The Hidden Value of Numbers: Generative AI for Tabular and Time Series Data | The Six Five Summit

Hey everyone. Welcome back to the six five Summit. Daniel Newman, here, CEO of the Futurum Group, excited for this next session here in our data DevOps observability track.

We got Dev ever Shaw. He is the CEO of Ikigai. We're gonna be talking about AI of course, because that's what the world wants to hear about.

But Dev Everett, wanna thank you so much for taking some time to join. I know you're busy building, starting up, making things happen. I love it as an entrepreneur, one to another.

How's it going today? Excellent, Daniel, thank you for having me here. O of course.

It's great to have you here. So listen, you know, this event has companies like, uh, Amazon and Microsoft and Google, and we also have exciting startups like Iggy Guy. Gimme the quick one minute, you know, for everyone out there that isn't familiar with the really interesting and exciting work that you're doing.

Absolutely. So, uh, I'm CEO founder of Iki Kai, also professor of ai at MIT, where I've been teaching for 20 years. Uh, at Iki Kai, we bring generative AI for tabular time, cities, data, structured data.

Uh, our mission is to enable everybody to thrive in the AI era, right? With the view rather than get drowned. And our focus is enterprises.

Well, listen, I first of all, nice resume. I mean, look, one of the most esteemed, uh, universities, schools on the planet, uh, one that, uh, someday if I work really hard, maybe I'll have the chance to, uh, come and, uh, you know, stop by and say hi to your students. I, I, I love it.

But, uh, listen, um, LLMs are all the rage. Everybody's talking about LLMs, LLMs, LLMs. We're going from, you know, we've got small models, we've got bigger models, we've got gigantic models, you've got trillions, billions, millions of parameters.

Um, but it's not the only AI that's out there. You know, LLMs are good for some ai. Talk a little bit about just your general viewpoint on kinda what LLMs and AI are, and then what are kind of the rest of the four decades of AI capabilities that we've been building, and and when are those approaches more appropriate?

Absolutely. Uh, so LLMs, uh, diffusion for model for images, multimodal models, these are great models for unstructured data. See the world in which, uh, unstructured data is four decades, we've been spending a lot of time trying to learn what are the life mathematical representations for those things, and we could not figure out, and this modern take is let's just take all worlds data and find, uh, just represent it in a sense, in a succinct manner so that when a new data example comes in, let's look up effectively what is the right matching answer to that and producing answer.

That's what these things do. On the other end, when you think about tabular times, Sydney's data, four plus decades of, uh, research, actually I would say a hundred years plus of research has provided us methods, uh, approaches, models, representation, so that when those data points come in, we understand the structure within it so quickly and we can extrapolate from it. And if you think about enterprises data, core enterprise data, whether it's about people, whether it's about product, whether it's about finances, whether it's about sales customers and all that, it's primarily ular time sitting data.

And those are the things where we don't need large language models or diffusion models. We need a very good mathematical presentations that can allow you to extract patterns, insights from it and make decisions with it. Do scenario analysis and all that nice sink.

Yeah. So there's a big opportunity in, and in businesses actually, you're kind of hearing the, the, the tide turn a bit on conversations. We all know that the LLMs are excited also becoming a bit of table stakes.

Of course, some, you know, some definite nuance between company to company on the training and the quality and the outputs, but at the same time, it's become largely available to everyone for almost nothing. 0, it's the next generation of search. It's a new format, but essentially it's, it's Google's next thing is like you search and now you get generative responses, but still in many ways search the value.

The whole market talks about the value of a is when a company can actually use proprietary unique data that sits inside of one of their various systems of record or that sits on the, you know, devices of people who have built out endless PowerPoints and, and, and, and, uh, Excel spreadsheets and have written millions of emails and done presentations with graphics and representations. That data is where the, that's the money data as I like to call it. That's where companies differentiate.

You've built a novel, novel approach. You're trying to help companies handle this and, and some of this and some of what I mentioned, but those datas the things that you talked about. The large graphical model.

