
<Laugh>. Excellent, excellent. So, so so what I left out of your LinkedIn bio, which is where I read it. Why don't you give us more of your, your background, then we'll go into what ZAMA and F H E and everything else is about, because you say you started at 10 and made a social network at 14. Pretty cool.

Dr Rand Hindi (00:05:33):
Yeah, I mean, I started coding when I was 10 years old. That was in 1995. And, you know, at the time, we didn't have all the tools we have today. I think, you know, my first website that I built was, I don't even think we had c s s, right? So like, you know, all the layout was like, you know, using H T M L. P H P was just the beginning. And so I started doing a lot of these things for fun and with a friend. We started posting pictures of you know, high school parties that we went to on a website. And that turned into a social network, which was a great experience because the social network in the nineties looked pretty similar to what we have right now. Just a lot less you know, fluid and a lot less fun with a lot less people.

(00:06:21):
Well clearly. So yeah, so I think, you know, that's how I got into developer into development. And then after that, I went to study computer science. But the thing is that I was so bored learning how to code because I had already been coding for a few years by the time I went to college. And that's when I started getting interested in artificial intelligence. So I started doing machine learning in 2003 and eventually ended up doing a PhD in machine learning applied to biology. So that was, you know, sort of like the, the the continuity of my interest in everything related to data. And then after that effectively started being an entrepreneur, I I did one job interview out of college. And I remember the the interviewer at the end said, look, you're pretty good technically, but you're never gonna fit in a company. You should probably start your own. And I was like, I knew it. You know, I knew that I was like an entrepreneur. And from that point, I only started companies. So I was running a machine learning company with a focus on privacy that got acquired a few years ago. And now I'm running a new company, Zama, working on this new technology called the Homomorphic Encryption.

Doc Searls (00:07:32):
So so you've been in, you've been in Paris or Francis whole time is, I see you done a whole lot of things in, in Paris. I,

Dr Rand Hindi (00:07:40):
I grew up in Paris then I lived in London and in New York for a few years, and then came back to Paris just before Covid actually when I sold my previous company.

Doc Searls (00:07:51):
So, so, so, so tell us about about zma and what you're, what you're doing with that. And, and we can branch off another topics from there.

Dr Rand Hindi (00:08:02):
ZMA is a company working on a technology called a homomorphic encryption. So homomorphic encryption is a type of encryption technique where you can do processing on encrypted data without actually having to decrypt it. So imagine, for example I am a user of chat G P T, right? Everybody's talking about chat G P T. Right now, when I'm sending a query to chat G p T, all of my data is being processed by open AI's servers, right? They're the one running the actual machine learning models. And then they send me back a response and I can go back and forth. What that means is that the company who's running the service has access to all of the data of the users using that service. With homomorphic encryption, you could do the same thing, but keeping the data encrypted end-to-end. So I would encrypt my data on my side, on my machine, send it to, you know the provider of the AI service.

(00:08:58):
They would process it blindly encrypted, produce a response, which itself is encrypted. So Chad G p t gets an encrypted input and produces an encrypted output, and I can decry the output. So from my perspective, nothing changes, but now the data is encrypted during transit and during processing. And so why is that important? It's important because when you can do that, it doesn't matter if a government subpoenas the company for data, they don't have the keys to decrypt the user data. It doesn't matter if a hacker breaches the server security because there is nothing for them to steal. There is no data visible. There is, it doesn't matter if an employee goes crazy because they cannot snoop in on any kind of data effectively, it removes the need to trust a service provider with the security and the privacy of their data we're sending to use the service.

Dan Lynch (00:09:47):
Wow. that, that, it's really interesting concept, Brenda, I, I, I like it. It, it goes beyond the, I mean, we've heard a lot about people's storing data unencrypted, you know, you see these, you see these stories all the time. Somebody's got into a database and got x amount of credit card numbers, passwords, whatever it might be, cuz they're not storing the things encrypted. But you're going beyond that, you're saying actually, cuz even then people would decrypt the data to process it and then encrypt it again to store it. But you're saying it never gets decrypted exactly. On the server.

Dr Rand Hindi (00:10:16):
Exactly. The, the way we like to think about it is, you know, how on the web we went from, you know, no encryption at all to encrypting transmission of data between client and server. So H T T P to h t TPSs mm-hmm. Well, imagine if the data stays encrypted even during processing. So imagine if, like you could extend, you know, H T T P S to end-to-end encryption for anything you do online. And this is really what we think about when we think about homomorphic encryption. We see it as end-to-end encryption for everything we do online.

Dan Lynch (00:10:48):
Hmm. Excellent. How, how does this affect kind of processing speeds and stuff? Cause I imagine it puts more load onto your C P U, does it in some way?

Dr Rand Hindi (00:10:57):
That, that's a great question. You know, when, when we talk about a technology that sounds like a magical solution for too many problems, the question is always why aren't we using it? And the, in the case of homomorphic encryption, there are three reasons. The first one is that it was way too complicated to use unless you had a PhD in cryptography, it was impossible for you to use this technology. The second problem is that you were very limited in terms of the kind of operations you could perform. So you could maybe add numbers, you could multiply numbers, but you couldn't like, you know, run a database on encrypted data or run a machine learning model and encrypted data. And finally, the third problem is performance. It basically takes about a million times longer to process data encrypted than if it's not encrypted. So, you know, great idea, but if something that used to take a second now takes 11 days, it's not exactly user-friendly and webscale you know, if you think about it. So these are exactly the things that we had to figure out in our company effectively by building tools for developers that makes it easy to use homomorphic encryption, even if you know nothing about cryptography, while also working on making the technology able to do any kind of computation at any kind at a much, much, much faster speed.

Dan Lynch (00:12:11):
Hmm. Excellent. I, I was reading it earlier on your docs are really good, by the way. I love the stuff you've got on the website. It was, it was great. Even for, even for someone who's not an expert in this, for me, I found it really fun to read some, I loved that one of your blog posts started with, I must confess, I've committed a crime. And I'm, it's like something, I was like, I'm gonna have to see, see more of this. I wanna know what happens. But yeah, it was really good. But I was looking at the, yeah, the stuff you were saying about it could be maybe five years or so before the, the kind of the hardware catches up. Is that a fair kind of timeframe?

