Controversial runPythonCode() function debated on Linux kernel mailing list

An alien intelligence is probably the best description. It suffers from the same problem the rest of the internet has, lack of non-verbal communication. You can understand what a cat is thinking better than what ChatGPT is thinking.

Human children learn right and wrong from their parents and peers. There's no reason to think it will be any different for AI. Except for human children we have various mechanisms for punishment to encourage correct behaviour. How do you punish an AI chatbot? I'll guess we'll need simpler chatbots that do tell the truth to train those than don't?

Then again, perhaps we should heed the SFIA quote:

Keep it simple, keep it dumb,
else you'll be living under Skynets thumb.

We probably won't though. As a race we're exceedingly good at f*cking things up first and trying to fix them afterwards.