Math bg, tinkerer, AI enthusiast, enjoyer and occasional creator of art and music.

  • 3 Posts
  • 5 Comments
Joined 1 year ago
cake
Cake day: June 10th, 2023

help-circle





  • Both are fascinating to me. I have also met a lot of people who are interested in both. I don’t think it’s a coincidence. There’s something really interesting about the way we can use symbols and signifiers to encode and transmit and preserve information. Any kind of information.

    Coding requires you to say precisely what you mean. Give clear instructions. Define exactly the pieces you are working with. There is really no room for ambiguity, and there’s something really satisfying to the logical side of my brain about that kind of rigidity.

    But that’s exactly why linguistics is interesting to the other side of my brain. Human language is full of fuzzy categories, changing definitions, unwritten rules, unspoken connotations, creative repurposing, borrowing, taboos… You can add dimensions of meaning with text, your voice, your eyes, the movement of your body. You can pack so much nuance into a single word or phrase; a subtle hint can mean so much more than what you are literally saying. You can intentionally encode a message so that it is NOT understood. There is infinite malleability in human language.

    This is why it’s so exciting to see such progress in natural language processing. Large language models kind of blur the lines and begin to “understand” and respond to the ambiguities of the way we use language (at least in a kind of probabilistic sense). But they are also learning our programming languages. Right now, we can converse with AI models that can write (basic) code for us and with us, and make changes based on our conversational language. Imagine one day programming in plain language without that intermediary step!