@kravietz2048 @CR1337 No, he showed the same issues with only couple of minute old release again
English
Jiri Bernovsky
476 posts
The “hacking” report is nonsense because the author was testing a demo app with a mock authentication backend - whose only purpose is to make it work as a proof of concept. This was all written in the test app installation manual - he just chose to ignore it initially. The next day he corrected himself, but media picked up the initial hype 🤷
English
A short story:
1. European Union 🇪🇺 is releasing their mobile app to check people's age online
2. Ursula von der Leyen says "It is fully open source. Everyone can check the code."
3. Cybersecurity experts check the code and break the app in 2 minutes.
4. Chief Spokesperson Paula Pinho tells reporters, "Yes, it is ready. Maybe we can add, 'and it can always be improved'"
English
@Corejec @DarPutinovi Ale na začátku vytahuje pero z vlastní kapsy, když to jejich nepíše. Je to silný moment a měl by tam být, protože ne všichni si všimnou dvou per, ne všichni vědí, že vrací vlastní pero.
Čeština
@_Curious_Human_ @DarPutinovi Je jasně vidět, že do krabičky s protokolarnim perem přidává vlastní. Na konci jsou tam dvě.
Čeština
Český velvyslanec Pavel Bechný měl při podpisu memorand mezi ČR a Chile potíže s nepíšícím připraveným perem. Prezident Pavel mu proto nabídl pero vlastní, které po ceremoniálu nechal v krabičce na stole. Myslím, že všichni víme...😉
Čeština
@vonderleyen @EmmanuelMacron Hacked within 2 minutes, unsecure.
English
Thank you @emmanuelmacron for organising this discussion on the safety of our children online.
With the DSA, we have EU-wide rules.
And now we have an EU-wide app.
It’s piloted in 🇫🇷 🇩🇰 🇬🇷 🇮🇹 🇪🇸 🇨🇾 🇮🇪
And soon available to all.
Online platforms are held accountable.
Parents are reassured.
Emmanuel Macron@EmmanuelMacron
✅ Thanks for joining the movement.
English
@stevibe Man, you need to include info on used quants otherwise you're just confusing people:)
English
Qwen3.6 35B-A3B: smarter, but forgot how to use tools?
Running 6 Bench Packs on BenchLocal across 3 open-source Qwen models.
✅ ReasonMath: 92 vs 85 vs 86 — 3.6 wins
✅ InstructFollow: 97 / 97 / 97 — tied
❌ ToolCall: 83 vs 97 vs 100 — 3.6 tanks
Qwen3.5 27B still the tool-calling champ. 3.6 clearly leveled up reasoning, but tool use took a hit.
DataExtract live now.
BugFind + StructOutput next.
English
Jiri Bernovsky retweetet
As Czech FM Macinka visits Israel and claims to support moving Czech embassy to Jerusalem (but avoids any firm commitment), look at his 2-year old clip in which he slams the embassy move idea as the "biggest absurdity" and "complete farce" promoted by "incompetent politicians"🤡
English
@sudoingX I have asus pro art 16 w/5090, it's capped at 120 W and basically useless with gemma4. I get like 10 tps.
English
compiling llama.cpp with cuda 12.8 on the rog strix scar 18 5090.
this is the mobile 5090, 24gb vram in a laptop. same vram class as the 3090 desktop i run benchmarks on, just portable. i can run these tests from anywhere now.
first up carnice-27b vs qwen 3.5 27b dense vs gemma 4 on this exact machine. same class, same hardware and real agentic coding tests. data dropping soon.
English
@theo I just installed Hermes in Termux (proot Ubuntu). Working just fine even with local models.
English
I want to control my agents from my phone. I'm sure you do too.
What workflows do you want? Do you have one that works now? Tell me all about it.
English
@PernotLeplay @MistralAI They can't even train competitive model and he would already like to participate in European over-regulation and create a monopol. Sure.
English
Is @MistralAI CEO Arthur Mensch becoming the champion of EU tech sovereignty?
Mistral published 22 proposals to boost EU tech independence vs 🇺🇸&🇨🇳
In them Mensch pushes to develop European AI to ensure our armies can’t be turned off by foreign adversaries.
The proposals are, mainly:
🇪🇺“European preference” clause: public bodies must prioritize European cloud and AI providers when purchasing.
💶 Tax advantage for companies using European AI infrastructure.
🏢 Government procurement as a lever to keep critical workloads on European-controlled infrastructure.
👨💻“AI Blue Card”: a fast-track European visa for AI researchers and engineers, delivered in 15 days maximum, valid 4 years.
📊 A centralized European database of public domain works to train AI models with better access.
BFM Business@bfmbusiness
Arthur Mensch sonne l’alerte 🔔 Le patron de Mistral AI publie 22 mesures d’urgence pour éviter le décrochage technologique de l’Europe. 🎙️ @AnthonyMorel
English
@EU_Commission List results, achievements, something real that actually already works.
English
One year into our AI Continent Action Plan, we are delivering on our promise to work for a sovereign, trustworthy digital future.
