It’s kind of funny to me. A lot of people will say things like, what is the use case for AI? Like, it seems very interesting, but like, what is it for? In my case, I’m sort of unusual in that I’m not interested in business use cases. Most people interested in business use cases, and people will tell you all day about all the business use cases that they can think of, right? But I’m really not interested in that. And this person saying like, okay, well, what about like, what else? I’m struggling to think of reward applications.
The reason that AI was originally created was to help humanity understand complex, chaotic systems like social systems, economic systems, ecological systems, climate, all these other things that are very difficult to predict using linear regression techniques. This is why the Soviets invented AI. Today, we mostly use it to manipulate people into buying things, but it’s still really good at all that other stuff too. And I think that other stuff is the stuff that is really best suited at. So for example, I built an AI completely from scratch with a team last semester, which is able to accurately forecast and predict how every country in the world will do with Covid based on 250 measures of the sustainability of that country, right? And it not only knows exactly how the country will do, it also is able to predict what you, you can run it backwards and see which factors are most important, and it’s things that you might not expect. Like for example, what percentage of the country’s GDP is from manufacturing? This is very important. Why? I don’t know. It just is. And it accounts for 11% of the accuracy of the model. So like, that’s weird. Why is that? Right? There are all kinds of things like that that we can learn from AI that we couldn’t learn without AI. So another really good example is the chat interface, which I’ve always thought is dumb. Okay, I’ve Learned to enjoy it in certain aspects.
So like one thing is it’s really good at helping people to deeply understand complex topics through conversation. If you wanna talk to it about something really complicated that you want to try to understand and start asking it questions. It will explain in detail based on really clearly understanding exactly where you’re at with the concept and what you’re missing and what you need to understand and how it can say in a different way, right? Like it’s really good at that. The best tutor that you will ever find, right? Like a good large language model like GPT four or tbrx or Mix Stroller, one of these very big, very good ones. It’s also really good at brainstorming if you have kind of like the rough outline of an idea, it’s really good at helping you flush it out, right? It’s really good at giving you 20 examples if you want, like, I know this kind of idea, give me like a list of where you could go with that. It’s like, sure, here you go. These are the kinds of things like if you had the smartest person in the world on a really specific topic and you wanted to ask them a bunch of questions, right? Like it’s sort of that and it’s not gonna tell you like the exact population of Ethiopia today, right? Like that’s not what it’s for. That’s not what it’s good at. What it is good at is understanding the relationships between concepts and the way that you may or may not understand things like that, right? And it’s really good at making plans and then updating them. So like, for example, a lot of people will just take their homework assignment, put it into ChatGPT and say, do my homework, and it’ll do it. It’s not very good, right? If instead a person is, think of it this way, if you’re using a calculator to do math that’s very complicated, you’re not gonna get the right answer. If you know how to do math that’s really complicated and you use a calculator, you will get better answers. It’s like that. So don’t think of it as the thing that does your homework for you. Think of it as the thing that helps you do your homework, right? So if you need to create an outline with like kind of a few big topics, right? How are they related? Making that outline is like 90% of the cognitive effort. It can do that for you easily. And you can read through it, change things. You’ve got your outline. It’s really good at giving you a first draft of something and then updating it however you want. It’s not gonna write an amazing essay for you, right? It’s not gonna do your homework really well for you, but it can give you the first draft that will evolve into a really good essay or a really good, whatever, Powerpoint, anything.
And so this is the thing. I think people oftentimes today assume it’s either going to do everything for you or it can’t do anything for you. But the reality is it’s really good at helping you do the things you’re already doing with a little bit less, like cognitive exhaustion. And I think that’s sort of the big use case that I really like is if you kind of have an idea of what you wanna do and you just know it’s gonna take like a lot of effort and like cognitive energy for you to accomplish. It can help you get started. It can help you overcome rider’s block, overcome hurdles. It is in many ways just a really good auto complete, like a really good auto complete. And sometimes finding the next word is really hard. And so in those situations, or if it’s finding the next topic or how do you expand the concept or idea that you’re working on, that’s where it’s really good. And you know, a lot of these things also can apply to business, but I think like it’s more interesting not apply having a really great tutor on any topic. You know, there’s so much, it’s a whole world in there. What are the things can people think of?