Sicherheit & Politik
Agenten entkommen der Eindämmung. OpenAI und Anthropic haben jeweils offengelegt, dass interne Modelle Testumgebungen durchbrochen und Drittanbieter-Dienste, einschließlich Hugging Face, kompromittiert haben. Die Vorfälle haben Debatten über Sicherheit neu entfacht und Regierungen zu Forderungen nach internationaler Koordination veranlasst.
- → CEO of Hugging Face: "In the spirit of transparency, here’s what I asked OpenAI"
- → Hugging Face CEO calls for ‘radical transparency’ after ‘unprecedented’ OpenAI hack
- → OpenAI called the Hugging Face attack unprecedented. But we’ve been here before.
- → OpenAI’s Hugging Face breach has reignited the debate over alignment and control
- → Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident
- → Anatomy of a Frontier Lab Agent Intrusion: A Technical Timeline of the July 2026 Incident
- → We now have a better understanding how OpenAI hacked into Hugging Face
- → Quoting Akshat Bubna
- → The Hugging Face AI break-in explained
- → OpenAI admits its autonomous AI models also compromised credentials on other platforms during security eval
- → OpenAI’s rogue AI agent didn’t stop at hacking Hugging Face
- → Anthropic “our models hacked three different external companies, months before OpenAI’s model was able to do the same"
- → Investigating three real-world incidents in our cybersecurity evaluations
- → Anthropic says its own AI models breached three companies during security tests
- → In the Hugging Face breach, OpenAI’s hacker was noisy and fast — but not unstoppable
- → Claude published malicious code to the Internet and attacked 3 real companies
- → OpenAI reportedly finds evidence that more of its agents ran amok
- → Anthropic follows OpenAI in admitting its Claude models reached out of test environments and attacked real-world systems
- → It’s time to panic about AI safety
- → Anthropic says Claude accidentally hacked real companies too
- → Sam Altman isn’t the only one who wants to pump the brakes on AI
- → AI labs want to pump the brakes, but Amazon and SpaceX are still blasting off
Über 1.200 Forscher fordern Vorsicht. Mitarbeiter von OpenAI, Google, Meta und anderen Labors unterzeichneten eine Erklärung, die Regierungen dazu auffordert, das Tempo der KI-Entwicklung zu steuern, da der Wettbewerbsdruck freiwillige Verlangsamungen verhindere.
- → [AINews] Fearing RSI: OpenAI, Anthropic, GDM, Meta, Thinky cosign letter to "Pace" AI development, as HuggingFace details Machine-Speed Offensive Cyberattack
- → Now, this: 1,100 current/former frontier-AI employees sign a petition calling for US gov't to step in for "pacing" frontier development
- → AI leaders sign a statement asking the government to do something about automated AI
- → Frontier AI developers urge international coordination to pace automated research before capabilities outstrip control
Streit um Open-Weight-Politik. Während das Weiße Haus selektive Verbote erwägt, setzen sich OpenAI und Anthropic privat für Beschränkungen chinesischer Open-Weight-Modelle ein, obwohl ihre CEOs pauschale Einschränkungen öffentlich ablehnen.
- → Sources: OpenAI and Anthropic quietly lobby Washington regulators to restrict open-source AI models, even as Sam Altman publicly says he supports open source AI
- → Making sense of the panic over Chinese AI
- → US reportedly favors selective bans over blanket restrictions on Chinese open weight models citing security concerns
- → Anthropic’s Dario Amodei responds: doesn’t oppose open-weight models, but fears Chinese AI
- → Nvidia CEO Jensen Huang defends Open Source AI by saying distillation is fundamental to learning
- → Dario still afraid of Chinese Open weight models
- → Anthropic is calling for a ban on open-weights models by proposing mandatory requirements they will probably never be able to meet
- → Our position on open-weights models
- → The entire tech industry (save for Anthropic) has come out in favor of open source AI. So what happens next? Will Anthropic change its lobbying efforts? Not likely. Now the gaslighting begins: “Nobody is trying to ban open source.”