Dr Rand Hindi (00:12:43):
Yeah, I think five years is the, the timeline we have in mind for being able to run a very large AI model like chat g p t on encrypted data, because, you know, this is kind of like the worst case scenario of what you want to do. Encrypted the hardware, which is gonna be accelerating homomorphic encryption by a thousand x at least is gonna be available in a couple of years. But we think we need a little bit more than that in order to get to, you know, large language model AI scale usage. But in the meantime, we can do many other things, right? We can already use homomorphic encryption for blockchain because a blockchain is a slow, expensive computer that actually runs very simple programs. So it's kind of like the exact opposite of AI in the cloud where things are huge, fast and cheap. So we can run smart contracts and blockchain using homomorphic encryption today. You can also use it for all kinds of like, smaller scale applications like, you know, retrieving a valued and encrypted database or filtering images encrypted end-to-end and even maybe doing spam detection on your encrypted emails, for example.

Dan Lynch (00:13:53):
Hmm. Excellent. I was, I was, when I was reading through your stuff before, I noticed that I was thinking about the whole quantum computing situation and how that could, could impact all of this. And I noticed in your documentation it said that homomorphic encryption helps to mitigate against attacks from things like fr from quantum computing and stuff. Can you tell us a little bit about that and, and how that kind of stuff works?

Dr Rand Hindi (00:14:19):
It doesn't really help mitigate against quantum computer attacks, it's just that homomorphic encryption is a scheme which is post Quentin. What that means is that even a Quentin computer cannot effectively break the the encryption itself. So if you look for example, at R S A, which is based on vectorizing large numbers a quantum computer can factorized number very efficiently. So that's why R S A is broken with a quantum computer. Mm-Hmm. But homomorphic encryption like the other by the way, new standard for post quantum cryptography are using a different type of mathematical problem. Something called learning with error which effectively makes it impossible for a quent, even for a quantum computer to break this encryption. Like there is no algorithm that you can figure out with a quantum machine that could break that. At least not that we know of right now.

Dan Lynch (00:15:10):
Mm-Hmm. <affirmative>. And, and would, would, I mean, we talked about the, the advancing hardware make this stuff easier to do, to do the encryption as a process. Would quantum be app applicable in that realm as a, as a way of speeding up what you're doing for servers and stuff?

Dr Rand Hindi (00:15:25):
Absolutely. Absolutely. The main bottleneck when you're doing homomorphic encryption is computing what's called a Fourier transform. So effectively you're trying to multiply two large polynomials, which are two vectors of numbers mm-hmm. <Affirmative>. So you take two like list of numbers, you try to multiply them very quickly, that takes a while. If you do it, you know, element by element. So you, you only something call that a FIA transform a Quintin computer can actually do a FIA transform very efficiently. So honestly, I'm actually looking forward to quantum computing, not because of everything that it will break, but for everything it will speed up.

Dan Lynch (00:15:59):
Yeah, that's a great point cuz a lot of people, I mean, I, I have some friends who were very into cryptography, some mathematician friends and stuff, and, and they're we had a discussion about this a actually at a Linux user group not so long ago. And lots of people were worried about that. You hear a lot of fear Mong grain and stuff about the, the Quantum's gonna destroy encryption. It's gonna, you know, once we've you know, the hackers have got quantum capabilities in their computers, they're just gonna cut through all our encryption. But it sounds as though it, it, it's a good double-edged sword. It's gonna work both ways in some ways.

Dr Rand Hindi (00:16:30):
Listen, if in 10 years you didn't upgrade your system to move to a post Quentin cryptography scheme, then you're probably not very good at security and you deserve to be hacked. Honestly, <laugh> mm-hmm. <Affirmative>. I mean, it's, I, it's a terrible thing to say, but like, imagine if you're not patching your servers for 10 years, hey, you know, you're gonna get hacked for sure. So it's the same thing with cryptography. No cryptography is broken. You have to upgrade to a new modern way of doing cryptography, which is better. And this is the same thing here. So I'm not afraid of Quentin computers in fact, I'm investing in companies doing Quentin computers. It's something I've been actively looking at. Yeah, I'm very excited about the promises of that technology. I think every problem it creates, we can figure out we need technology.

Doc Searls (00:17:16):
So I have some timeline questions, but first I have to let everybody know that this episode of FLOSS Weekly is brought to you by Kolide. That's Kolide with a K Kolide is a device trust solution that ensures unsecured devices can't access your apps. Kale is a big news. If you're an Okta user, Kolide can get your entire fleet up to 100% compliance. Kolide Patch is one of the major holes in zero trust architecture, device compliance. Think about it. If your identity provider only lets known devices log into apps, but just because a device is known doesn't mean it's in a secure state. In fact, plenty of the devices in your fleet probably shouldn't be trusted. Maybe they're running on an out of date OS version, or maybe they've got unencrypted credentials lying around. If a device isn't compliant or isn't running on the Kolide agent, it can't access the organization's SaaS apps or other resources.

(00:18:15):
The device user can't log into your company's cloud apps until they fix the problem on their end. It's that simple. For example, a device will be blocked if an employee doesn't have an up-to-date browser. Using end user remediation helps drive your fleet to 100% compliance without overwhelming your IT team. Without Kolide IT teams have no way to solve these compliance issues or stop insecure devices from logging in. With Kolide, you can set and enforce compliance across your entire fleet. Mac, windows, and Linux Kolide is unique in that it makes device compliance part of the authentication process. When a user logs on with Okta Kolide alerts them to compliance issues and prevents unsecured devices from logging in, it's security you can feel good about because Kolide puts transparency and respect for users at the center of their product. To sum it up, COE's method means fewer support tickets, less frustration, and most importantly, 100% fleet compliance. Visit Kolide.com/floss to learn more or book a demo. That's K O L I D e.com/floss. That's K O L I D e.com/floss.

(00:19:37):
So so ran when I, when I look at the, the Zama website and I look at the banner across the top, the menus there, I'm used to seeing products <laugh> right off top with most companies, right? And I see community, you've got a lot of that documentation, GitHub and more, and there's a lot of great explanation there. And I'm wondering what, what the, I mean, you said that it's gonna take several years before homomorphic encryption becomes normalized for a variety of reasons you just listed. What are the landmarks here and and how are you, what are you looking at for the business of ammo as it slowly moves or rapidly moves toward whatever it's going to become?

