The Data Stack Show - Shop Talk: Why AI Is Not Another Crypto

Episode Date: June 16, 2023

The Data Stack Show is a weekly podcast powered by RudderStack, the CDP for developers. Each week we’ll talk to data engineers, analysts, and data scientists about their experience around building and maintaining data infrastructure, delivering data and data products, and driving better outcomes across their businesses with data.RudderStack helps businesses make the most out of their customer data while ensuring data privacy and security. To learn more about RudderStack visit rudderstack.com.

Welcome to the Data Sack show, Shop.Costas. We have talked with people who built amazing data technology at companies like Netflix, Uber, and LinkedIn. But you and I actually don't record our talks about data very much. But we actually talk about data together a ton. And so Brooks had this amazing idea of just recording some of the conversations that you and I have before and after the show about data and our opinions on it. And really, this has been one of my favorite things that we do. So welcome to Shop Talk. It is where Costas and I share opinions and thoughts on a personal level about what we're seeing in the data space. And it really is simple.

We ask one another a question and the other one tries to answer it. So without further ado, here is Shop Talk. Welcome to the Data Stack Show Shop Talk, where Costas and I talk shop. Costas, it's been a while since we've done one of these. Lots has happened. I have a new baby in my household, so that's kept me away from the microphone for a little bit. But this is exciting.

Oh, yeah. Yeah, I miss pre-styling with you, so let's do it again. Okay, this was the question on my mind, which is a broad topic and kind of sounds cliche. We actually haven't talked a lot about AI in terms of like the latest hype, right? I mean, we've talked a lot about, we've talked a lot about ML ops. And we've talked to some extent about, you know, sort of AI applications at scale with several guests,

but those tend to be more technical conversations around why ML is difficult, right? We've touched on the ethical, which is interesting, but LLMs are, you know, it seemed like overnight they went from being something that was kind of cool to something that, you know, there are all these blog posts about, you know, there's petitions being signed by Elon Musk. Yeah. You know, Sam Altman's talking to Joe Rogan. Andreessen Horowitz is saying, you know, this is going to save the world. Other people are saying it's going to burn the world down. How are you thinking about this?

I've actually been wanting to, I almost texted you this the other day, but I was like, oh, we need to talk on the phone. How are you thinking about it? Yeah, how I'm thinking about it. How I think or how I feel. It's like two different things. Not necessarily like aligned. That's probably how a lot of people,

actually, that's probably a pretty articulate way to explain a lot of people. Like, how do I think about this? How do I feel about this? Maybe this. Yeah, yeah. I think like, first of all, the problem with like what is happening right now, now with AI is that I think there is a kind of

hype fatigue to the people out there, especially after the crypto craziness. So we had this so revolutionary thing coming that is going to change everything. Democracies are going to change. Banks are going to change. Everyone's going to be multi-billionaire or whatever. And suddenly everything collapses. And instead of that, we just have many people going to the court. I think it's interesting because you have this experience and immediately, like, as you still like experience that, you get into like an extreme hype that comes with AI and then you're like, is this different?

Is it the same? Because you can hear that, right? Like, there are like people saying, oh, it's the same thing. Like it's at the end. Yeah. Like people are like overselling this thing and at the end it's the same thing like it's at the end yeah like people are like overselling this thing and at the end it's going to be okay like same thing like ask crypto right like what is the use case of me like yeah what are we going to do now if you want my opinion like i

don't think it's the same thing it's quite different i don't know i don't know. I don't, I'm not sure that like we have figured out yet. Like what's exactly, let's say, the source of the value is going to be and how it's going to be delivered. But at the same time, I think that like the hype also hurts the whole. I agree. Like, it's almost like, I don't know, like, I feel like almost, I want to advise people to just quit, sorry, Twitter, like don't listen to the people there. Like, it's like, no, don't do that to yourself.

Yeah. Now, there are like things that are changing with these technologies that I don't think we hear that much about them because probably they are coming from industries that are a little bit boring to most people. Anything that has to do... I was talking with people that they're doing some amazing stuff in SureTech, for example. And there's like a lot of value.

Now is this value like going to make like humans obsolete? No, a hundred percent no. Like it's just, it's going to be, let's say different and like in some ways, like much more efficient and to be honest,, much more pleasurable to do your work. Because mundane tasks, that's okay. You don't like doing. So there's a lot of hype. I think it's hard to separate signal from noise.

I think we are going to see new business models coming out of that stuff, which is probably something I'm very curious to see how this is going to emerge. And I'm waiting for the hype to go away. So I'll keep working. And by the way, I'll give you an example, right? Like you hear like anyone talking about self-driving cars anymore? No, I was actually going to bring, I'm so glad you brought this up. That faces similar cycle. And there are actually some signals that made a lot of people say, oh, well, this whole

thing is over, right? Like GM shut down, you know, their self-driving car program or whatever, right? But there's still companies like Waymo doing incredible stuff, you know? Like, okay, I mean, in San Francisco, if you go out for a walk, you're going to see Waymos and Cruze

driving on their own all the time. I had my my father, like visiting from Greece and like, he saw like a car without like a driver and he was like, what the fuck is this? Like, yeah. Like,

you know, like we take it for granted, but like, it's like, it's not the reality, like for the 99.9% of the population out there. And the reason I'm saying that is because progress is made. But progress, let's say, the real progress happens much slower than what like the hype cycle is, right? So yeah, of course, like self-driving cars are out there. But it's not like one day to the next will substitute everything with self-driving cars.