Talk about, just gimme the quick distinction from A LGM. We're gonna make new names today and an LLL. Absolutely.

Uh, very nicely put by the way. Um, in terms of lgs, uh, what we do, we think of data as a tabular timestamp data. Think of, uh, as a mental model, think of a giant spreadsheet that's changing over time.

Okay? For example, the demand of your product in terms of what sales you have made in a given channel, given location for a given product. That's changing over time, time series, uh, kept in a nice structured manner for this type of data.

What you want to do is you want to learn the relationship between every pair of cells, okay? So that if there is a cell that is a value that is missing, you can sell it up. There's a cell that is value, but which is unusual.

You can detect anomalies. There's a part of data that was having some type of model, and now post covid model has changed. You wanna understand change points.

You wanna understand what are the relationship, what are the similarities between multiple behavioral things like that product in that region behaves like this product in that region, or that product is 40% this and 60% that. Those kind of embeddings, uh, those kind of, uh, what I would call calculus of these structured time series data is what large graphical model enables in a domain agnostic manner. Large graphical models are primarily based on mathematical representations.

And for that reason, as you present, small amount of data tries to look for the right representation for it so that you can do all sorts of extrapolation very quickly in a very efficien manner. These models are primarily, uh, representation. They're not like, sort of learn from world's data for that reason.

They're very efficient, computationally efficient, they scale well. We run on, we are proud that we run on CPUs, not GPUs. Um, and once this is at the core, you can do all sorts of interesting things, whether harmonizing data together when it's in multiple data environments, simulating, forecasting, predicting to doing scenario analysis on top.

So all of that you can do with this. Okay. Okay.

So I had a, uh, I had a question, but now I have a different question because you just opened up the, you know, I'm a, I'm a, you know, CHIPS are cool again, by the way, you know, remember when semiconductors were not cool? Um, they are cool. I mean, you know, you just said something.

And by the way, I've said this on the record many times. I get some lashback from the Nvidia fan boys, and I, I, I'm a big fan of Jensen. I I've gotten to know him.

Incredible, incredible leader, but a lot of AI is still done on CPUs. Talk to me just really quickly, 'cause I don't want to go too far down this rabbit hole, but why did you choose that is what I think the world would love to hear, why a, a company like, like yours has gone down the path of using A CPU as opposed to A GPU. Um, so at the end of the day, um, we are using CPUs can because we can, and that's sufficient, Okay?

And the reason that's sufficient is because we are not taking what I would call, uh, an approach where go through entire world's data to learn the patterns. We are just focusing on your data and to identify patterns. We are using, uh, the universal mathematical representation that we have identified over years of d years of research through my research at MIT and our collaborations.

No, that ma that makes a ton of sense. I mean, it's okay sometimes to just, uh, be like, look, it does the job and it's way more cost efficient right now. Way more power efficient right now.

Um, there's way more access. You're not gonna wait nine months or 18 months to get access to them, or you're not gonna pay a huge tax in the cloud to get access to. And it does, it doesn't work.

A lot of inference is done on A CPU. I mean, a ton of inference and a ton of inference will be done. We're not just gonna shut all these data centers down that we've been building over the years, we're gonna put 'em to use.

It's kinda like older storage technologies became cold storage and then we'd warmer and then we'd HBM and, and we're seeing this happen. So talk a little bit, just gimme a couple of the like really clear use cases for lgn, if you could just put 'em in context for people. Absolutely.

So, um, back to the and use of AI. In enterprises. Enterprises, there are two things that they wanna do really well with data and ai.

One, figuring out opportunity to grow. Two, run their operations efficiently. In either case, it's about form of balancing act.

Think of supply chain, supply and demand. Those are the two sides of equations. As you think about demand, you wanna forecast demand really well because that's uncertain.

