Because I've only been working with AI agents, the information has been too scarce or limited. What I tried this time is...
Doom
Limmyda
You know Doom, right? The ones with displays are divided into whether or not Doom runs. In a way, Doom can be seen as the minimum performance that allows you to see the minimum specifications of the display and the operating system.
I've tried challenging Doom a few times with my entry-level autobot, but I gave up because it wasn't enough learning to digest Doom. This time, I'm going to proceed with a cloud model.
I'm going to run it in an overnight infinite loop and give it memo functionality.
Will it really pass Stage 1....
Doong!!
I'm curious to see if a general-purpose model can actually clear Doom.
If it can be cleared, I'll use this data as training data for my autobot later.
Yoohoo!!
