11 ms·

> In the arc of intellectual history it defines a broad, new, computational paradigm for formalizing the world. I am in awe of what Wolfram achieved with Mathe

> In the arc of intellectual history it defines a broad, new, computational paradigm for formalizing the world. I am in awe of what Wolfram achieved with Mathematica. And also the size of their ego.

castles 3y ago
Are they wrong :) ?
throwup238 3y ago
I was honestly expecting Wolfram himself to have claimed the invention of LLMs or transformers, probably by saying it’s really a scaled up implementation of some other function in Mathematica.
- seanhunter 3y ago
  Joking aside, he sort of implies this in his essay explaining GPT models[1] At some level this reminds one of the idea of universal computation (and my Principle of Computational Equivalence), but, as I’ll discuss later... And his "Principle of Computational Equivalence" is [2] "There are various ways to state the Principle of Computational Equivalence, but probably the most general is just to say that almost all processes which are not obviously simple can be viewed as computations of equivalent sophistication." Which a cynic might say is mighty convenient, because this is non-specific enough that you can apply it to basically anything and say you invented it and/or it's equivalent to something you invented and if it doesn't quite fit, you can use the "various ways to state" clause to weasel-word your way into something which does. I find Stephen Wolfram frustrating for this reason. Benoit Mandelbrot is another guy who constantly seems to claim he invented everything, eg the Efficient Markets Hypothesis (which Mandelbrot basically claims he invented because he gave Eugene Fama some advice about the price process for stocks when Fama was a PhD student even though Fama/French was after that and Mandelbrot's idea of the price process for stocks is very obviously and demonstrably wrong if you know about market microstructure and/or look at trade marketdata[3]). [1] https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-doing-and-why-does-it-work/ https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-... [2] https://www.wolframscience.com/nks/p716--outline-of-the-principle/ https://www.wolframscience.com/nks/p716--outline-of-the-prin... [3] At a microstructure level price movements are made up of individual trades which jump around wildly with gaps in both the time and price dimension, and it has a base level beyond which you can't "zoom in" any more- it isn't some kind of fractal scaling in time and/or volatility which is what Mandelbrot wants it to be.
  - throwup238 3y ago
    And the penny drops! Thanks, I didn’t know he already did it. Totally on brand though.
    - defrost 3y ago
      Every mention of Stephen Wolfram deserves a link to Cosma Shalizi's epic take-down of a review ("A Rare Blend of Monster Raving Egomania and Utter Batshit Insanity"): http://bactra.org/reviews/wolfram/ http://bactra.org/reviews/wolfram/
      - seanhunter 3y ago
        What an amazing review-Thanks for that. It really brings together a lot of sad threads in my mind from that time. Having been the kind of guy who waited patiently for hours for home-made CAs to wiggle about on my VGA monitor and/or wrote little artificial life simulations etc for years in lonely isolation I bought "A New Kind of Science" with great expectation when it was in huge piles in every bookshop and was kind of devastated by how empty it was. I had really hoped for so much.
  - chaxor 3y ago
    This is absurd and cannot be true. Shmidhuber beat both of these guys easily with his paper from decades before these amateurs even claimed to start working on it.
    - seanhunter 3y ago
      And tehy didn't cite him. It's a travesty.
pjmlp 3y ago
So what if he likes to put things that way, there is hardly any competition to Mathematica, in what it is capable delivering. Yet many of those that complain about Wolfram, will idolatrate Steve Jobs, and they aren't (weren't) that different in praising themselves.

> In the arc of intellectual history it defines a broad, new, computational paradigm for formalizing the world. I am in awe of what Wolfram achieved with Mathematica. And also the size of their ego.

castles 3y ago

Are they wrong :) ?

throwup238 3y ago

I was honestly expecting Wolfram himself to have claimed the invention of LLMs or transformers, probably by saying it’s really a scaled up implementation of some other function in Mathematica.

seanhunter 3y ago
Joking aside, he sort of implies this in his essay explaining GPT models[1] At some level this reminds one of the idea of universal computation (and my Principle of Computational Equivalence), but, as I’ll discuss later... And his "Principle of Computational Equivalence" is [2] "There are various ways to state the Principle of Computational Equivalence, but probably the most general is just to say that almost all processes which are not obviously simple can be viewed as computations of equivalent sophistication." Which a cynic might say is mighty convenient, because this is non-specific enough that you can apply it to basically anything and say you invented it and/or it's equivalent to something you invented and if it doesn't quite fit, you can use the "various ways to state" clause to weasel-word your way into something which does. I find Stephen Wolfram frustrating for this reason. Benoit Mandelbrot is another guy who constantly seems to claim he invented everything, eg the Efficient Markets Hypothesis (which Mandelbrot basically claims he invented because he gave Eugene Fama some advice about the price process for stocks when Fama was a PhD student even though Fama/French was after that and Mandelbrot's idea of the price process for stocks is very obviously and demonstrably wrong if you know about market microstructure and/or look at trade marketdata[3]). [1] https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-doing-and-why-does-it-work/ https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-... [2] https://www.wolframscience.com/nks/p716--outline-of-the-principle/ https://www.wolframscience.com/nks/p716--outline-of-the-prin... [3] At a microstructure level price movements are made up of individual trades which jump around wildly with gaps in both the time and price dimension, and it has a base level beyond which you can't "zoom in" any more- it isn't some kind of fractal scaling in time and/or volatility which is what Mandelbrot wants it to be.
- throwup238 3y ago
  And the penny drops! Thanks, I didn’t know he already did it. Totally on brand though.
  - defrost 3y ago
    Every mention of Stephen Wolfram deserves a link to Cosma Shalizi's epic take-down of a review ("A Rare Blend of Monster Raving Egomania and Utter Batshit Insanity"): http://bactra.org/reviews/wolfram/ http://bactra.org/reviews/wolfram/
    - seanhunter 3y ago
      What an amazing review-Thanks for that. It really brings together a lot of sad threads in my mind from that time. Having been the kind of guy who waited patiently for hours for home-made CAs to wiggle about on my VGA monitor and/or wrote little artificial life simulations etc for years in lonely isolation I bought "A New Kind of Science" with great expectation when it was in huge piles in every bookshop and was kind of devastated by how empty it was. I had really hoped for so much.
- chaxor 3y ago
  This is absurd and cannot be true. Shmidhuber beat both of these guys easily with his paper from decades before these amateurs even claimed to start working on it.
  - seanhunter 3y ago
    And tehy didn't cite him. It's a travesty.

pjmlp 3y ago

So what if he likes to put things that way, there is hardly any competition to Mathematica, in what it is capable delivering. Yet many of those that complain about Wolfram, will idolatrate Steve Jobs, and they aren't (weren't) that different in praising themselves.