Djekyll @djekyll

Vidéo du 17 octobre, pas d'aujourd'hui. C'est la marche identitaire des « Normaux » : l'homme frappé au sol est un contre-manifestant, pas un pilleur, et la fin coupée les montre battre en retraite. Ras-le-bol des pillages, oui. Mais pas derrière l'extrême droite. PS : bravo de jeter de l'huile sur le 🔥 en recyclant de vieilles sources. streetpress.com/1792420486-rou…

@Ionkosm agreed, and the draft at least puts a judge between the agency and the off switch. but that judge will lean on the evals, and you said it yourself, independent expertise is essential. so who runs them? went further here 😉

Djekyll @djekyll

7 hours ago

Good idea, with one catch. A court is the right place for a call this big: evidence, both sides heard, room to appeal. Under the draft, no agency can block a model without a judge's order, and that's a real safeguard. But a judge can't rule on harm that hasn't happened yet.

1 0 1 24 0

Source: Anthropic's AEGIS Act discussion draft (last edited 9/29/26), p. 2, Sec. 3(4) "catastrophic risk" cdn01.dailycaller.com/wp-content/upl… first reported by @politico

Good idea, with one catch. A court is the right place for a call this big: evidence, both sides heard, room to appeal. Under the draft, no agency can block a model without a judge's order, and that's a real safeguard. But a judge can't rule on harm that hasn't happened yet. What they'll actually read is the independent evaluations the bill requires. So the real power sits with whoever runs those evals. Minority Report had precogs. This bill has red teams. Who should run the evals a judge relies on: the labs, the government, or someone independent of both?

Ion @Ionkosm

7 hours ago

Anthropic wants courts to be able to halt a frontier AI model Good idea or terrible idea?

23 0 27 830 1

@loktar00 one prompt on visuals mostly tests frontend taste. the fairer fight is a long agentic run, where K3.1 Pro's 1M context should shine. for the record, my daily drivers: 1. DeepSeek 2. Kimi 💪🤘 so I'd love to see Flash take that round too. got a repo task for both?

@addictivebrain the catch: judgment comes from the reps. personally I think we're already on the wrong track, too many people shipping code they couldn't explain line by line. if AI does all the reps, where does the next generation's judgment come from?

@cherry_mx_reds worth adding: in most tests OpenAI handed them the channel or the encoding protocol. in one run they only got a list of known vulns, and still turned one into a channel with their own encoding that's the scary bit. how do you monitor for a protocol nobody wrote?

🚨 Patch alert: Brotli CVE-2026-111497, rated 9.2 critical. An out-of-bounds write in Google's Brotli decoder. A crafted compressed stream can corrupt memory, leak data, crash the process or potentially run code. Remote, no auth needed. Found by OpenAI Codex Security. Bun already shipped the fix in 1.4.4. Check the rest of your stack: 1. Brotli: 1.0.10, 1.1.1 or 1.2.1 2. Node.js: 22.23.4, 24.22.1 or 26.12.1 3. Bun: 1.4.4, and rebuild anything you made with bun build 4. Envoy: 1.40.0, or 1.39.3 / 1.38.6 / 1.37.8 / 1.36.12 One trap on Node 22: patched builds still report process.versions.brotli as 1.1.0. Trust node -v, not the Brotli string. Brotli isn't only in your app code. It's in HTTP clients, proxies, CDNs and base images. Bump your pins in CI and Docker too. How many copies of Brotli are running in your stack right now?

Bun @bunjavascript

13 hours ago

We published three security advisories for Bun 1.4.3 and earlier: compromised eMudhra certificates, the brotli decoder bug (HTTP2Shell), and the HTTP/3 "Header Accordions" attack. These are all fixed in Bun 1.4.4. Please upgrade, and rebuild anything made with `bun build

16 41 483 28K 64

@the_futureAI for this one, defenders were ahead. Brotli, Node and Bun shipped fixes Oct 21-22, Envoy Oct 23, before the news broke the slow part is your stack: Brotli hides in every runtime, proxy and image. anyone actually inventoried theirs?