Dr Rand Hindi (00:20:19):
Monomorphic encryption is not a new idea. Actually it's been theorized in the seventies by the inventors of R S A because R s A is actually homomorphic when you use a specific version of it. So it's something people have been wanting to build for a long time. My co-founder Pascal PA is also a very famous cryptographer who invented the first additive scheme in the nineties. So the first time you could add encrypted numbers, homomorphic and so the holy grail was can you actually do additions and multiplication on encrypted data? Because if you can do that, you can basically compute anything. And that only happened in 2010 with someone called Craig Gentry, i b m at the time, who invented the first fully homomorphic scheme, meaning you could do both additions and multiplication and encrypted data. So this was a huge revolution because nobody knew if this was even possible.

(00:21:11):
 So from that point, it was a whole bunch of new scheme that were much, much faster. And the one we build at ZAMA specifically is called T F H E. And it's like a modern way of doing homomorphic encryption, which is much faster and much more convenient than what existed before. So like the timeline right now is we are with homomorphic encryption where AI was 10 years ago. The technology works, we figured this out, we're building the tools, and now we have to think about the use cases and applications and make that, you know, mainstream. And this is what's next for Zama effectively. So when we started a company, you know, in 2020, we had a choice. We could either build a specific business case that was applicable with the state of homomorphic encryption at the time, or we could make a long-term bet that this technology was gonna be used everywhere eventually, and focus on building tools for people to use this technology in their applications. In other words, instead of addressing companies, we're addressing developers. And, and so this is why the company isn't like, you know, a product solution enterprise kind of navigation bar. It's really a developer's first company. Everything we do, we open source. Everything we do, we try to maximize the developer experience so that this technology gets included in as many things as possible.

Doc Searls (00:22:39):
Wow. There's it it, it's interesting to me that you're saying that this is it, it's going to be universal at some point. And I, I'm old enough, so I remember, I mean, one of the first clients I had was in the early eighties, and it was a fiber company. It was called Fiber Land, and they were saying, fiber's gonna be everywhere, right? <Laugh>, it was 40 years ago, but it did, it is now. I mean, functionally it is, I mean, the first time I actually did some work in Paris in the mid nineties. And at that time, Cisco had one, I think 500 megabit fiber line crossing the Atlantic that they were working off of, you know, and, but now it's coming to many homes and it and Crea. But the same has happened with, I remember George Gilder in like the midnight, you saying, data's gonna be free, never charged for data data's, they're gonna have abundant data.

(00:23:28):
We're gonna have abundant speeds. We're gonna have abundant processing. I just gotta do a laptop I'm working on here. It's got, I dunno, is it a trillion transistors in it? Something like that. And, and now we saw it with blockchain to a limited extent, but really with AI now we have giant ai and, and everybody, the top tops of their heads are coming off because they're thinking what could be done with AI and and encryptions, or it's been one of those words that's been around for a long time. I worked with Phil Zimmerman on mm-hmm. On P g P a thousand years ago also when he had a, briefly had a company and he was busy fighting r s A at that time. But encryption it, I mean, it, it's really interesting to me that I haven't, I haven't, I hadn't heard about about homomorphic encryption already.

(00:24:18):
You've sold me on it now. I can see how it moves problems into new spaces, <laugh> and solves it in some others. But I'm, I'm really interested in how community is being developed now, because that's, that's what you're looking for. You're looking for scaling community, which is really interesting for us, because that's what we've got here, <laugh>. Right? So this is a, a bit of a sales call for you. I hadn't thought about it that way before the show, but that's really what's going on. I mean, if this is, if this is the future, how do people, how do you sort of onboard this, have it fanned out into cr into GitHub and, and and turn into business? I mean, it, it's interesting to me also how many businesses you're investing in. So, so somebody, how, how, how is, how is community happening? It's hard to make. I mean, I, I've been, I've seen so many, I've covered so many communities with Linux Journal, they're all different mm-hmm. <Affirmative> and even within sub-communities, and you can fork everything and those turn into new communities. So, so how's it going for you with that?

Dr Rand Hindi (00:25:20):
The nice thing about homomorphic encryption is that because nobody actually could use it. There are very few people who were actually in this community to start with. So when we started a company, there was maybe a couple hundred people in the world interested in this technology. Now there are a few thousand. So it was very easy to get everybody around the table very quickly. But building, you know, an open source community is very hard. It's something I've done in my previous company as well, and it requires exceptionally good libraries, co libraries that people can use. So, you know, if you don't think about developer experience, the same way you think about user experience when building a consumer product, you know, you're gonna make the job of a developer hard to integrate this. And nobody likes to do hard work. People like to do pleasant work, right?

(00:26:07):
So, you know, that's the first thing we said. We said, we have to make this so easy that a developer who never heard about homomorphic encryption can integrate it in their products without changing anything about their habits. So that means, for example, being able to take Python code and convert it into a homomorphic equivalent so that you don't have to learn a new language. That means, you know, automatically integrating with a whole compiler tool chain so that, you know, you can convert your machine learning model into a homomorphic equivalent without actually changing your toolkits. So that's the first thing, making it easy and fun for developers to use. The second thing is, and I know this isn't sexy, but it's documentation. Look, I'm, I've been a developer my whole life. Every time I had to learn a technology and I looked up the documentation and I see something like impossible to understand that isn't comprehensive.

(00:26:57):
You're like, okay, maybe if it's my job and I have to do it, I'll have to learn it. But if, if I want to do it, like to try it and to have fun, I'm probably gonna give up. So I think, you know, documentation is like, if you have bad marketing on your website, people are not gonna try your product. If you don't have good documentation, developers are not gonna try your library. So I think that's like a super critical part. And finally, to be honest, you just have to want to be involved. I think the fact that I'm a developer and I'm the CEO of the company, it just creates a culture of like, wanting to do what's right for our users, which are developers. And that means I like to hang out to developers. I like making demos. I like, you know, answering support questions on this core when people are trying to do something because, you know, it's, it's what I would've liked if I was on the other side of the table. And I think if you have that, honestly, you can start growing an organic community of people who believe in what you're doing. And then, you know, from that, marketing takes over and then it becomes, you know, a bigger machine. But the initial part, the first thousand users of your product or service, you know, there has to be organic. You cannot pay for them.

