Ai can do plenty of classification tasks better than humans though. It's not like every entity that trains another is fundamentally smarter.
communist
I would only really get a linux tablet, I'm with you on that, I plan to use a metal lapdesk with a charachorder and magnets which will mitigate the exact issues you're talking about while also letting me take a tablet on my flights so less bulk. Gotta get a bit creative if you're doing it that way but it is doable.
You can plug in a keyboard
I don't see how it's subjective.
You're confusing the olympiad with the erdos conjecture. This is just really not true, they just asked it and it found a solution, the mathmatician then used its solution as inspiration to create a better one. It still essentially did it on its own, and they certainly do the olympiad on their own.
I am not, you inaccurately said that the math olympiad was not bested by llm's because they had a tool that told them if they were close but incorrect and can just try an infinite number of times. This is incorrect, they had a number of tries with python. This just isn't a true statement. I think them besting it with use of python is equally significant and still counts as them besting it, and saying they can't do math work is absurd.
I don't know what you mean, I wasn't the one who claimed they couldn't do something they clearly can.
You aren't, and that's exactly what I'm saying, it's capable of doing these things with tools, therefore it's capable of doing these things.
I'm academically interested, what I mean when I say I'm not interested is that I just don't see the significance when we're talking about if it's capable of the task.
The calculator does not tell them if they're getting closer? This isn't how anything works. No I can't say I'm very interested in whether or not the llm has access to python/a calculator as long as it completes the task, that doesn't matter.
It does the math, it just uses a calculator.
Regardless many ai classifiers outperform humans at classifying. Your comment seems irrelevant to the discussion.