We have deployed 19 AI factories, simplified rules, and launched the Data Union Strategy to unlock the potential of shared data ↓
link.europa.eu/nj3VH9
English
Love seeing these new 🦞 homes - share your setup and we'll start posting them.
andrew chen@andrewchen
built my openclaw a new home (mac mini + nvidia dgx spark in a 8U mini rack)
English
@bridgemindai My qwen3.5 122b does 60-80 tps based on quant, on 4x3090, 7 years old hw. Spark is the issue.
English
Spent the last few days vibe coding on my NVIDIA DGX Spark.
Here's what I learned.
Qwen 3.5 122B took one minute and nine seconds to respond "Hi how are you doing".
Unusable for vibe coding.
Gemma 4 was fast but built a dot instead of a first person shooter game.
GPT-OSS 120B was the sweet spot.
Fast, capable, and actually produced working HTML.
Open source models running locally are not replacing Claude Opus 4.6 or Codex with GPT 5.4.
Not even close.
But they're getting better every month.
The new DGX Spark Bench is live on bridgebench.ai.
Real-world benchmarks for local models on local hardware.
This is just the start. Full video below.
BridgeMind@bridgemindai
Claude Code rate limited me so hard I bought a $5,000 NVIDIA DGX Spark. Arriving tomorrow. A personal AI supercomputer. Anthropic cut off OpenClaw users. Slashed Claude Opus 4.6 rate limits. Told $200/month Max plan customers to use less. Then gave us a credit as an apology. This is what happens when AI companies have too much power over your workflow. One update and your entire stack breaks. Local models are the only infrastructure no one can throttle. No rate limits. No 529 errors. No surprise policy changes. Tomorrow I'm testing the DGX Spark live on stream. Running local models through real vibe coding workflows. The goal is simple. Never depend on a single provider again.
English
@Scobleizer Yep, I switched from openclaw a month ago and never looked back. Great with my local models.
English
My X feed is a constant stream of posts about Hermes like this one.
Zainan Victor Zhou@ZainanZhou
I tried a few hours Hermes Agent from @NousResearch , so far a few things I really love💗 (compare to @openclaw and even native @claude_code 1. self-fix and healing, when it try fix a problem, it remembers and learn from it automatically 2. better communication: in both TUI and Slack it prints out middle steps while finishing the task. @openclaw til today still can't reliablly communicate with Slack, which in part contribute to this issue github.com/openclaw/openc… and it seems pretty obvious Hermes has better concurrency management. 3. MUST BETTER SECURITY MODEL: instead of asking for permission each time, hermes actually only pause and ask when something is dangerous. So far, I think that's why people who have tried Hermes says OpenClaw: "here is the another fix" Hermes: "it just works" (actually not always but when it does, such as external dependency failures, it actually attempt, try and report much better" Kudos @Teknium and team
English
@_Honza_Dvorak qwen3.5 od 27b dense, přes 35b, 122b, po 397b v ruznych kvantizacich. Qwen-Coder-Next, ..., pro Hermes, OpenClaw, Claude Code Router, Claude Code, OpenCode, Perplexica, OpenWebUI, správa Ubuntu, MiroFish, správa Docokeru, psani scriptu, kodu, prace s textem. Musim to promazat :))
Jaký lokální LLM modely provozujete?
K čemu je používáte?
Čeština
Musím říct, že Gemma 4:31B je až překvapivě použitelná (na RTX 3090). Čekal jsem, že se to ani nehne, ale sype to výsledky velmi solidně.
Čeština
@sudoingX Just day after I spent 4 hrs on local deployment of Honcho for Hermes:))
English
holy shit hermes agent v0.7.0 just dropped and your memory is now fully pluggable.
7 providers out of the box from cloud to local sqlite. don't like any of them? build your own and plug it in.
credential pools. multiple API keys per provider with automatic rotation. key gets rate limited, next one picks up. stack this with the fallback provider chain from v0.6.0 and your inference never stops.
camofox anti-detection browser. stealth browsing with persistent sessions. your agent researches without getting blocked.
inline diff previews on every file change. secret exfiltration blocking catches credential leaks in URLs, base64, prompt injection. every release adds another security layer.
the repo crossed 24,000 stars. it was 13,000 two days ago. #3 fastest growing on github this week. 6 major releases in 22 days. 263 PRs merged in the last 6 days.
this is what happens when a founder gets on the field instead of managing from the sidelines. teknium merges contributions same day. i know because all 3 of my PRs were merged the day i opened them.
Teknium (e/λ)@Teknium
Deeper dive into some of the updates in v0.7 Memory: We have begun transitioning each of the systems in Hermes Agent to work through defined interfaces so that the core code is more maintainable, and more providers for everything can be supported. We started with memory: Now there is a plugin type that is for memory providers, pulled Honcho out of the core codebase, made it a plugin, and added 6 more systems that were waiting as PRs as new options. Now you can access your favorite much easier, and if your favorite isn't built in, it is easy to create a new plugin that does!
English
Getting a lot of comments about this.
Those of you who’ve used it - better than openclaw?
Nous Research@NousResearch
Hermes Agent runs in your CLI and through messaging platforms like Telegram, WhatsApp, Slack, and Discord - picking up and transferring sessions wherever you go. nousresearch.com/hermes-agent It has advanced agentic capabilities: command over subagents, programmatic tool calling, advanced filesystem/terminal control, agent-managed skills, browser use, scheduled tasks, and more.
English