Why Are We Not Talking About This?

I really wish we'd stop anthropomorphizing these things. It isn't "thinking", it's processing. They don't "lie" or "hallucinate", they generate incorrect output. Concepts like thinking or lieing literally do no exist in this context.

The I in LLM stands for intelligence, and the E in AI stands for expertise.

PERIOD OUTSIDE THE QUOTATIONS MARKS GANG "UNITE".

"Questionable licensing" is being very generous. Using non-profit research loopholes to ingest everyone's IP (no licensing unless an entity with big enough pockets threatened them), and then turning on the dollar generator should be criminal.

A good 33-50% of my programmer job is re-interpreting the naive instructions from non-programmers, and programmers who have not actually solve the problem, in to a workable solution that takes in account all the hard-to-knows or unknowns from the original assumptive idea.

The AI bros are missing the forest for the trees, in that programming languages are far simpler than natural languages, due to having very small grammar and being context free. They think learning a programming language is hard because it's unfamiliar, but they've already put more effort into programming with LLMs than it would take for everyone to just learn basic python.

“We just need to tell it all the details and it’ll write the code for us.” Those programs exist. They’re not LLMs, but called “compilers,” and the details are known as “code.”

sad, we didnt get to see what chat gpt recommended for embedded

lmao, called the 1hr long video

So basically AI is the "I'm feeling lucky" button from Google, which would just take you straight you the first result.