[ChuckMcM]

Ruh-Roh. Chuckling aside, this is a really good thing. This is, btw, EXACTLY how enterprises figured out they weren't ready for "The Internet" in the dot com boom. They were spending millions on servers and kit and the ROI was crap.

[kotaKat]

The computer told them they ran out of tokens. The executives thought they could just put a quarter in the slot. The computer instead laughed and told them to put the whole financial quarter in it.

[indigodaddy]

Would humans be cheaper?

[stavros]

It's not about humans vs LLMs. It's about where the sweet spot of LLM usage by humans is.

[timcobb]

This is going to be very different though. These companies will just start investing in in house cheaper inference that's been on the front page here the last few weeks.

[ChuckMcM]

Oh absolutely. When asked by executives how I think they should adopt AI I've always been upfront about its strengths and weaknesses, and unlike a lot of technologies you can try out a lot of different things on much cheaper technology. As long as they throw out the notion that "more expensive" AI is "better" AI, it gives them a good way to compare workflows with and without machine learning augmentation.