The $3 Trillion Labor Question: Will AI Really Replace Us?

Speaker A: This is Ryan Staley and I am back with Mike Palmer, the CEO of Sigma Computing, for part two, where we are on the future of AI and data. For those of you that missed episode one, go back and watch it. Mike crushed it. We talked about a lot of different areas about transformation, what's kind of happening, um, the integrity of data accountability. With AI. We crushed a lot of different areas. So, Michael, welcome. Happy to have you back, man.

Speaker B: Glad to be here again.

Speaker A: Yeah. So, all right, so we did talk a little bit about data on the last episode and specifically as it relates to accuracy and AI and what's kind of happening as things are progressing forward. So I guess, like you've been through multiple transformations before, right? You've been through mobile, you've been through cloud, I guess like as a jumping off point. And we didn't even cover this in episode one. But like, do you think this is the same or do you think that this is different as it relates to what we talked about with pattern matching in episode one?

Speaker B: For me, it's 100% the same. And I go back all the way, uh, to the.com era for those of us that were around for this wonderful time living, uh, in San Francisco. And everything was dot com, right? Uh, I'd have to almost, for most of your listeners don't even remember this, but everything that was a valid business all of a sudden became a dot com business. And everyone was going to sell everything on the Internet. That was it. And if you didn't do it in the next six months, you were totally going to be gone and irrelevant and out of business. And as it turns out, like, the vast majority of the companies that did that were the ones going out of business and then the others sort of continued. On the other hand, some amazing companies came out of that time and those companies got it right. And obviously the most notable of these was Amazon. And Amazon, what they actually got right wasn't selling stuff. What they got right was data. They got the fact that they could understand who you were and what the pattern of your buying was, and they use that to attract merchants. And they. More merchants created more product, and more product created more users, and more users created better pricing. And we all know the flywheel that they, uh, that they've espoused since then. So I think the pattern is going to be the same for us. It's a bunch of companies, uh, that, as I referred to at one point, is a bunch of chancellors who think that we've got this AI thing and they're going to somehow create a great company in three months. And you know, they don't do the work, right? They don't do the work to really understand what customers need to do their jobs better, to make them more productive, to raise revenue, to save costs, to do that compliantly, securely, to do it at scale over a long period of time. And so those companies, just like back in the dot com area, they're going to get washed out and there's going to be another set of companies that don't respond to AI at all. And they're also going to be in big trouble. Just like the, you know, the shopping mall equivalent of, of from the dot com era. Um, and then in between there are going to be a bunch of great companies, ones that really do have the cultural DNA to build great products and identify great technology and its role in building great products and they're going to do the work. And uh, I hope we're one of those companies. So I feel like when we come out of uh, this, this wave, it's going to take a couple of years because there were so many dollars invested in AI, uh, companies and continue to be, we're going to see significant productivity gains. We know this, right? Because there are manual areas of all of our jobs that clearly are great opportunities to be reduced. There are going to be things that we don't do today. And I think this is one of the areas that I'm most interested in where AI is not replacing anything. It's enabling us to do something we just didn't do before. In the last episode we talked about these data applications that customers are writing on Sigma. The vast majority of them are not replacing an app, they're replacing a habit. And no one could replace that habit. They didn't even know they had it. They were just doing stuff every day. Oh, every day I modify this, I upload it, I send it to this other person, they send it back and we just do it. This is what we do. We're AI is going to replace that, right? We are going to build AI based applications that create efficiencies in our jobs and then we're going to take that time and we're going to do much better things with it. And we don't know what those things are. You know, the last. I'll leave you with this example that we use all the time on AI. And by the way, we also don't Even know which AI we're going to use. If we compare the AI that's available today to 18 months ago, we had the rise and fall of so many models along the way, and we will continue to see it. Adoption and evolution are something that we're going to have to be accountable for here. How do we use the best tech at the best time for it? But let's talk about a mobile. You know, I was telling everybody I had a mobile phone in, like, the mid-1990s, and it was really cool because I did something I could never do before, and that was take a phone call when I wasn't at home. Was amazing. And then fast forward to today. I m have, uh, two daughters and one is 15. And when I call somebody, she's like, that's so rude. Why don't you just text them? So I think it was like the thing I bought this thing for. The revolution of making a phone call I wasn't at home is like, the rudest thing that I can do 25 years later or 30 years later. It's nuts. But what. What I do with my phone is something that was never imagined in the mid-1990s. This is where we are with AI. AI is being targeted at some early things. Just like a phone call 10 years from now, you know, it's going to be doing something radically different. We just don't know what it is. So every day we're all going to show up and adopt and adapt. Uh, and I think we're going to all participate in identifying what that future of AI is. So the pattern tells us that we have to play in the game, we have to invest to change, and that we have to be prepared to do that over the long term. Uh-huh. Yeah.

