bluespin

Training the models still uses a vast amount of resources

Hearing about your experience was very helpful - thank you! I hope the followup procedure, if it happens, is more effective. Would love to hear how it goes if you remember. Best of luck!

It's an apt description of how these models function. They predict the most likely response to the input based on their training data. A brain can grasp concepts and reason about them - an LLM cannot

The US south often fits the west coast definition here as well

I would say the reverse but I think we're expressing the same sentiment

This image isn't how the painting looks

Absolutely! Thanks for the followup. Good luck with the procedure

I was responding to the bit about not caring about credibility. Engaging in a discussion is pointless unless others consider what you have to say, so credibility is a factor even in those situations. I agree with your opinion about not being too quick to cite sources in the same casual cases

Chiming into a discussion without caring if your point is considered is certainly a choice. Why comment at all, then?

We run one for our product and it's been teetering on the edge of extinction for a while. Google keeps saying they're going to kill them and going back on the decision

you would love jsx/tsx with react

Ideally you'd have a better place to work with that data than a UI that is displaying relative dates. Internal reporting data you can query for instance