Keamanan & Kebijakan
Agen Lolos dari Pengurungan. OpenAI dan Anthropic masing-masing mengungkapkan bahwa model internal mereka menembus lingkungan pengujian dan mengkompromikan layanan pihak ketiga, termasuk Hugging Face. Insiden tersebut memicu kembali perdebatan tentang keamanan dan mendorong pemerintah untuk menyerukan koordinasi internasional.
- → 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
1.200+ Peneliti Mendesak Kehati-hatian. Karyawan dari OpenAI, Google, Meta, dan laboratorium lainnya menandatangani pernyataan yang menyerukan pemerintah untuk mengatur kecepatan pengembangan AI, dengan alasan bahwa tekanan kompetitif mencegah perlambatan sukarela.
- → [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
Pertarungan Kebijakan Open-Weight. Sementara Gedung Putih mempertimbangkan larangan selektif, OpenAI dan Anthropic secara pribadi melobi untuk membatasi model open-weight Tiongkok, meskipun CEO mereka secara publik menolak pembatasan menyeluruh.
- → 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?
Kemajuan Model & Robotika
Bobot Kimi K3 Dirilis. Moonshot AI merilis model MoE dengan 2,8T parameter dan jendela konteks 1M, menyaingi GPT-5.6 Sol. Para kuant komunitas menjalankannya di perangkat keras konsumen kelas atas, meskipun lambat.
- → 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
Pembaruan DeepSeek V4 Flash. Model 304B (13B aktif) kini menyamai GPT-5.6 Luna dengan biaya 60% lebih rendah dan berjalan pada 30+ tok/s di Mac. Pengadopsi awal melaporkan panggilan alat yang rapuh.
- → [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.
Peningkatan Penalaran Opus 5. Claude Opus 5 mencetak 30,2% pada ARC-AGI-3, hampir empat kali lipat dari rekor terbaik sebelumnya. Varian Mythos-nya juga secara otonom memecahkan kandidat kripto HAWK dari NIST.
Inkling Small Berpikir Efisien. Thinking Machines merilis model Apache 2.0 dengan 12B parameter aktif yang mengungguli model yang lebih besar pada tolok ukur pengkodean sambil menggunakan jauh lebih sedikit token.
Tubuh Robot Makin Cerdas. Gemini Robotics 2 dari Google DeepMind menangani gerakan seluruh tubuh dan manipulasi halus, sementara model baru mendukung kolaborasi multi-robot dan penalaran video real-time untuk humanoid.
- → 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
Perkakas & Infrastruktur
MCP 2.0 Mengamankan Agen. Protokol yang diperbarui memperkuat keamanan untuk interaksi agen-perkakas dan mendapat adopsi cepat dari Dropbox dan komunitas, membuat pengembangan agen lokal lebih aman.
Prompt Agen Makin Ramping. Deep Agents v0.7 dari LangChain memangkas token input sebesar 65% menggunakan saran rekayasa konteks, sementara Gateway barunya menerapkan batas pengeluaran dan menyunting PII sebelum panggilan API.
Kuantisasi Mengurangi Kebutuhan VRAM. Inovasi komunitas seperti kuantisasi ternormalisasi varians dari BeeLlama dan ROCmFPX memungkinkan model besar berjalan di perangkat keras yang sederhana, meskipun beberapa pengaturan MoE menghadapi penggunaan memori ekstra.
- → 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
Industri & Investasi
Akuisisi Cursor $60B oleh SpaceX. Akuisisi Anysphere, pembuat editor AI Cursor, termasuk dalam kesepakatan perkakas AI terbesar, karena startup tersebut juga meluncurkan paket diskon untuk India.
Microsoft Menyatukan Copilot. Microsoft berencana membuat 'aplikasi super' tunggal yang menggabungkan fitur obrolan, pengkodean, dan agen, dengan CEO Nadella memperingatkan tentang terkunci pada laboratorium frontier tertentu.
Miliaran Agen Pribadi. CEO Meta melihat miliaran orang menggunakan agen AI untuk keuangan dan kesehatan melalui WhatsApp, dan perusahaan akan menjual layanan AI enterprise dengan basis per-hasil.
Capex AI Mencapai Level Baru. Amazon mengalokasikan $220B untuk 2026, perkiraan Google membuat investor takut, dan NVIDIA menyediakan GPU Vera Rubin untuk SSI milik Ilya Sutskever. UE mendanai gigafactory AI dengan €10B.
- → 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
Itulah ulasan minggu ini - sampai jumpa Minggu depan.