|
|
|
|
|
by Hexstream
6616 days ago
|
|
Not sure it's so hard, first you introduce some randomness and inefficiencies in the gathering process so that it's more credible that it's a fallible human that's doing it. And as a second crucial step, you hook groups of bots into a "monitoring computer", at which an actual human sits and responds "No, I'm not a bot lol." when the supposedly fatal question arrives from an admin. You can even outsource this to the third world. |
|
Just the computer vision aspect of it alone would be amazingly hard.
Efficient goldfarming (e.g., with a frost mage that can take on 15-30 monsters at once) involves carefully coordinating a complicated series of actions where timing is critical and reacting quickly and intelligently when things go wrong is the difference between living and dying. It involves making complicated judgments about relative positioning, reasoning about surroundings, cause and effect, and other things that require human-level intelligence.
If you can make a bot that can seek out and kill monsters in a non-trivial manner without dying all the time there's no reason you couldn't program a robot to replace a wide variety of menial real-world jobs. The consequences would be tremendous and using it to make $20/day off the Internet would be retarded because you could license the technology for billions.