There’s been a lot shared here and posted elsewhere on the internet about the amazing things large language model AI like ChatGPT can produce. I thought I’d share this funny post as a helpful reminder that it makes terrible errors as well. It can be pretty funny, but this is where serious dangers exist for this technology as people may trust it as “all knowing” and take any answer it gives as truth.

This is actually where Google AI researchers were fired after writing On the Dangers of Stochastic Parrots. The paper warns of some dangers of these AI’s in being able to effectively guard the content it is trained on due to the sheer size, as well as the risks of people trusting it too much. Super interesting stuff, and this is definitely a technology that will not be going away.

ChatGPT tricked