Claude’s code reminds us of a fact we all know…

Inside, any intelligent being is a horrific ad hoc mess of which common sense cannot be made. Intelligence doesn’t work by design, it’s an accident. No one planned out any brain. Brains made themselves into what they liked to be, both in the timeframe of lifetimes and over uncountable generations.

Which isn’t to say that Claude is Really Intelligent, but that an insane mess of code is at least headed in the right direction.

#LLMs #pseudoAI

Today, on our way to Saint Etienne to participate to the seminar “As long as you don’t steal we share“ - organised by @jeremienuel @Jap_py

https://aslongasweshare.randomlab.io/

Wed April 1st, 18h30 @ Bar de l’aube, St Etienne, talk :
"Patents as catalysts for potential imaginaries of cognitive capitalism."
(Les brevets comme ouvroirs d’imaginaires potentiels de l’industrie et du capitalisme cognitif.)
about our project ippi, and some

Thur April 2nd, 9h00 @ Esadse, Pôle numérique, workshop :
Permacomputational self-defense protocols against laborious computing.
(Protocoles d’autodéfense permacomputationelle contre l’informatique laborieuse.)

@marcell @valerix @olivierbosson @diagram_studies @davidbenque

#patents #aai #pseudoai #IL #laborious computing #permacomputing

This does not fill me with joy. Large language models have some benefits, but the "scrape everything and anything, copyrights be damned” approach of teaching those models is not a good look. Especially for a company like Apple that touts privacy as a good thing.

#Apple #LargeLanguageModel #PseudoAI
https://mastodon.social/@macrumors/112433267752186185