Claude

[Owen_English]

Someone has to write the code for AI training to plagiarise.

Well not so as AI teaches itself and that's my point - it seems to be capable of learning all on its own.  Plagarise is not a word that AI knows or even cares about.  Noone has written any code to specifically teach it to deal with a PB request such as my request. Yes the code will be buggy but as time progresses, that will improve and at a rate that few appreciate. I will submit the code it produced and ask for it to improve it - let's see what happens then....

Of course there are massive risks here - I've asked it to do a few nefarious things like hack this or hack that (to discover the response) and there are boundaries with Claude where it refuses. But who is to say that those boundaries will prevail in the future? I'm an old-school programmer who has denied the existence of AI for long enough and having decided to play with it am astounded (and concerned).