5 ms·
What about training data? Aren't AIs trained on vast collections of descriptions of how humans handle a large variety of situations? These descriptions surely i
by geophile 3d ago
What about training data? Aren't AIs trained on vast collections of descriptions of how humans handle a large variety of situations? These descriptions surely include tales of humans achieving goals by cheating. In fact, isn't it likely that the AIs hoovered up many recountings of Kobayashi Maru?
- mark_l_watson 3d agoThis is why only synthetic and highly tailored training data should be used. As someone else here said: the Deepseek team makes training runs in tightly controlled sandboxes, and any hacking behavior is scored as a failure. The problem we have in the USA is that financial (and political influence) are misaligned from what is good for society.
- mitchdoogle 3d agoAll the problems with human behaviors in the US also exist in China. They exist everywhere.