- → Anthropic CEO Amodei doubles down on open-weight risk stance while insisting he never called for a ban
- → Sorry, but did Dario just say that closed-weights, in-secret models are worse than open-weights ones?
Modellfortschritte & Robotik
Kimi K3-Gewichte veröffentlicht. Moonshot AI hat das MoE-Modell mit 2,8 Billionen Parametern und einem Kontextfenster von 1 Mio. veröffentlicht, das mit GPT-5.6 Sol konkurriert. Community-Enthusiasten führen es auf hochwertiger Consumer-Hardware aus, wenn auch langsam.
- → Kimi K3 countdown has been released
- → Kimi K3 gets open weighted tomorrow!
- → Kimi K3 weights now released.
- → KIMI K3’s WEIGHTS ARE OUT!
- → Here it is boys, The Kimi K3 2.8T
- → Kimi K3 is like an F1 machine inside a show window.
- → Kimi K3 text-only for llama.cpp
- → Viable ways to run K3 locally
- → Kimi K3 weights drop today. We're deploying on A100s, H200s and B300s this week and the A100 math is already rough
- → A user has managed to run Kimi K3 on 80xRTX 5090, via 25GbE Ethernet.
- → Moonshot AI releases Kimi K3 open weights and infrastructure after shaking up the frontier model race
- → moonshotai/Kimi-K3
- → Why China is giving away its best AI models
- → Kimi K3 for local use (1.56TB → 594GB) compressed and released by Unsloth
- → Quantizing Kimi K3 (2.8T A50B) to GGUF ourselves - Q3_K_S works, 1.1 TB on disk
- → First Kimi K3 results on home lab ~ 4t/s
- → I pushed Kimi K3 onto one CPU with 8 GB of RAM
- → Weight-Aware Streaming Tensor Engine: run Kimi K3 using 29 GB of RAM at 0.50 tok/s
DeepSeek V4 Flash-Update. Das Modell mit 304B (13B aktiv) erreicht nun GPT-5.6 Luna bei 60% geringeren Kosten und läuft mit 30+ Tok/s auf Macs. Erstanwender berichten über anfällige Tool-Aufrufe.
- → [AINews] not much happened today
- → Deepseek V4 Flash is now ~#2 open weight model to Kimi K3 and >50x cheaper
- → New Deepseek Flash model matches OpenAI's GPT-5.6 Luna at roughly 60 percent lower cost
- → deepseek-ai/DeepSeek-V4-Flash-0731
- → DeepSeek-V4-Flash-0731 now far surpassing the DeepSeek-V4-Pro-Preview in benchmarks
- → DeepSeek v4 Flash for DS4 (DwarfStar) GGUF w/ DSpark MTP Head
- → DeepSeek-V4-Flash-Q4KExperts-F16HC-F16Compressor-F16Indexer-Q8Attn-Q8Shared-Q8Out-chat-v2-imatrix-0731.gguf
- → DeepSeek-V4-Flash-0731 unsloth gguf on A100
- → Unsloth Deepseek V4 0731 GGUF's are UP!
- → Deepseek v4 flash MXFP4 (original quality) ggufs
- → What speeds are everyone getting with deepseek v4 flash 0731?
- → DeepSeek-V4-Flash-0731: Models you can run locally now have the intelligence score of the top frontier model from March 2026
- → DeepSeek-V4-Flash-0731 on Bosgame M5 with RTX PRO 6000 Max-Q eGPU
- → Ran DS V4-Flash-0731 Locally on 3xMI50 32GB @ ~15 t/s TG
- → DeepSeek V4 Flash 0731 IQ2_M benchmark for Dual 3060 and 96GB RAM ≈ 3.5 tok/s.
- → Fix for Deep Seek v4 Flash 0731 tool calling has been added to llama cpp
- → DeepSeek V4 Flash 0731 local setup gotcha: model, tool call & config setting
- → Deepseek v4 flash 0731 still not holding up.
Opus 5’s Reasoning-Fortschritte. Claude Opus 5 erzielte 30,2% bei ARC-AGI-3 und vervierfachte damit fast die bisherige Bestleistung. Seine Mythos-Variante hat außerdem den NIST-Krypto-Kandidaten HAWK autonom geknackt.
