Anti-shibboleths

August 5, 2023

In the few areas where I have a clue about what people are thinking about, there's words that make me wonder whether this person is either clueless or assuming I'm clueless.

One example is that in New York City, if you ask someone what their favorite type of food is, and they respond with "Asian", it sets off a few alarm bells. I'm not a very sophisticated restaurant-goer, and even I know that a description like that is so broad as to be almost meaningless.

A more prominent example today is using the word "AI". I think there are smart people who use the word precisely (e.g. as in "AI risk") but usually people using it at this moment in time mean something like "RLHF-tuned large language models" or even specifically ChatGPT. I see this most on Twitter, though I've encountered people in real life who will tell me about what "AI can do now". I'm still deciding if this is a useful heuristic, but I change how I view someone's idea based on whether they say LLM or AI.

I explicitly don't want to say that you need to understand a lot of math to use these tools in creative and useful ways. Specialization is important and not all researchers want to do product design - but it's very useful in any field to be precise with terms.

This post is now outdated. As with all things, technology adoption moves faster than we can imagine, and now it feels standard to see 'AI' embedded in products in ways I didn't expect. I think now (and probably before) using the term 'LLM' is more confusing than just saying it's 'AI'. Unsurprisingly, I didn't realize that end users don't care about how the text is generated. You can disregard this post entirely.