WolframAlpha

A pocket calculator is super with respect to you when it comes to multiplication and division and arithmetic.

And the average Wolfram Alpha running computer, Wolfram Alpha is a mathematics and thinking engine that existed before the LLMs.

To me, it was always, in fact, it was probably just as impressive as when I first saw ChatGPT.

Anyway, Wolfram Alpha was able to do integration and differentiation, lots of sophisticated mathematics in a really simple way.

You can ask it sophisticated mathematical questions and it would spit out the answer with an explanation.

That was like well over a decade ago it could do that kind of thing.

Yeah.

Okay, so I might as well go to, um, Wolfram Alpha.

[laughs]

[laughs] Okay.

Okay.

[laughs]

Okay, now I'll, I'll spawn another, I'll spawn another window and try to get Wolfram Alpha on that one too.

And, and now we'll try the, uh, um, the other guy.

And I'm gonna get onto this in a second.

It is better at math than Wolfram Alpha.

I know that's a big claim, but it's far less buggy.

I found Wolfram Alpha crashing very frequently.

Anyway, here are the two solutions, and isn't that incredible? From a photo, essentially, it then extracts out the math problem, including the two square roots, and then solves it.

This is all within the same window of ChatGPT.

Now, obviously, I need to do far more experiments to see if it actually can recognize any particular images, and maybe I'm putting it down too harshly, but at the moment it does seem to hallucinate if you ask it about too much of the detail of an image.

Next, do you remember how one of the key weaknesses of GPT-4 is that it can't really count things, especially not characters, words, et cetera? And even more so, it can't do division.

You know, Mathematica came out in 1988.

Wolfram Alpha came out in 2009.

These both automated a large part of the sort of pure mechanical aspects of mathematics.

And, um, the, the fact is that people sort of made a fuss at those times saying, "Wait a minute, what do we teach in math anymore?" Because somebody can just go to Wolfram Alpha and get the answer and see a step-by-step computation.

I will say that we, we made a, a conscious decision not to, uh, let Wolfram Alpha kind of randomize the steps it gives so that that's kind of our version of the steganography that people are talking about with modern AI to basically say, you can tell if it was a Wolfram Alpha-generated answer because, you know, any time you type in the same question, you get the same set of steps to, to get the answer.

We could easily have made it so that you wouldn't get the same set of steps.

We decided not to do that.

We have a, an MCP server that's a cloud-based MCP server.

Uh, perhaps not as energetically on some of them, but, um, uh, it was kind of, um, the, the, uh, the path of my life was, was, um, uh, somewhat determined.

Um, and then, uh, in the 1990s, she lived through the, the decade when I was working on New Kind of Science, um, and Wolfram Alpha and, uh, in more recent years, the Physics Project and the things that have come from that.

I have to say that in, uh, I've talked about many things on live streams.

I have tended to not talk about my family and so on, um, uh, in these, in these kinds of settings.

I mean, I would occasionally tell her questions that people had asked on these live streams, which she agreed were quite interesting.

So that's about as close as it got.

I think, I mean, she, what was she? I mean, she was certainly She was involved with the very first kind of live stream that I did, which was the launch of Wolfram Alpha in 2009.

She was certainly there.

Was it CliffNotes? No.

Did y'all use Wolfram Alpha? Do you remember that?

The hell is that? [laughs]

No.

Oh, kids were big on Wolfram Alpha.

It was, this was before AI.

You, it was a website that you could just put any math problem into and it would give you the answer.