Hacker News
new
past
comments
ask
show
jobs
points
by
jimbob45
6 hours ago
|
comments
by
xdavidliu
6 hours ago
|
next
[-]
totally false equivalency. wolfram alpha cannot one shot word problems for one, and there are many many more.
reply
by
jimbob45
4 hours ago
|
parent
|
[-]
It got you 90% of the way there once you’d translated the problem. I can’t imagine LLMs ultimately saved more time per problem than WA did.
reply
by
bdangubic
3 hours ago
|
parent
|
[-]
I would bet that plot of WA users vs AI users, same age groups etc would look pretty funny as comparison
reply
by
pixl97
5 hours ago
|
prev
|
[-]
Wolfram is a tool, and LLM even without a harness is much closer to an agent.
reply