Business, Innovation and Managing Life (June 10, 2026)

**Stephen Wolfram** (0:01)
Hello, everyone. Welcome to another episode of Q&A about business innovation and managing life.
Let's see.
See a question here from Cathy. Does using AI help you learn, or does it short circuit the learning process?
It's an interesting question, and I don't think we fully know the answer yet. I mean, if you have been given an exercise, where the AI can just do that exercise, maybe it's the wrong exercise. I mean, this is an issue that came up years ago with calculators, came up again with Mathematica, came up again with Wolfram Alpha. It's the, if the meat of what you're being asked to do is something that has been automated, maybe that isn't the key thing that you should be being asked to do. And the whole point is, once that thing has been automated, there's now more things that you might be able to do. The expectations should be raised. You know, we're about to have this issue because we have our summer programs coming up where we have our summer research institute and our summer research program by high school students. And this is sort of the first year where lots of AI based Wolfram Language programming is really doable.
And many of the projects that have been done in previous years, a version of them could probably be done by just feeding the prompt to an AI and saying do something. So that, as far as I'm concerned, that translates into what I expect should be possible now is something bigger and stronger than what was possible in the past because there's a new tool that's available. Now, in terms of learning things and how AI contributes to that, I have to say for myself, I'm still not necessarily fully developed in my thinking about this, but when I want to learn new subjects or want to learn about something, it's asking an AI to summarize it for me works quite well. And particularly when I tell the AI, this is, I know about this and this and this, summarize this for me, it's quite helpful. And it's more helpful than just reading a generic, this is a summary of that story, which doesn't have that sort of prior information of, oh, I know about whatever. I was recently asking it something maybe yesterday about some particular development in mathematical quantum field theory. And I'm like, I know about traditional quantum field theory and this and this and this. Tell me about this new thing. And I did a pretty good job of explaining the new thing with reference to the things that I really knew about. That's very helpful.
I think that this question again, so for example, what I've been trying to work out is for summer programs for instance, what should our expectations be? So I think one of the key things is that, for example, Wolfram Language is a notation for expressing thoughts computationally. And as such, it's part of the sort of communication channel between humans and computers. It's not something where in sort of traditional programming languages, you're just writing a blob of code that's for the computer and not for the human. With Wolfram Language, you're writing something which is where you're using the language as a notation to communicate thought and a way to sort of crispen up your thought. And so I think one of the key things is if you write something, that's a computational essay, for example, you're expected to understand everything that is the input that you wrote there, which means the English text, the computational language text. Those are things that you're expected to understand, and you should make go to the effort to write them, just as you would go to the effort to write the English text in a way so that it's understandable. So also go to the effort to write the computational language in a way that's understandable, which means not just an AI spewing up blobs of code, nor is it an AI spewing up blobs of text. If you produce something where your computational essay is just a bunch of AI stuff, that's probably not going to be terribly useful. Maybe the AI could say a few things in ways that are clearer than you could say them. So long as you have kind of fully processed them yourself, I suppose it's okay to present that. But there's no point in rewriting the sentence just for the sake of it not being AI written. But if you don't even understand the sentence, then that's absolutely useless. I mean, I've seen that a bunch with people recently, including some people I've worked with, where they're producing AI stuff that I don't think they understand, and this is not very helpful. I mean, it's very easy to get computers to produce things that one doesn't understand.