Re:AI benchmarks are a bad joke – and LLM makers are the one...
by nradov on Saturday November 08, 2025 @11:17AM
(#45857746)
LLMs can probably be taught or configured to use external tools like Excel or Mathematica when such calculations are needed. Just like humans. There are plenty of untapped optimization opportunities.
[ Parent
| on HN ]
Re:AI benchmarks are a bad joke – and LLM makers are the one...
by davedx on Saturday November 08, 2025 @02:15PM
(#45859142)
Chatgpt spins up a python sandbox for any complex calculations. It’s been able to do that for a while now
Re:AI benchmarks are a bad joke – and LLM makers are the one...
by joegibbs on Sunday November 09, 2025 @03:46AM
(#45863997)
I tried making a spreadsheet application and found that they’re not that great at working with 2D data, especially if there’s a lot of it. It’s harder to do search for a large spreadsheet than a large text file - you might get a range of thousands of numbers, how do you search that? And things like headers or important information may not be anywhere near where it’s focused which means it needs to read a ton of irrelevant context. For small sheets it works perfectly though, it’ll have to be something I’ll take another look at in the future.