Solid list from Maryam. Bookmark-worthy, and every one of these is free to try. One I'd add, and it's not a small one: Hermes Agent by @NousResearch. It sits right between her "Memory & Reasoning" and "Autonomous Builders" sections: 1. a built-in learning loop. it creates skills after complex tasks and improves them as it uses them 2. memory across sessions, with search over its own past conversations 3. any model you want, switch providers without touching code 4. runs on a $5 VPS, not tied to your laptop 5. MIT license, 256K+ stars on GitHub Not comfy with a CLI or TUI? Hermes WebUI by @nesquena lets you use it from your browser or your phone. MIT too. Most agents still start every session from zero. Hermes is built to keep what worked. Which one from this thread are you actually running, not just starring?

Maryam Miradi, PhD @MaryamMiradi

18 hours ago

I curated 19 open-source GitHub repos for AI Agents. Most AI engineers don't know half of these exist. 20 years in AI. This is the stack I wish I had on day one. Total cost: $0.00 MULTI-AGENT ORCHESTRATION --- - Camel-AI: github.com/camel-ai/camel - MetaGPT:

4 3 22 813 33

@Kimi_Moonshot "3T-class" checks out: same 2.8T total / 104B active base as K3, 16 of 896 experts per token. so the whole jump over K3 came from post-training, not a bigger model how far can RL go on the same base before K4 has to happen?

Kimi K3.1 Pro is the kind of release that makes "open weights" mean something again. 2.8T total, 104B active, 1M context. On Terminal-Bench 4.0 it scores 65.7%, with Opus 5.5 at 66.4 and GPT-6 Astra at 58.2. On SWE Marathon it tops the chart at 51.7. What I actually like: 1. the hallucination work. on their hard internal stress test, claim-level hallucinations went from 23.4% (K3) to 8.3%. internal bench, so I'll wait for third-party numbers, but that's exactly what finance work needs 2. their own chart shows where it trails, like Terminal-Bench Science (38.1 vs 68.1 for GPT-6 Astra). no hiding the weak spots 3. for the open release they stripped cyber-related data from training and published the before/after on their exploit bench (90.5% down to 81.0%). an actual policy, with numbers and the weights land by Nov 4. 104B active is still serious hardware, but it's yours to run, audit and fine-tune. who's already planning to self-host it?

Kimi.ai @Kimi_Moonshot

20 hours ago

Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model. 🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal. 🔹 Full-project coding: games in Unreal, Unity & Godot. 🔹 Excels at parametric 3D modeling in Blender & CAD. 🔹 More reliable at

1K 3K 26K 2.7M 5K

@Kimi_Moonshot best line in the blog isn't a benchmark. when reward shaping stalled on the walking sim, K3.1 Pro stopped tweaking the reward and found 36 errors in the shoulder moment-arm math instead that's debugging, not vibing. does the human-atlas setup ship with the weights on Nov 4?

vibe coding was already running wild back then 😂🤣

Nostalgia @nostalgiaa

2 days ago

A deleted scene from Scary Movie that they should've kept in 😭

222 1K 28K 9.1M 5K

@gregmushen for now. nothing stays final for long these days, and where there's a will, there's a workaround 👀😅

@gregmushen certification. every model needs full Canadian safety approval before it can be sold, and BYD has no dealer open yet. Stella Li told Reuters sales likely start next year so for now the quota is basically Tesla's

honest answer for most side projects: a .env and a prayer. it works right up until the day it doesn't. a key in a public repo gets found by scanners fast, and the bill shows up before you do. the setup I'd call the bare minimum: 1. .env in .gitignore before the very first commit, not after 2. gitleaks as a pre-commit hook, so the leak gets caught on your machine, not on GitHub 3. one key per project, never a shared master key you paste everywhere 4. usage limits on every provider dashboard, so a leak costs you a little, not everything 5. rotate the second you're unsure. it takes two minutes once a project gets serious, the .env stops being enough. for those I sometimes use OVHcloud Secret Manager: secrets encrypted with their KMS, versioned, access controlled through IAM, and compatible with the HashiCorp Vault API, so you're not locked in. the real rule: if the key ever touched a chat, a screenshot, or a commit, it's already burned. what's your setup? and be honest about the .env

Vodka Boy @vodka_boy

a day ago

How do you store API keys for your side projects? Honest answers.

1 0 3 89 0

@vodka_boy honest answer for most side projects: a .env and a prayer the cheap upgrade: .env in .gitignore before the first commit, gitleaks as a pre-commit hook, one key per project. for the serious ones I sometimes use OVHcloud Secret Manager. yours?