"LLM did something bad, then I asked it to clarify/explain itself" is not critical analysis but just an illustration of magic thinking.

Those systems generate tokens. That is all. They don't "know" or "understand" or can "explain" anything. There is no cognitive system at work that could respond meaningfully.

That's the same dumb shit as what was found in Apple Intelligence's system prompt: "Do not hallucinate" does nothing. All the tokens you give it as input just change the part of the word space that was stored in the network. "Explain your work" just leads the network to lean towards training data that has those kinds of phrases in it (like tests and solutions). It points the system at a different part but the system does not understand the command. It can't.

Petite bouteille à la mer:

Je vais faire un stage d'observation des pratiques d'égalité Filles/garçons dans un établissement scolaire de Bruxelles dans le cadre d'Erasmus du 06 au 11 avril prochain et je cherche un hébergement pas trop cher pour la semaine, si jamais vous avez des pistes n' hésitez pas.

Le repartage est doux et contribue à tuer Musk à petit feu.

622 fois la limite légale: chez Temu, les vêtements pour enfants sont ultra toxiques

https://korii.slate.fr/biz/622-fois-limite-legale-temu-aliexpress-shein-fast-fashion-vetements-toxiques-enfants-phtalates-plomb-cadmium

Regular reminder that Bluesky is a VC-funded company which inevitably is headed for the same fate shared by all VC-funded companies. It must maintain hypergrowth until it either:

- folds
- sells
- IPOs

Maybe tuning into another re-run of the social media VC show is appealing to some, but it certainly isn’t for me.