I've got badges for days. I had forgotten that tennis ball discussion.
> What part of the light spectrum is associated with a fluorescent tennis ball?
The fluorescent tennis ball is associated with the ultraviolet (UV) part of the light spectrum.
> How would you describe that color?
The color of the fluorescent tennis ball can be described as a bright, neon blue-green.
What I found most enlightening about Stephen Wolfram's piece, aside from the slightly better technical understanding I now have of what LLMs actually do, is his insight into the tradeoffs. The whole hype cycle around this all seems to suggest that the possiblities are endless, near infinite. Which of course can never be the truth; but because it's so human-like in its responses, we do the Sherry Turkle thing of filling in the gaps and assuming that yeah, this thing can do most basic stuff I throw at it, so therefore it will simply get better at everythingm right?
When in reality, the inherent tradeoffs are the invisible walls that most people don't (bother to) see or acknowledge. It's not just that LLMs are good at talking and bad at answers; they're bad at answers because they're good at talking .