Dan Lynch (00:28:05):
Mm-Hmm. <affirmative> that, that's, yeah, it's a really interesting point. I think you mentioned that this has grown from say, a couple of hundred people who were interested in it when you started off into like few thousand, and then that could snowball into more, I'm interested in the I think among developers, among tech people. We love the idea of encryption and privacy and all the rest of it. But I, I was mentioning to doc earlier about the, the kind of, without getting political about the situation in the UK right now, where the government is trying to effectively outlaw encryption in some places, and that's a bit dramatic, but you know what I mean? They, the whole classic arguments that we've had in the past in other, in other places where the governments and other large entities have said, we need backdoor in your encryption and so on.

(00:28:47):
 I, is there a kind of a ah, I'm trying to of a good word for it. Is there a kind of a, a, a a a public kinda mindset that needs to change? Do we need to win kind of hearts and minds of people to the idea of privacy in a way? Do you think that's something that could happen? And how does that intersect with, say, business? Because I was thinking about, say, like, you mentioned models, examples such as voice assistance, for example. Like, I don't know, like Google for, let's say Google's got voice assistant, they have what's the, you know, incentive for somebody like Google to say, let's make all of the, the this private with home and offering encryption when a large part of their business is getting your data, and I mean, possibly selling it or using it to advertise to you and stuff like that. Is there a, is there a, a kind of a, a case that you have to win people over a little bit there somehow?

Dr Rand Hindi (00:29:37):
So the thing about businesses is, the nice thing about encryption is it doesn't prevent you from using data. You just use it without seeing it. So you could still, you know, do a personalized recommendation. You could still do advertising, which is targeted. You're just not gonna know who you're targeting with what, so the user would still get the targeted ad. It's just that Google wouldn't actually see the data of the user and wouldn't know which ads it actually sends to the user. Mm-Hmm. <affirmative>. So that's important. Homomorphic encryption doesn't prevent any data usage. It simply removes the need to secure the data. So it, it's, it's a solution. It's not like a, something that prevents from doing something. But let's put that aside for a second. I actually don't believe that people should care about privacy. I know that's a bit of a contradictory thing when you're running a cryptography company, but a lot of people believe that privacy is something you have to take into your own hands.

(00:30:31):
I don't buy that. I think that people shouldn't care about privacy, not because it's not important, but because it should be guaranteed by design in every product they use. And so the question is, who are building those products? It's not companies, it's developers in those companies. So really the person you want to convince is the developer. If developers believe that privacy is necessary for their users, they will push to put privacy and cryptography in their products. They're developing. Developers are the enablers of privacy. Not companies, not governments, not people, developers. So if we do a job right, developers will push back against governments. People will not even know it's happening the same way by the way that, you know, with the first crypto wars doc, you know, you talked about Phil and p g p remember the cypherpunks in the nineties when they pushed back against the ban on encryption from the US government, you know, with a clipper chip and all of that. You know, it wasn't the main, the mainstream people that cared about that. You know, we didn't have like people in the street pushing against that saying, Hey, we let us encrypt our data. It was cryptographers, it was developers, it was, you know, technologists that pushed against it. And that won, and that forced the government to change their stents. It's exactly the same thing now. Exactly the same thing. Now. This is just the crypto wars of the nineties, just a second time over.

Dan Lynch (00:31:57):
Yeah. But it, it's like the sequel actually say <laugh>, it's a, it's a sequel. Yeah, yeah, definitely. I wanted to, to yeah, to kinda give you some space to make that point. Cuz obviously on this show, we, we strongly believe in privacy and and so on. But it's something you hear a lot now from politicians and, and other people that it's the old classic argument. You know, you're not doing anything wrong. You don't need to hide it and all that rubbish. But there we go. So I'm interested in the tools. You talked a little bit about developers getting 'em involved and stuff, and I had to look through your GitHub stuff and all that that you've got on there, which is really great. And you've got a, a few different things on there that I thought it might be worth kind of talking about a highlighting you, first of all, you've got a, a product called Concrete or project called Concrete, which is really interesting. It's, it's kind of a compiler for all this. Can you tell us exactly what it is and how it works, or It gives us an overview.

Dr Rand Hindi (00:32:44):
The idea of concrete is that we take some high level code Python for example, and we automatically convert that into homomorphic equivalent, meaning we convert that into a circuit of cryptographic operations that you can run on encrypted data. And so it's a compiler because it takes some high level language, converts it into some intermediary representations in this case, like a homomorphic representation, and then it compiles that to binary code for whatever targets hardware you're using G P U C P U asic in the future. So effectively by using a compiler, you remove the need for the developer to learn all of the underlying layers of cryptography and to optimize the hardware. You basically just like tell them, Hey, just write Python code like you always did, and literally just add this one line function to your code, which is compile circuit, and you're done. That's it. That's the only thing you have to change.

Dan Lynch (00:33:39):
Hmm. Excellent. And, and what's the kind of adoption been like for that? Do you, I mean, are people taking it up? Developers interested in this?

Dr Rand Hindi (00:33:47):
Definitely. I mean, it's, it's pretty much exponential. I mean, it's not as big as AI is a field yet, so I, I think, you know, it's still gonna take us a few years before we get a hundred thousand, you know, people using it. But we went from like zero to 3000 people using it in like a year and a half. So that's actually pretty fast when you consider it that this is such a niche new technology. So I'm very, very confident that we're gonna see more and more developers start to use that, especially now that we've got those tools available.

Dan Lynch (00:34:21):
Mm-Hmm. <affirmative>, right? Yeah, definitely. I mean, that was one of the things I was so impressed by is the, the tools that you've got on there which you talked about lead in developers towards that kind of thing. I was really interested in the, you've got the rust implementation as well for UI stuff, which is quite cool. Yeah, that's, I love that looks really cool.

Dr Rand Hindi (00:34:37):
I, I mean, Ru Rust is my, is my, is my Love, I mean, menu loves <laugh> 2015 you know, like everyone in the nineties I started coding learning c and c plus plus, and that was great at the time, you know, I have to say honestly, but it wasn't really meant for modern applications. So eventually, you know, I moved out from, from c plus plus I started doing some Scala which was really nice as well. Scales a beautiful language, but if you want very high performance, you have to use something more low level. And the, and the jvm the scales targeting is not as good as, you know, just running code directly in the machine. And so when I found Rust, it was like obvious, like rust is like c plus plus performance with all the modern features of a language and the kind of like, fun that you would get with like functional strongly type languages.