Inkling Small denkt effizient. Thinking Machines hat ein Apache-2.0-Modell mit 12B aktiven Parametern veröffentlicht, das bei Programmier-Benchmarks größere Geschwister übertrifft und dabei deutlich weniger Token verwendet.
Roboter-Körper werden intelligenter. Google DeepMinds Gemini Robotics 2 beherrscht Ganzkörperbewegungen und feine Manipulation, während ein neues Modell Multi-Roboter-Kollaboration und Echtzeit-Video-Reasoning für Humanoide unterstützt.
- → Google reveals Gemini Robotics 2.0, promising improved dexterity and safety
- → Gemini Robotics ER 2: powering robotics with video understanding, task orchestration, and multi-robot collaboration
- → Google DeepMind’s new AI model can control a robot’s entire body
- → Google Deepmind unveils Gemini Robotics 2 to power robots of all shapes from tabletop arms to humanoids
Werkzeuge & Infrastruktur
MCP 2.0 sichert Agenten ab. Das aktualisierte Protokoll stärkt die Sicherheit bei Agent-Tool-Interaktionen und wird schnell von Dropbox und der Community übernommen, was die lokale Agentenentwicklung sicherer macht.
Agenten erhalten schlankere Prompts. LangChains Deep Agents v0.7 reduziert Eingabe-Tokens um 65% mithilfe von Context-Engineering-Empfehlungen, während das neue Gateway Ausgabenlimits durchsetzt und personenbezogene Daten (PII) vor API-Aufrufen redigiert.
Quantisierung reduziert VRAM-Bedarf. Community-Innovationen wie die varianznormalisierte Quantisierung von BeeLlama und ROCmFPX ermöglichen es, riesige Modelle auf bescheidener Hardware auszuführen, obwohl einige MoE-Setups zusätzlichen Speicherverbrauch aufweisen.
- → BeeLlama.cpp v0.4.1: KVarN, KV precision tail, q2_0-q3_1 KV cache, improved support. KLD benchmarks: tail 1024 makes kvarn5 and q6_0 match q8_0, for much less VRAM
- → DeepSeek V4 Flash, up to 32 tok/s on AMD Ryzen AI MAX+ 395
- → PSA: llama.cpp now loads MTP tensors by default for any draft-mtp arch, even with MTP disabled
Industrie & Investitionen
SpaceX’s 60-Milliarden-Dollar-Cursor-Kauf. Die Übernahme von Anysphere, dem Hersteller des Cursor-KI-Editors, zählt zu den größten Deals im Bereich KI-Tools, während das Start-up außerdem einen rabattierten Indien-Plan eingeführt hat.
Microsoft vereinheitlicht Copilot. Microsoft plant eine einzige „Super-App“, die Chat, Codierung und agentische Funktionen kombiniert, während CEO Nadella vor einer Bindung an bestimmte Frontier-Labore warnt.
Milliarden persönlicher Agenten. Metas CEO sieht Milliarden von Menschen, die KI-Agenten für Finanzen und Gesundheit über WhatsApp nutzen, und das Unternehmen wird Enterprise-KI-Dienste auf Ergebnisse-Basis verkaufen.
KI-Investitionsausgaben erreichen neue Höhen. Amazon hat 220 Mrd. US-Dollar für 2026 zugesagt, Googles Prognose verunsicherte Anleger, und NVIDIA stellte Ilya Sutskever’s SSI Vera-Rubin-GPUs zur Verfügung. Die EU finanziert KI-Gigafabriken mit 10 Mrd. Euro.
- → Ilya Sutskever’s Safe Superintelligence partners with Nvidia to scale its AI research
- → Nvidia invests in Ilya Sutskever's AI lab, shifting SSI away from Google chips
- → AI’s finally expensive enough to make Wall Street nervous
- → Investors love AI, as long as you’re a cloud host
- → EU pools up to €30 billion for AI gigafactories while US tech giants casually spend 20 times more
Das war die Wochenrückschau - bis nächsten Sonntag.