ChatGPT already lies to us and we’re just getting started.

in section .v I make this point that ChatGPT cannot be used for summarizing as we always have to cross-check what it reports back (as else it can return total non sense)

https://proofinprogress.com/posts/2023-02-01/the-ai-crowd-is-mad.html

Yeah. Mf tried to sell me on the idea that square root of 4 and square root of 16 were the same thing. Can't trust machines these days.

Seems like there will be many flavors of AI based on the biases of the creators

I tried ChatGPT to help me write some code (I am not a developer myself), and it took me 2 hours to troubleshoot and go back and forth with errors to ChatGPT, until I gave up and found working code on Stack Overflow 🤷‍♂️

This idea broke my head. Real danger ⚠️

Verifying ChatGPT is accurate is fertile ground for startups. Or perhaps it is taking the output of ChatGPT (or Large Language Models generally) and updating it so it is correct. Or showing where it needs some human input. 

The startups would be vertically focused bringing expertise and perhaps data to the equation.