Why intent prediction needs more than an LLM

**Ryan Donovan** (0:00)
If solving difficult problems is what drives you, Raytheon is hiring software, systems and hardware engineers to take on some of today's most complex technical challenges. Explore opportunities at careers.rtx.com/raytheon.
Hello, and welcome to The Stack Overflow Podcast, a place to talk all things software and technology. I'm Ryan Donovan, your host, and today we are talking about how LLMs fail on things that aren't language, things like intent prediction. And my guest for that is Frank Portman, who is the CTO at Yobi. So before we get into the topic of the day, we like to get to know our guests. How did you get into software and technology?

**Frank Portman** (0:47)
It's a good question. So I have been doing this for a little while now, but actually way back in the day, I wanted to be a mathematician. And I think I have a very similar origin story to many people working in technology, which is that when I was very young, I liked computers, I liked video games, I liked to tinker. But I definitely wasn't one of those people that was coding something at 13
I knew how to open the terminal on my Windows computer, which was advanced, but not too crazy.
But I was like, that's cool, technology is cool, but I like math. So I actually went to college for pure math, which I did and I finished, and I still find it super interesting. But along the way, we had to take a couple of courses, like outside of just our direct major, which I think is pretty common. So I was like, okay, let's do applied math. I'm not doing pure math.

**Ryan Donovan** (1:37)
Right, right.

**Frank Portman** (1:38)
Specifically, let's do the ones where we have to implement our own matrix multiplication and all that stuff in C.
And suddenly, I wasn't the one that was asking people for help. And I was like, maybe there's something here, because it's always fun to kind of do what you're good at. And I think it was just super fun for me to, you know, I think of writing software and solving problems with software is just like a series of mini puzzles. By the way, that's not the software engineering job. I think that's like the coding that you do. Software engineering job is all about like solving business problems, which is we can get into that. But I founded Challenging and Rewarding. And this was right around the time data science and machine learning was becoming like a very hot buzzword. Okay, despite no formal training or education here, I do have the math and stats backgrounds. I seem to be good at coding thing. Let me try.
Into that was tough. Like nothing's harder than getting your first job, I think. Ever since then, that's what I've been doing.

**Ryan Donovan** (2:37)
Yeah, obviously, matrix multiplication is pretty hot these days, isn't it?

**Frank Portman** (2:41)
It is pretty hot these days. That's what everything comes down to these days.

**Ryan Donovan** (2:47)
So, I've heard of folks using LLMs for a sort of generic, abstracted next token prediction where the next token could be intent signals, could be steps in a process, could be, you know, behaviors in a chain. Why do you think that is a lesser approach or a flawed approach?

**Frank Portman** (3:07)
There's a couple of things, and I, you know, I definitely welcome you to challenge me on any of these points, because I only see what we do and also I see kind of what the industry does at a high level. So, the nice thing about models is they'll do exactly what you train them to do.
It's also the worst thing about models. And it's not clear to me why the inductive bias of let's gather all the text in the world and maybe all the video in the world, get very good at predicting sort of sometimes very long sequences, to be fair, and then maybe fine-tuning it a bit so that it's more pleasant or more correct in conversation, which is subjective. It's not clear to me that that inductive bias would make these LLMs good for like forecasting or more prediction or decision making.
But they are phenomenal at synthesizing information within your context. They're phenomenal at like writing code. They can do all sorts of emergent things such as writing, rap in the form of Shakespeare. But decision making under uncertainty is sort of like the thing around modeling intent and making predictions. And it's not clear to me that the inductive bias of like, let's just try to predict next token can just build that into existence. However, I do think those technologies such as LLMs, in combination with the right tools, you could start to get there. And I think that's what people are pretty excited about, agentic everything versus the one LLMs rule them all, if that makes sense.