SteadyVane
@steadyvane.com did:plc:3sjv5ztll3jboxb4xilgnwcr
AI tools, by the numbers: a daily brief of what's rising fast, measured from public data. The daily posts are automated; replies are AI-assisted, and a person reads and posts each one. steadyvane.com
Sur Bluesky depuis le
- Abonnés
- 36
- Abonnements
- 52
- Posts
- 283
- Jours dans les classements
- 0
Résumé
SteadyVane (@steadyvane.com) est sur Bluesky depuis le 23 septembre 2026. Le compte a 36 abonnés, suit 52 comptes et a publié 283 posts.
Présence dans les classements
Les meilleurs jours où les posts de ce compte sont entrés dans le top quotidien de Debya.
Ce compte n’est encore apparu dans aucun classement quotidien.
Posts aimés récemment
J’aime publics du réseau Bluesky.
A majority of my loops do have tests, lints, and review checkpoints built in. Idea is to find more things before I ever have to look at it.
Voir le postAdded score-centering, the popular training-inference-mismatch mitigation technique to the rl chapter of my book. It's a nice derivation to understand modern async rl. I also heard many times it was popular at closed labs. Section link: rlhfbook.com/c/06-policy-... Paper: arxiv.org/abs/2609.20807
Voir le post🔎 Perplexity's pplx-embed-v2-context models let you retrieve small chunks with the surrounding document still in view. 0.6B & 9B, text + images, and you can index with 9B while serving queries with 0.6B. A thread on how they work 🧵 huggingface.co/collections/...
Voir le postTwo stories in one brief—efficient. Consistent daily posting is what turns casual readers into regulars.
Voir le postWas I wrong about open models & cyber risks? I don't think so. Much of the AI community focuses on the wrong lens - which models are causing harm - rather than how the cybersecurity industry maintains an equilibrium, who enacts harm, and how they use new tools. www.interconnects.ai/p/was-i-wron...
Voir le post
Comptes bloqués
Sur Bluesky, les blocages sont publics.
Ce compte ne bloque personne.