As you plan for supply, you wanna figure out what are the right planning and controlling actions that you can take among choices you have available. Uh, if you go to finance, you're trying to figure out how do you reconcile your data that sits into multiple environments so that we can unify it. You can do a bunch of things on top of it, whether it's after that, forecasting your cash, forecasting your consumption.

If you're a consumption based company, uh, or planning and doing scenario analysis from an outcome driven perspective, uh, if you are, uh, thinking, uh, about, uh, insurances and insurances, you're to do claims auditing. How do you use AI to do that? Uh, if you're thinking about, uh, financial services, you want to, uh, do a fraud detection, fraud management risk score that is comprehensive across customers and transaction across all sorts of data altogether.

So these are the type of use cases that we have been seeing across supply chain banking and financial services, insurances, and, uh, the traditional CFO's office. Excellent. And by the way, you know, you, it sounds like you're hitting the highly regulated industries and also, uh, sounds like you can, you can help with some of what I'd say persona based role-based, uh, which is super important.

I mean, just ask our CFO of the questions that I'm asking him and how much AI could help if, uh, if it, if it was properly implemented and, and we're getting there. Every company's getting there. It's a, it's a process.

So I got about a couple minutes left with you here, and I really appreciate you spending the time. I'm gonna ask you to put your professor hat on for a minute here in the last, yeah. I'm gonna ask you kind of a big question.

Hopefully you can answer it in a, in a somewhat condensed manner, but key considerations for the development of a longtime AI strategy. And I'm gonna couple this considerations for businesses developing a strategy, and at the same time, how does, how do you see policymakers and, and business leaders making sure that we manage it? So how do we deploy and then how do we manage it responsibly?

Great. So there's sort a few things here. Uh, one is as we use ai, uh, whether we use, uh, internally or externally with ai, exception is a norm.

So plan for it. Rather than thinking of that as a surprise, if going back to compliance, uh, heavy compliance slash recognition, heavy industry, you want to keep expert in the loop, not out of the loop. So the best way to think in my mind, bringing AI is to have experts within the loop, not out of the loop.

Let machines do a lot of work for you, but make sure that there is a spoonful of int intelligence that is brought in by experts constantly. So that's one part. Second is data is where AI stacks and data is where AI ends.

So be very mindful of how do you utilize data, how do you mix it up or how do you avoid mixing data? It's very, very important to have attribution, uh, in terms of where you are generating AI insights. Okay.

And then finally, the third thing is, uh, unlike other things, AI is rapidly progressing with regulation. The, it's the norms that matter and those norms are not yet defined. So what you need to do is to, you need to start defining those norms along with your, uh, other fellow community industries and try to self start to sort of get into the self-regulation things.

For example, even though we are a small startup, we decided to, uh, follow this ourselves. And we have set up our own AI council. This AI council is a collection of, uh, awesome academics both from, uh, uh, social sciences, the law and policy and ai.

And they help us meet a quarterly and they help us provide what are the things we should do or we should not do, and then implement. Yeah, listen, you did a good job and you played the professor role perfectly there. Dere.

I would love to sit in on one of your lectures at some point. Thank you. This is a big problem.

I remember just recently, I was on CNBC and they were asking me about, you know, basically, can this be regulated? And I think it's gonna be people like you building companies like you, working with policymakers, business leaders, to figure out a, how do we do this technology to keep innovating and lead the world, but at the same time, do it safely, do it securely, do it responsibly. It will be a tug of war.

Industry will always move faster than regulators. There's really nothing we can do about that except work closely, stay informed, you know, try to open the box, be transparent whenever possible. And it sounds like this is exactly what you're doing.

Deborah Shaw, CEO Ikigai, thank you so much for joining me here at this year's six five Summit. Let's talk again soon. Thank you Daniel.

Thank you For having Alright everybody, we appreciate you tuning to this segment. It was a fascinating one, lgs, we've got a new acronym now. Check out Ikigai and stay with us for all of our six five coverage studio.

Sending it back to you.