(00:35:26):
Like in my previous company, we started doing machine learning and rust. So when we started Zama, it was a no-brainer that the cryptography should be in rust. But even beyond that, you, even beyond just you know, promoting Rust as a language, if you think about what breaks security, it's often memory issues, honestly. Right. even open SSL recently you had a memory buffer overflow discovered, like open s ssl, like everybody's using that, right? So moving from cc plus plus to Rust is a way to also solve that problem. So I think, you know, if you're gonna be writing cryptographic code where you need performance and an extremely high level of security, then you don't want to have to think about memory issues in your code on top of it.

Doc Searls (00:36:16):
You brought up since on Rust an earlier show we talked about the controversy going on right now in around the Linux kernel where a lot of people want rust in the kernel for the reasons you just gave, I think it's closer to the medal and all that and opposition to it by Linux himself to some degree. Do you have any thoughts about that? I mean, it, it, and, or it might just be, you know, you know, fit to purpose if the purpose is good for that, you know, works. I just wonder if any thoughts about it? I

Dr Rand Hindi (00:36:51):
Think anything that goes into the Linux kernel has to be thought about extremely well. You know, like it's you're changing something so fundamental that, you know, you cannot afford to make a mistake and rewriting the whole kernel in rust could be done, but you know, it's years of work that you need to potentially recheck because, you know, the, the, the kernel code has been checked and rechecked and rechecked, right? So there is a kind of say in engineering that if it works, why change it in a way? However new features could be built in rusts. There is no reason to use c plus plus to build new features into the kernel. You could ride them in rust directly. But then you have the issue of interfacing between C and Rusts. And while this is completely possible and it works fine, it does add quite a bit of complexity, right?

(00:37:37):
 And it's not necessarily very well integrated into all of the like, tool chain for like, you know, building for CI for all of those things. So I think there's still a little bit of like sea rust interfacing work that needs to be done if you want to build large projects that combine both languages. But I mean, hopefully, you know, that's something that's gonna be sold. Anybody doing Rust has that issue. We have that issue. By the way, our compiler uses a framework called M L I R, which is in c plus plus. So we literally have that c plus plus compiler framework compiling Titan codes into Rust Code and calling a Rust library behind the scenes. So it's doable, but, you know, our compiler is nowhere near the level of complexity of Linux, right. <Laugh>, I think, you know, we're talking about very different code bases here.

Dan Lynch (00:38:27):
Yeah. There's a lot of lines in, in the kernel, obviously. So one thing I noticed, Ronda, I was looking through your stuff. You, you've announced a, a second round of the Bounty program that you, you guys are doing seems to be going really well. Can you tell us a bit about that? And if people are interested how they could maybe get involved?

Dr Rand Hindi (00:38:46):
Yeah, actually that's part of, you know, how to also engage more with the community. Usually when people think about bounty Bounty programs, they think about security bounties. You find a bug, you report it, and you basically get a bounty. But we think about it as interesting problems for people to solve as well. So it's more of like, you know, a, a quarterly competition. So every quarter we put out a list of interesting problems that people can solve. So it could be building an application and using, you know, homomorphic encryption. Like for example we were asking people to build a RegX engine in homo in f h e cool use case, right? You need that if you wanna do any kind of like encrypted string manipulation, for example mm-hmm. <Affirmative> all the way to, you know, improving the cryptographic primitives in the open source implementation to even, you know, building F p g acceleration for homomorphic encryption.

(00:39:36):
So depending what you're interested in, you could try to solve one of those bounties and basically get your work integrated into our stack as well as of course, you know, getting the price and the price. You know, it goes from, you know a thousand dollars for writing a tutorial that we put on, on our blog to a hundred thousand dollars. If you're solving one of the bigger problems that we're currently working on, I think we even have a bounty where you can get like hundreds of thousands of dollars depending on how good your implementation is on acceleration, for example. So yeah, I, I think, you know, the Bounty program, if you're trying to learn f h e, it's a great way to get started because you'll have like a bunch of cool challenges to work on. If you're a cryptographer doing this professionally, and you want to like, try your new ideas, there are more complicated ones that could, you know, be a good use of your research time or, you know, if you just have a cool idea and you need a side project, take a shot, you know, it's it's it's it's fun and potentially quite profitable as well.

Doc Searls (00:40:40):
So, I, I wanna go a little bit further into F H E as an, as an ecosystem and all. But first I have to let people know that we have this thing called Club TWiT. And Club TWiT is a, is a great part of the way our entire network works and joining. It's a great way to support the network as a member. You get access to ad-free versions of all the shows on twit, as well as other benefits. There's a bonus TWIT plus feed that includes footage and discussions that didn't make the final show edit, as well as bonus shows we've started, such as hands-on Mac, hands-on Windows, ask Me Anything. And Fireside chassis, some of your favorite twit guests and co-hosts. And because I haven't mentioned, Lennox said this is important. As FLOSS Weekly listeners, you may be interested in checking out another Club TWiT exclusive show, the Untitled Linnux Show, hosted by our own Jonathan Bennett. So sign up, join Club TWiT. It's only $7 a month. Head over to twit tv slash Club TWiT and join today. And we thank you for your support. So I, I, I wanted to go back to, you've mentioned F H E A number of times is fh e.org is, is this website. There's, you've got a conference, you've got talks, you've got meetups you mentioned earlier about working with communities. So tell us a little bit more about how, how, how f e works and how you're integrating that with zma.ai as well.

Dr Rand Hindi (00:42:10):
FH e is just the acronym for fully homomorphic encryption. So that's how, you know, we call that it's like, it's like, you know, AI for artificial intelligence, iMorph encryption mm-hmm. <Affirmative>. and when I started a company, you know, I was thinking about a good way to rally the broader community, not around the company Zama, but rather around homomorphic encryption. And so very fortunately, to be honest, like I was able to put my hands on the F h e org domain. You don't get those three letter domains, you know, <laugh> per sale that often, especially not for such a low price, to be fair. And so I was like, let's just build a community around that. You know, it's a, it's a good name. Let's just start doing meetups. And so initially it was just a monthly meetup, people talking about homomorphic encryption, presenting a paper, showing a demo, and that grew very quickly, you know, from like, you know, a hundred people the first time to a thousand people, you know, like six months later.

