Not for nothing, but couldn’t this be used to have AI play a game for 80,000 simulated hours and flag all the bugs? Human playtesters are important and have value, but no human should have to do the work of criss-crossing an enormous game map thousands of times just to see if the character model gets stuck on a random vertex sticking out somewhere, and yet it seems to be a distressingly common occurrence in more than a few games I’ve played.
Not for nothing, but couldn’t this be used to have AI play a game for 80,000 simulated hours and flag all the bugs? Human playtesters are important and have value, but no human should have to do the work of criss-crossing an enormous game map thousands of times just to see if the character model gets stuck on a random vertex sticking out somewhere, and yet it seems to be a distressingly common occurrence in more than a few games I’ve played.
Croteam did something like that for “Talos principle”. You can read here (points 9 and 10 in the article).
Nvidia is doing a lot of that. It’s slightly better than nothing. It also quite expensive unless you’re Nvidia.