Thread 106176267 - /g/ [Archived: 302 hours ago]

Anonymous
8/7/2025, 5:35:52 PM No.106176267
img_introducing
img_introducing
md5: 14fcf983a372b64aecac1040ca09ded9🔍
is it possible to train a robot in a simulation to hunt down humans and force them into the dataset caverns?
Replies: >>106176351 >>106176453 >>106176552
Anonymous
8/7/2025, 5:43:45 PM No.106176351
>>106176267 (OP)
Not for you if you have to ask here
Replies: >>106176439
Anonymous
8/7/2025, 5:51:45 PM No.106176439
>>106176351
>Not for you if you have to ask here
but i was told by the worms in my head that this board is the ultimate source of technological insight.
Replies: >>106176504
Anonymous
8/7/2025, 5:53:31 PM No.106176453
>>106176267 (OP)
Theoretically yes but it would be extremely difficult with current technology. Just tracking humans would be a problem, let alone running, capturing without killing, and then giving clear instructions while somehow keeping humans from running away.
Replies: >>106176578
Anonymous
8/7/2025, 5:57:25 PM No.106176504
>>106176439
Yes, and you just got the full gist of it.
Next!
Anonymous
8/7/2025, 6:01:00 PM No.106176552
>>106176267 (OP)
Not currently
Anonymous
8/7/2025, 6:03:57 PM No.106176578
>>106176453
i ask cuz i recently saw that simulations can be to train robots using a vr headset that hand tracks you as guidance. you could just reenact capturing a human and then it runs like a thousand different simulations in parallel each with a thousand variations in parameters or some shit. then the end result is an optimized hunting bot. Can a consumer access such processing power?