(00:43:01):
And at that point we're like, okay, you know, let's try to put more resources into this community so that it's not just a meetup, but is like really the central focus point for anybody who's interested in homomorphic encryption. You wanna learn about homomorphic encryption, go to the ft org website. You'll have a bunch of resources to get started. You've invented something new, you wanna present it, come to one of our weekly paper sessions and talk about it to all other cryptographers. We have a yearly conference as well that we do physically. So, for example, this year is a month ago. It was in Tokyo. So you know, really fun places each time as well. And really fxc.org isn't Zama, it's just that we effectively moderate it and pay for it pretty much. But it's something that's completely open. Anybody could participate.

Doc Searls (00:43:50):
So I, I find myself imagining as the real internet of things, not the Apple, Google and Facebook are the things and the Google of things, but the real internet of things where all of our property, the, you know, the microphone I'm talking to, the, you know, the, the loose keyboard, the other things become members on the internet and and encryption happens between them, but the encryption becomes homomorphic at some point. That success you're talking about is gonna take, you know, a few years to happen, but will happen, is gonna evolve a lot of big companies. Mm-Hmm. <affirmative>, and they're gonna be doing things, they're gonna be part of your community. And then Analytics Foundation shows up, <laugh> and the Linux Foundation says all these big companies are all developing things different, their own different ways. Can we mind if we take over this f.org thing you've got that's sort of a, I mean, I'm not saying that's a bad future. I think that is a typical future for anything that starts to take off. We're seeing it happen right now with personal wallets as self-sovereign identity is taking off as a thing. There's the Open Wallet Foundation, and in fact, I joined the Open Wallet Foundation for the purpose of making sure they care about individual people and not just about big companies all fighting each other. So I'm wondering if you've imagined out how that goes as, as the community grows and your events grow.

Dr Rand Hindi (00:45:18):
You know our event, our main event for fg.org, I think that's a good testimony to how little Zama is biasing this. We had 16 speakers, only one of them from was from our company. Effectively, anybody can speak, you just have to submit a top proposal, and there is an independent committee that isn't from Zama choosing who's gonna speak at the F H U org events. And so this year we had people from various universities. We had people from one of our competitor competitor talking at our ad event that we actually host and pay for. Try to imagine that we had people from Google, people from Intel, people from, you know, like all kinds of companies. And this is great. This is what we want, you know, we want this to be about homomorphic encryption. GI ZMA is just very fortunate to have their resources and the money to make this takeoff, but down the line, I'm really hoping that this will end up as an independent organization and foundation with its own budget and its own people kind of the same way that, you know, rust ended up having the Rust Foundation spinning out of Mozilla down the line.

(00:46:29):
Right? you could imagine that. So ZMA has incubating fhc.org, but eventually we're hoping it's its own thing. And, you know, and, and whether we partner with the Linux Foundation with Apache or with someone else, that will depend on how, how much rich in terms of developer that can bring, the more people are using homomorphic encryption, the better I think, you know, down the line. But again, you know, it's about developers. This isn't a community about companies. It's not a community about business. It's a community about cryptography and about using cryptography in your products.

Dan Lynch (00:47:11):
That's very, very cool. Y you've probably just answered this, but I'm gonna ask it anyway. You've mentioned competitors there and, and other companies and stuff. Are there many other companies in this space that, you know, I'm not expecting you to ma to name them necessarily, but are there other people doing this?

Dr Rand Hindi (00:47:26):
There are many companies actually. Especially actually, I think I would say there are three main companies us. There is a company in Korea and a company in America. Great people, all three of us, you know, have raised a lot of money and we all have like, you know, great photographers on board. So I think that like, you know, we're equivalent in terms of technical capabilities and in terms of skillset, but we're very different in terms of strategy. You know, we are the bottom up developer first company. The American company is a top-down enterprise company. The Korean company is more of an academic company right now. They're a little bit younger. And even though everybody wants to make a ton of money eventually, because there's a point of a company, let's be honest we're just having very, very different ways of getting to that point. And even once we get to that point, like, you know, I like them. The, they're good people, they're nice people to hang out with, I think we're gonna share the market more than compete on the market. This isn't a winner, takes all kind of market, you know, I think this is gonna be split between you two or three main players and a bunch of smaller players eventually.

Dan Lynch (00:48:32):
Excellent, excellent. I was wondering, thinking about applications for for this technology, you mentioned a few well you mentioned some already on your, on your documentation on your website, you mentioned a lot of, of potential future ones. What, what's, can you think of an interesting or, or, you know, maybe outlandish kind of future possible application for this that excites you that you wanna tell us about?

Dr Rand Hindi (00:48:53):
So, I, I think short term blockchain is a perfect application for that. Hmm. long term AI is gonna be the most, I would say impactful use of homomorphic encryption. Why? Because if you're gonna be talking and sending data to artificial intelligence systems all the time everywhere, right? You need to make sure that your data is secure. So there is no, i, I don't, I don't believe that there is mass adoption of AI on the scale that is necessary for this technology to really bring it value. Unless you figure out privacy. And this is what's happening now. Italy is banning strategy, G P T, because of of privacy. There are a bunch of people having data leaks because the data they sent ended up being used in the training set, which reappeared in a prompt from someone else. Like, you need to figure this thing out and this is what homomorphic encryption does. But this is gonna take a, a little bit longer because of the performance requirements. So I estimate that strategy G P T is doable, homomorphic in five years

Dan Lynch (00:49:54):
Mm-Hmm. <Affirmative>

Dr Rand Hindi (00:49:55):
In the meantime, you can do smaller machine learning models, you can do that today with our tools. But long term, if you want like C G P T encrypted end-to-end, that's five years short term blockchain. Why a blockchain by design is public. Everything you do, every piece of data, every computation is public because you need, you know, the people doing the computation to agree on the result. And so the way they did that is just sharing the result publicly, right? So that people could verify it. But that prevents you from having confidential smart contracts. So for example, if you want to hide what assets you own or if you want, you know, to do like a seal bid auction, you want to bid on something without everybody knowing how much you're bidding, because otherwise it's too easy to outbid you. If you want to, you know do finance on the blockchain confidentially without everybody knowing what you're trading and what you're doing or even simple things, if you wanna play poker on the blockchain, you need to hide your cards, which you cannot do right now. So I think blockchain has a strong need for confidentiality and privacy right now. And because blockchains are slow and expensive and because smart contracts are very simple programs, we can make it work even without an accelerator. So blockchain is a perfect early adopter for homomorphic encryption machinery, and AI is the large scale, long-term kind of like use case that we really need to crack for humanity to be better off.