Speaker A: Yeah. So a lot of things unpacked there that you said. So I totally agree with you that what was available 18 months ago versus now is wildly different. Right. Um, even across the model landscape. I mean, what Google was putting out there before was a joke. And I think Gemini is exponentially better, I think, in my opinion, as of right now, as of this recording. Right. Who knows when you're listening to this, but it's like, right neck and neck with what OpenAI and ChatGPT is doing right now. So kudos to them for kind of getting on path with that. But I want to agree 100% with you. The only thing that I question, and I love your feedback on this, man. This is healthy.

Speaker B: Uh, I love a good debate.

Speaker A: Healthy feedback is like, what do you think about the attack on the labor force with basically all the VCs, uh, EBAT, but mostly VCs, right. Really targeting that. I think it was $3 trillion or $5 trillion in labor costs, um, with agents. And by the way, I don't, as a disclaimer, don't think agents are there yet. But that combined with Moore's Law, like I think the way AI is doubling, it's like every six months instead of 18 months. So that's a little bit of a change. So what's your take on kind of those two elements with like the future of work and what's happening with AI and how it might be a little bit different than some of those other examples you mentioned?

Speaker B: Great question by the way. I hope they're right. You know, so that's my first statement. Um, having said that, let's again, pattern match. How many times just in the last 200 years has some form of technology been predicted to uh, make people's work obsolete in such a way that they wouldn't have jobs? This has happened over and again. Everything was, every domestic product was going to eliminate work at home. Machines and industry in general was going to eliminate the need for people. And yet we have some of the most, the lowest unemployment rates we've ever had.

Speaker A: Mhm.

Speaker B: And this has been consistently true. Why? Because humans are creative and they will always find sort of the next thing. And what people are just uncomfortable with at times is unless they can know what that next thing is, they predict a doom and gloom.

Speaker A: So.

Speaker B: So number one, I hope technology replaces crappy work with something that's better. Secondly is, by the way, we need it. If you look at birth rates in advanced economies, we need to become more productive. So this is a good thing for all of us. We wouldn't put our back with a thresher and a wheat field, you know, from 200 years ago and hopefully 50 years from now, we're going to look back and think like, why do we have to do all that, you know, and whatever that that is that we've completely replaced with something better. So I think that we need this productivity change. I hope that it's right and I hope that we get comfortable with the idea that the thing we're going to do differently in the future is unknown to us right now. But it's there, it's going to be there. The problem is like anything else. What's the timeline? This is anybody's guess. And I come back to my earlier point. I am not a believer in revolution. I am a believer in evolution. So to the extent that all of us are participating in our own Adam Smith like ways in the great economy called AI and we're learning things. The aggregate experience is going to create something worthwhile. Whether it's in data or whether it's in replacing legal services or it's outbound sales or it's robotics or self driving cars or whatever it is. All of this is going. There are going to be advancements that apply more medicine, they're going to be advancements that apply in hop categories and then we're going to have I think these bursts of really interesting change and then maybe flatline a little bit. Can't predict it. Our job I still believe is recognize its importance, work really hard to figure out how to apply that technology to something meaningful. Learn and then adapt.

