More things that I’ve noticed about integrating LLMs into my workflow:
- People like to compare LLMs in their own field of expertise (to see whether they will be replaced, or just because that’s what they can test). But what I tend to use them for is things that I am bad at. I’m okay at writing, so I don’t really see much improvement there. I’m really not good at programming, and so I’ve seen an impressive improvement in my productivity by using an LLM to augment what I’m doing.
- It’s nice to be able something dumb questions without fear of looking stupid. Often the benefits seem to come from just the “rubber duck debugging” effect of just spelling out your thinking. The usual pattern is I ask a question, the LLM gives me a suggestion, I explain why that wouldn’t work, LLM apologises and offers something else. I point out why that wouldn’t work either, but start outlining something that could work. LLM commends me on my creativity, and starts spelling out what I could do to make that work, or its limitations.
- And by contrast, as the fear of looking stupid throught dumb questions declines, I’ve found myself feeling more confident asking questions in other contexts.
I also spent some time today catching up on that last piece of hype, Meta’s VR bid. I don’t like to dismiss anyone’s work, but it’s strange how Meta has been shifting tone from Oculus’s gaming vibe to something more … generic? Flat enterprise? People poke fun at Mark Zuckerberg’s avatar, but honestly it’s really hard not to look like cyberzuck in the new environment. It’s just got this very bland feel to it. Also, the rough edges from the old Oculus Quest software still seem to pervade the whole platform, but without the wow factor to drive it. It was kind of fun to mess around trying to get your hands to work on the Quest. In this new world, I mostly spend my time trying to link user accounts and clicking on privacy options. I feel like I’m moving slow over broken things.
(350 words)