Dan Lynch (00:51:22):
Hmm. Excellent. I'm thinking about, yeah possibly cuz of the stuff that I tend to get involved with like edge computing stuff as well. I know obviously you wouldn't, you wouldn't be processing, cuz we talked about the need for fast hardware and stuff, you know, you're not gonna be processing stuff necessarily as quickly on iot devices. Things like Doc mentioned is that a challenge for the future? Maybe something that could effectively, you know, well,

Dr Rand Hindi (00:51:47):
So I used to do a lot of edge computing, by the way, my previous company we used to do machine learning on the edge. Mm-Hmm. <affirmative> specifically for price reasons. So our, our kind of like idea was, well, if we cannot protect the data on the cloud, let's just not send the data to the cloud and process it locally on the device. So, you know, we were the first company, for example, to show how you could do a voice assistant, so kind of like a Siri Alexa on a raspberry pie, for example. So, you know, you could do it on like your washing machine or tv. And that company got acquired, by the way, by Sonos was one of our customers. And now your Sonos actually uses an edge based voice assistant, which is the technology my company built, my previous company built.

(00:52:28):
But sometimes you cannot run the computation on the edge because maybe you need multiple users to participate in the computation, or maybe because the compute, the computation is too big to fit a micro washing machine or like a small edge device. So you still need to do stuff in the cloud. Otherwise we wouldn't be using the cloud. And this is where homomorphic encryption comes in. So I think, you know, it's not about edge versus cloud right here. It's more about whatever you can do on the edge, just do on the edge. There's less stuff to do in the cloud that's better for everybody, right? But if you need to do something in the cloud, we have to enable you to do it privately. So, you know, one example of that, that is, you know, for example, all of those health connected devices you know, you have an Apple watch, I've got like an URA ring I've got like a connected scale.

(00:53:18):
 All of these things are connected to a dashboard, which isn't a cloud, and you need that dashboard to me in the cloud because you're pulling data from multiple devices. So it's impossible to do that on the edge, like it has to be in the cloud. Well, that is a great use case for homomorphic encryption. Just get, your device could be sending encrypted data. You aggregate everything, you do all your computation in the cloud encrypted. So you do all of like your machine learning, prediction, statistics, trendlines, all this stuff you want about your health, and you decrypt it in the browser directly of the user. So you're keeping the encryption decryption on the edge. But all of like, the computation and and aggregation of the data, which has to be done in the cloud, is now done privately with homomorphic encryption.

Doc Searls (00:54:02):
Hmm. You know, we're getting down toward the end of the show here. And I, I want to at least visit, if you've got a, a wonderful subs which is, you know, talks about your deal flow, it's called Deep Tech Deals. You are involved in a lot of deals, and I'm, I'm impressed by that. And if I see a theme, I mean, there's one is that, as you put it, you're, you're very long on Ethereum. And, and I think because can program on it so let's, this is that at least briefly. And, and what you're thinking is about where that goes and where cryptocurrency goes. Cryptocurrency is kind of an doldrums now in the public imagination, but I don't think it is among developers. So just visit that for a a couple minutes and we'll see what we got.

Dr Rand Hindi (00:54:54):
Sure. When I sold my previous company I was fortunate, you know, to make some good money out of it, and I was like, you know, what am I gonna do with this funding? You know, I could put in a bank, but at the time it was like negative interest rates. So obviously that's not gonna happen. I could put it like all in stocks and crypto, but it's like super volatile. So it's like, you know what, I'm gonna invest in companies that I would've liked to build myself, but I just don't have time to because I'm already building something. Anyway. and so if you think about that, my personal interest is really in science and technology. I'm not a consumer guy. I don't know what's the next good, you know, social network to build, like, I don't know how to build the next TikTok, but I know what's gonna be needed for all those things to exist.

(00:55:37):
 And so I identified a few key trends there that I think are fundamental for the next decade. So one of them is what I call, you know, the end of Moore's Law. What comes after, you know, the, the C P U at some point, you know, you cannot be naturalize a transistor anymore, right? You're not basically getting the power out of just putting more transistors on a ship. And so what comes after that? Quantum computing photonics, but also open source silicon, right? Risk five, for example. Super interesting ecosystem that's popping up. And more generally, I would say like the unbundling of the C P U. If you take everything A C P U does and you put that into a specialized accelerator with a C P U core in the middle, that's powerful, right? So what if it wasn't just a C P and A G P U, but like more of like your chip on your Apple M one, you know, the Apple M one chip, it has 32 co-processor.

(00:56:33):
The C P U does very little of like the heavy lifting because you've got specialization. So I'm looking at a lot at that future of computing and Moores Law. The second thing is cryptography and blockchain. It's pretty obvious to me that the content, the concept of programmable money and the concept of like programmable assets in general is much better than like, you know, paper pushing type of like finance and stuff. Like, I don't understand why I still get paper receipts. I'm su I could, I could just get literally like an N F T as my receipt for everything I'm buying that would be so much simpler. I don't understand why I wanna make a bank transfer. It takes five days to go to the US why, what is happening during those five days? Like, someone explained that to me. And so when you think about money as a programmable entity, that opens up so many things, it just, it just streamlines so many of the rails of society that it's pretty obvious that, you know, this is something that's gonna be, you know, everywhere.

(00:57:32):
I don't know if this is gonna be Ethereum or Bitcoin or if it's gonna be some government, you know, digital currency that's gonna be adopted, but it's still gonna be running on some programmable, you know system. So big bullish on that. So privacy of course, very, very important. Privacy, I think is something that I've been very passionate about since I've been a teenager. So very bullish on that. And finally super bullish on Frontier Health. I think, you know, COVID has woken up the world to the fact that healthcare doesn't work. At least the way we've been doing it. We've been, you know, too slow in innovating on, you know, healthcare. And so now that people realize that there is a huge push to modernize everything better diagnostics, better treatments, better vaccines, better mental health you know therapies, everything's being overhauled. And that's like super, super exciting as well. So, so yeah, I think, you know, for me, I'm investing in what I'm hoping is gonna be an extremely technological driven future where we don't have any of the limitations we have right now. So yeah, it's faster machines, better infrastructure, and just, yeah, live a very, very old, very, very happy and very, very good health.