They need to make sure that they can drive without accidents and they will slowly roll them out and see how they do. At the beginning, it's going to be in more quiet times the beginning it's going to be like in more quiet times and then it's going to be you know like big times like all these things so it's happening but like that's the thing like that's like i think like part of like human nature you know like we get like super super excited like at the beginning and then like it's like oh okay boring you know like it's just a car without the driver. Yeah. Yeah, I agree. I think it's, yeah,

I think it's really unfortunate that the timing relative to the crypto hype cycle and, you know, that baggage just sort of being attached in a lot of people's minds, I would think, to AI and LLMs. Because I agree with you, they are fundamentally different. There's fundamental differences.

Yeah, and I think the crypto story is really bad because there are many, how to say that, bubbles or ugly stories in tech. But it's one thing when institutional investors lose their money. It's another thing when grandmothers and grandfathers lose their money it's another thing when you know grandmothers and grandfathers lose their money

yeah and that kind of hurt a lot let's say the faith of like the broader

society has towards like technology yeah we make fun of like the we work like situation

for example but at the end like okay like who cared about we work right like is it like your I don't, like your dad or your uncle, like, no one care

about like what happened there. But probably if you talk to them about crypto, like they're aware of that. Like they know stories of like people who got hurt by that because they lost their money, right? But losing the money is one thing. The long-term, let's say, negative effect of that is people losing faith in progress and tech. But again, as you said, AI is different, but it's also early. So we need to figure out what that means and how we can use it and how we can make it part of our everyday life. Yeah, I think it's interesting.

I mean, this is a drastic oversimplification, but if you think about broad impact of these technologies at a baseline level. Crypto for the average person sort of encouraged some level of speculation, right? I mean, that was the entry point into interacting with cryptocurrency, right? And that's really sad to me because the idea around some of the things that blockchain and finance could do is really interesting from a transparency standpoint and digital currencies and stuff. But it centered so heavily on speculation. I mean, not everyone would have said it that way, but like that was the initial experience. Whereas with AI and LLMs, what's really fascinating is that even on a very micro level, it's seen as a utility,

right? Like I can ask, I mean, this is such a small thing, but I was like, man, I'm working on some, I'm stubbing out like a fake dataset and I need 25 like randomly generated alphanumeric keys. You know, it's like, okay, it took two seconds. You know, a chat GPT, right? Where it's like, oh, like, okay, I need to figure out the Google Sheets formula or, you know, random number and I forgot the formula, you know, all that sort of stuff.

And that's a very small example, but there's sort of immediate utility on even a micro scale with LLMs, which is really interesting. And so to me, there's just a huge amount of interesting potential. And I'm with you. I can't wait to see after the first round of companies fail that are building really thin layers on it, that people will figure out how to do themselves. After all those fail and then people start building new business models, it's going to be fascinating. Yeah, 100%. What's fundamentally different is, I think you put it very well, with AI and ML, you have something that's just lowering the entry tremendously to anything

that has to do with interacting with technology. Right. I think that was probably the promise also of crypto. But at the end, what was insane to me with crypto was that, okay, you get something, you get an instrument of value creation that on its own is, like, extremely complicated, which is investment in financial instruments, okay? Yeah, it's like, I don't, I don't understand that stuff. Like, I don't know.

And, like, okay, I'm, like, I mean, I have, like, a brain that, that like can't think, right? And you get and you mix it with something which is cryptography. Which, by the way, like I did study like as part of my degree. And like, dude, like I don't understand what they are talking about. Like I don't. Like it's like, and they come to you, okay? And they're like, okay, let me give you my prospectus like to decide to invest in what I'm building. And they are like down in an algorithm that usually goes through peer reviewing from crypto experts to figure out if this thing is worth or not. So you take a complicated thing

with more complicated thing it needs and you give it to everyone. And okay, give it, but don't attach it to their wallet. You know what's really weird about that? Is that there was a huge meme emphasis in crypto, right? Like Dogecoin and all of these derivative coins that got created that were essentially derived from online memes. And I mean, as crazy as it sounds, there are people who made a lot of money when Elon Musk tweeted about Dogecoin because it moved up like some minuscule

amount, but percentage-wise was huge, right? And it was just, I mean, that obfuscated the complexity and seriousness of this being tied to your actual bank account in a way that I think was really unhealthy. And I mean, you're right. A ton of people lost a lot of money. Well, that's the thing. What happened with crypto is that crypto was like a zero-sum game. Yeah, people made money because some other people lost money.