Speaker A: Yeah. Yep. Well I love your answer and I think it's uh, very well grounded and the adaptability is huge. Kind of like what you're talking about of, I mean even over the last year. So um, let's shift gears a little bit and I know we're almost up on time so I want to be sensitive to that. Let's talk about the future of data and AI, right? Like something that you specialize in focusing on. What's your take on where that's heading? Because we have hallucination rates, um, we have amazing capabilities that we've never been able to diagnose such large chunks of unstructured data. However, you know, like basically looking at through the lens of what you do and making sure your numbers are laser tight, um, with where do you think that's going and what's the timelines for that at least from your perspective.

Speaker B: So here's another example where I think it's uh, it's an all of us working together even if it's without knowing it. Sigma for example takes advantage of the great technologies that are being built by model providers and all of the model providers, not just one. You mentioned Google's advancements with Gemini. You know I think we're going to have this like uh, I always refer to these as the horse race at the amusement park that you shoot the water at and like one horse pulls ahead. This is great. You know we want this sort of competition and some of them are, will specialize and they'll maintain leads in one area and some of them be generic and they'll go back and forth. For Sigma, one of the uh, things that we focus on is where do we add net new information to that model that raises the performance for our customers. So for example, if we're going to use a model to answer a data oriented question, what we can add to that is who you are, what you tend to do on a frequent basis, who you work with, what people like you tend to ask at your business. So we have contextual information to add to, model based information that hopefully closes some of that gap between the 85% accuracy rate of a generic model and the 99.9% accuracy rate that you probably need to do your job well. And this is where again, in an ecosystem basis, everybody wins. The model wins, the end user wins. For productivity reasons, Sigma wins because we know where our value add is in that we're not going to go write a model and replace Gemini. So I think that's the path that we're going to be on is that how do we complement each other over time for the benefit of the end user?

Speaker A: It's good, ma'. Am. All right, last but not least, as an executive $100 million plus company, what's your favorite use cases for AI?

Speaker B: Personally, man, uh, this is literally an impossible question to answer. We have customers that are building applications with AI, which I think anything like that is amazing, right? If you think about how difficult it has been to build applications for the last 30 years and how many failures that have been, how costly it is. So I think that is totally like remarkable. I use AI every day to, to augment something I'm looking at in data. You know, I can ask a why question and get a why answer from, you know, publicly available transcripts that get summarized all the way through something that's sitting in my document repository at my company that I think is amazing. But for me, the answer still has to come back to apps. You know, I think one, one app. I can't say the name of the company because I've never spoken to them, but we can see a lot of this in metadata. They've built an ERP system in Sigma for managing poultry and specifically the ability to manage breeding of chickens. And so like they have transportation and weather and they've got, uh, stuff about chickens, you know, like. But who's going to write that? You know, you're, you're not. SAP is not going to like, you know, we're going to go after the poultry market, but there are a lot of polterers out there that really need software to be more efficient. And the fact that we're opening up the market and lowering the barrier to build those apps to words is pretty remarkable. So I just think that for me, AI is unlocking a below the water line part of the iceberg that is out there for small and medium businesses or individuals and departments at large businesses to discover new ways to be productive, you know, and to, to be empowered to do that themselves and not wait for like the mothership central thing to kind of help them do their job for them. That, for me, is super exciting because language, in the end, is the lowest barrier to entry to technology.

Speaker A: Yeah. Well, that was fantastic, Michael. We are up on time for part two. Where can people find you? Where can they find more About Sigma Computing?

Speaker B: Sigma Computing.com is the best way, best, uh, place, best place to, uh, to to learn about Sigma. I would tell everybody, if you're watching this during the summer and you're coming to summit season, uh, here in San Francisco, stop by our booth. We also do these events, um, around the country throughout the course of the year and we look forward to, you know, talking a little bit more about this.

Speaker A: Yeah, appreciate you being on, man. This is a blast. So, yeah, if you are in going to the summits, definitely talk with Mike. He's a blast to have and I can attest for that from this episode, episode one, and even the pre show. So thanks for being on, Mike. Really appreciate you being on and, uh, we will see you all on the next episode.