Doc Searls (00:59:00):
Wow. Well, we're, we're down pretty much at the end right now, and Dan and I have been sharing it back and forth. Questions we still wanna ask <laugh>. I have, I have one about your portable mri. You were talking about healthcare as one of, one of the things you've invested in or toward frontier Health, like even the talk about that I'm in the US where healthcare is horrible. And the data side of it is especially bad. I live in four places, <laugh>, I have four sets of doctors. They communicate by facts if they do it all. So there's just a lot, a lot we can continue to cover there. Is there anything we haven't asked you? You could, you could answer real quickly in in the last minute or so, and then we, we can wrap.

Dr Rand Hindi (00:59:49):
I mean, I think it's good. I mean, this was interesting. I think perhaps, maybe I just want to reiterate something which I think is misunderstood and important. Developers are the ones who decide which technology gets used and which doesn't. Regulators, they try to make sure people don't misuse it. Consumers, they don't care. They just want something better, faster, cheaper, right? Companies, they just wanna make money. Developers choose which technology gets used. So if you really want everybody to have privacy, the real thing you have to do is convince developers that privacy is important. Developers are the garons of privacy for everybody else.

Doc Searls (01:00:26):
You know, back when when Netscape open sourced their browser, which eventually became Mozilla actually was completely rewritten as Mozilla. But when they open source that, I had a long interview with Mark Andreesen when he was still at Netscape, and he had this one liner he just dropped in the middle of things where he said, all technology trends start with technologists <laugh>. And it's like, and it's the most obvious thing, and yet it's so easy to forget. And all the policy makers and all the business people think they're in charge of it. And they're not <laugh>. It's,

Dr Rand Hindi (01:01:01):
They're not,

Doc Searls (01:01:01):
The geeks run the show. Well

Dr Rand Hindi (01:01:04):
We run the show in the, in on, you know, in, in, in secret, right? We're the one decide where the future goes.

Doc Searls (01:01:10):
<Laugh>. It's a secret society that could not be more public. It's just the weirdest thing. And that's what, that's our, those are our listeners and viewers. So there we are. We always close the two questions, which are silly but interesting. What are your favorite text editor and scripting language

Dr Rand Hindi (01:01:29):
<Laugh>? My favorite text editor is I mean, I'm still using Sublime Text. I think, you know, that's, that's fine. It's great. And what was the second Lang the second question? Sorry. Oh,

Doc Searls (01:01:39):
Scripting language.

Dr Rand Hindi (01:01:44):
Well, that

Doc Searls (01:01:44):
Was the tone.

Dr Rand Hindi (01:01:46):
Yeah, I guess I, I suppose I would say Python.

Doc Searls (01:01:52):
P Python, yeah. Okay. That's great. Well, listen we often, if not always say we have to have people back, but you are, you're on a topic that even our back channel says we didn't know about that, and, and they're entertained by it and energized. So that's, that's a good sign. We will have to have you back in, you know, a year or something.

Dr Rand Hindi (01:02:15):
And we can talk about all kinds of crazy deep tech stuff as well. You know, we can talk about healthcare and we can talk about

Doc Searls (01:02:20):
That as well. Yeah, I, I mean, it's just in, I'm gonna subscribe to your thing. I mean, it's it's you seem to have a post a month. That's a pretty good pace. The ones that are daily, Doug, I <laugh> the big things, the big things happen, not don't, not on the daily scale, but they, there's a different pace to it. So thanks so much rant. This has been really great, and I really appreciate having you on the show. Thank

Dr Rand Hindi (01:02:46):
You for having me. This was great.

Doc Searls (01:02:50):
So Dan that was, that was a rock and roll show. I like that.

Dan Lynch (01:02:57):
Yeah. Really interesting to talk to, to round a great guest and, and he knows about so much different stuff. I mean, as you said, we, we probably could have gone off into all kinds of different directions there, but we tried to kind of keep it focused on the whole privacy thing. I do think it's, it's a really interesting time right now for the whole privacy. I love his point that he made about developers deciding on the technology that, that, that gets used and, and all of these kind of arguments that are going on right now about privacy. I did very quickly wanna mention I, I know I'm jumping into plugs here, which is terrible. No,

Doc Searls (01:03:30):
Please go for it.

Dan Lynch (01:03:31):
But I did wanna quickly mention for, I mentioned at the start of the show about the the stuff that's going on legally in the UK right now. There's stuff going through Parliament to try and enforce people to open up their encryption and so on. If you wanna know about that or you wanna find out more, there's a open rights group which I'm a member of. It's a lobbying group in the uk who lobby the, the government on these kind of issues. And they're much more knowledgeable than, than I am. But if you go to open rights group.org, if you're in the uk or you care about policy in the uk I would say check out open rights group. They're doing some great stuff about the online privacy bill right now. Online safety bill, sorry, not the online privacy bill. That's what it's called.

Doc Searls (01:04:12):
Yeah. Yeah. It's, it's, it's I, I may have mentioned this before, but I kind of knew a former chairman of the FCC in the us who said that he'd met with every member of Congress. And the one thing, none of the two things, none of them knew anything about were technology and and economics <laugh>. So, and asking them to make laws. Oh my God, it's probably not a whole lot different in the uk. Well, thanks for that. Next week we will have Bruce Perzon. Bruce is one of the original real major figures in the open source movement. He's the one, he's not the one who actually named it, but he's the first one to kind of get the ball rolling. And and, but he's, his focus lately is on space. And he, I'm on a, I'm on a list or two with him, and he knows so much about it. So <laugh> there, there, here. We're thanking each other for this and that. So everybody come back next week. Bruce's parents will be on. He's always good. We were gonna have him on last fall. Health stuff showed up and we had to move around a bit, but he's gonna be on next week. It'll be really great. So come back then. Until then, I'm Doc Cs. Thanks for being with us.

Rod Pyle (01:05:34):
Hey, I'm Rod Pile editor, editor-in-chief, of Ad Astra magazine. And each week I joined with my co-host to bring you this week in space, the latest and greatest news from the Final Frontier. We talk to NASA chiefs, space scientists, engineers, educators and artists. And sometimes we just shoot the breeze over what's hot and what's not in space, books and tv, and we do it all for you, our fellow true believers. So whether you're an armchair adventurer or waiting for your turn to grab a slot in Elon's Mars Rocket, join us on this weekend space and be part of the greatest adventure of all time.