Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
candiddevmike
13 days ago
|
parent
|
context
|
favorite
| on:
When AI writes the software, who verifies it?
> and it wasn't long ago before LLMs couldn't do math
They still can't do math.
help
Hammershaft
13 days ago
[–]
Pro models won gold at the international math olympiads?
reply
otabdeveloper4
13 days ago
|
parent
|
next
[–]
[*] According to cloud LLM provider benchmarks.
reply
bandrami
13 days ago
|
parent
|
prev
[–]
They have trouble adding two numbers accurately though
reply
evrimoztamur
13 days ago
|
root
|
parent
[–]
Why are they expected to?
reply
haspok
13 days ago
|
root
|
parent
|
next
[–]
If you believe the "AGI is just around the corner" hype...
reply
bandrami
12 days ago
|
root
|
parent
|
prev
[–]
Going to be hard to convince normies it can do harder things than that if it can't do that
reply
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
They still can't do math.