Simple as that. But like technology is fundamentally a zero, sorry fundamentally a positive sum game. Yeah. Or at least it should be. It should be something that there's more value generated

than what is put into it. That's why, for me at least, I'm so excited about technology and why I really like working with that and building stuff with that. Although I will say, if you start MumfordCoin, I will invest heavily in MumfordCoin. Well, that's, yeah, I don't know. I think it's too late.

Maybe we need something with Mumford and the AI. I don't know. I think it's too late. Maybe we need something with Mumford and the AI. I don't know. I have to think about it. But anyway, enough with Google. Let's go back to AI in the mail. I have a question for you on AI mail. I was going there myself. Okay.

Okay. Let's talk about some of the fears around it. And I'm going to draw a comparison for you that I've been thinking about recently. So there is a big thread of conversation out there around AI and LLMs around, okay, no one's going to be able to know what's real. How do I sort through the information that can generate all this sort of stuff. And we can talk about the specifics there, but the comparison I wanted to draw was that actually I feel like Google

or really just online search in general, to me, warranted some of the same types of concern. Not that the search engine wasn't trying to mitigate things that were obviously fake, right? But there's so much crap out there on the internet that you can find just by searching, right? I mean, how do you discern what's real? And of course, the machine is doing some of that for you. The search engine algorithm is doing some of that for you, but it's like, well, that's literally like the MLM is doing some of that for you. Like the search engine algorithm is doing some of that for you. But it's like, well, that's literally like the MLM is doing some of that for you, right? I mean, in some ways, there's this huge negative reaction.

It's like, well, it's not like this is a new problem. We're experiencing it in a new way. And in fact, ironically, it's not like we haven't faced this in a new way and in fact a lot of the ways that this stuff is being generated it's trained on all of the information that's out there right a lot of which not a lot of which some proportion of which you know isn't good information so what's your take from that standpoint to me it's there's a there's i'm not saying they're i'm not saying we shouldn't think wisely about how we use technology but this is not

to me a new problem that we've faced. It's actually a problem that we've faced for a while. Yeah, let me I'll tell you my opinion and I'll get a little bit philosophical at the beginning but

I think it's important for what I'm going to say about AI and the things that you talked about. Okay, reality. You use the word real. What is real out there? And what I can trust and how I can figure out what is real and what is not real. Now, the truth is that reality is a construct of the end. Okay? Like, at a great... like, a big part of what we are experiencing every day, okay, it's narrated to us. Okay? And that's not like anything that I'm not trying to say anything.

There's no conspiracy in what I'm saying. Like, it's just how like humans work, like our brains work. Right. And like how we relate with like the environment around us. Now, there is, although there is a very important thing here, and that's the concept of accountability. When I interact with you, Eric, you say something to me and I trust you. You say something that is honest. It might be a lie.

And maybe I find out someday that this whole relationship we had, the two of us was like a big lie right but i know who's accountable for that it's you yeah it's like it does right the same thing applies also like with with google right like yeah but there's like a legal system there is like a framework there are like committees that they look at them and try to regulate them, blah, blah, blah, like all that stuff, right? And of course, we get used also to trust them. The problem with AI is that today we need to trust something that doesn't have a very clear agency in our system, right? It's not an organization that is legally governed by a specific framework. And we have seen it working and failing and working and failing,

and we know how to deal with it when it fails and when it works, right? It's not a human that we know how to deal with a human when this happens. It's a bunch of weights and connections that comes up with an output and okay, who is accountable for that? But say I go to a model and ask it for, like, I don't know, like, to cure my cancer, right? And it says something. And I decide that, like, I'm going to follow that and not the doctor, right? What's happening there? Like, what's...

Okay, that's an extreme example, which, of course, I'm accountable for my community. But, like, it's very different because, like, the interaction is very different, right? Yeah. So I think that's the problem that we are going through right now, that we have a new type of agent in the system that we are living in and we need to figure out how to regulate it and how to interact with it, how to deal with failure when we interact. All right. Well, thanks everyone for joining us for another Shop Talk. Costas,

thank you for sharing all of your thoughts, even if they're just a construct. And we will catch you on the next one. You know, Costas, we learned so much from the data leaders that we talked to, but I learned so much from picking your brain. And actually your questions really make me think really hard. So I appreciate Shop Talk. I think it makes me a sharper thinker. Well, it's fun.

Like, I think it's good to just sit and chat about the stuff that we experience. And yeah, I think like, I hope like people enjoy it. That's why I'll keep asking for people to reach out. Please do this. Come up, like you can do that. Like send an email. Yeah. Let us know how you feel and like what are your opinions of like

your experience with the show so please do that so me and derek we can keep being happy of course and of course we try to take the same types of questions to you know data leaders from all sorts of companies, large and small. So definitely subscribe to the main show if you haven't yet. Tons of really good episodes there and tons of really good thoughts from data leaders, you know, really around the world. So definitely subscribe if you haven't. And we'll catch you on the next Shop Talk.