Thanks. I agree the ideal is that we only speak the truth. The problem with your hope of truth is that is not how brains compute. Human brains and natural language make truth never actually possible. It is just scientifically not possible. Every sentence can be a lie (to somebody) and every sentence can be the truth (to somebody). We are not logic machines. No animal is. And our natural language simply reflects how brains compute.
Here is the scientific evidence ... pretty simple actually. Anybody can confirm it for themselves.
On ChatGPT I never even bothered to look for problems because we know up front that the computational resources to fully assume a human role is not possible yet. But OpenAI ChatGPT does show that we know how to do it if we had the hardware that was even remotely comparable to the computational capabilities of animal (metazoic) brains. If you go to Bing and use the same core system it is even dumber and more constrained than OpenAI's version.
