@themarkup I particularly love your table illustrating short, plausible questions, chatbot responses, and conflicting reality. A+ work, and I want to see more articles use that structure to describe these flawed chatbots.

Much more concrete than just talking about "hallucinations".

@themarkup Please can we stop calling bullshit generators ‘AI’? This kind of problem is a lot more obvious if you use correct terminology at the start:

‘We are going to deploy a bullshit generator to help inform people.’

Who would think it is a good idea when phrased that way?