Post

"surprisingly the first question was about loops" Extremely Not A Plant

Hmm, can’t imagine why they’d say the thing that involves spending at least three times as many tokens is what everyone should be doing

This sounds like it'll work about as well as my Satisfactory builds..

hype train never gunna fly off the tracks!

I've definitely seen dev influencers start pushing loops the last week.

The best part is that Steinberger and his followers, if they ever read this, won’t even understand the depth of the insult.

Every LLM has the same failure state where the context window eventually fills up and important information gets jettisoned, so let’s just speed that process along

Meh.. prediction machines can work they just aren't necessarily the most efficient.

Intel CEO advises customers that CPUs should always be fresh; replace every 400 hours.

I’m pretty sure Jensen Juan believes all money just belongs to him

lmfao

hahahah what the FUCK

Was that when he was in Taiwan recently?

Ed: surely we are at peak ridiculousness some dude: Jensen, hold my wallet

the people yearn for monarchy

wow cool they reinvented cron jobs and polling

Same word, entirely different usage. AI tokens are a measure of work done (and resources used) and are not arbitrary. The cost basis is similar to a utility so it only makes sense to charge metered rates. This is a reasonably understandable explainer of what a token is within the context of LLMs

One of my main clients requires us to use Claude, and I also loathe all the BS status updates it spits out about what it's doing when I prompt it. I told you to review this copy against the client's compliance guidelines; just produce the fuckin' word doc with the red lines, already.