Does anyone else think that LLMs have been regressing in the last 2 years. At least in the area of deep questions. Initially you could push the limits and actually bounce crazy ideas off of them and get some creative results.

Now they basically they always give you a canned answer that sounds like it was pre-approved by a committee.

They are not members of the party called the Conservative Party. It doesn’t stop them from being conservatives ideologically.

The article is very unclear on the kind of model they used. Several mentions of ChatGPT, but it doesn’t really sound like they used an LLM.

I really hope it’s not an LLM, this is a perfect case for specialized models trained just for surgery. I really wouldn’t want my surgeon to invent stuff when it doesn’t know what to do.

Good, keep convincing them that copyright is woke. And we don’t want anything woke right, wink wink

Humans suffer from the same problem. Racism and sexism are consequences of humans training on a flawed dataset, and overfitting the model.

Outsourcing your Defense Department to a foreign country is a choice. Not a smart choice, but a choice nonetheless.

It means max 220k, the big number is there to give false hope.

The difference between being technically right and knowing how to communicate.
There’s like a 100 different ways to reject a PR without referring to a contributor as a moron.

I think lately Linus learned how to do it correctly without going on rants.