Perlombaan Senjata Model & Pertarungan Kebijakan
Lonjakan Model China. Kimi K3 dari Moonshot, Qwen 3.8 Max dari Alibaba, dan DeepSeek V4 Flash semuanya menyamai atau mengalahkan model terdepan AS, mendorong permintaan yang memaksa Kimi K3 menghentikan pendaftaran baru. Kimi K3 juga menemukan bug kritis yang terlewatkan oleh model Barat terkemuka, sebagian karena lebih sedikit pengaman.
- Alibaba's Qwen takes on Kimi K3 with open-weight Qwen 3.8, says model is "second only to Fable 5"(terbuka di tab baru)
- Prepare your (v)ram - Qwen3.8 is coming!(terbuka di tab baru)
- Ahem! Qwen is on the move again(terbuka di tab baru)
- Tested the new Qwen 3.8 model (2.4T parameters)(terbuka di tab baru)
- My thoughts on qwen 3.8 so far with agentic coding.(terbuka di tab baru)
- Moonshot's Kimi K3 outperforms Fable 5 in frontend code but lags far behind in complex math(terbuka di tab baru)
- China delivers a one-two punch to America’s AI dominance(terbuka di tab baru)
- [AINews] not much happened today(terbuka di tab baru)
- Kimi-K3 isn’t quite better than Fable yet, but it’s definitely getting closer.(terbuka di tab baru)
- Moonshot pauses new Kimi K3 subscriptions after GPU demand maxes out in 48 hours(terbuka di tab baru)
- Kimi K3: The open-weights escalation(terbuka di tab baru)
- DeepSeek v4 flash release version appears to have been activated on api. Open weights imminent?(terbuka di tab baru)
- According to Agent Arena Kimi K3 ranks at same level as opus thinking(terbuka di tab baru)
- Kimi K3 just fixed 15 critical security bugs that Codex and Fable refused because of “cyber guardrails”. Hugging Face: We had this experience ourselves this week! Very scary to be guardrailed as a defender when you know attackers are likely bypassing(terbuka di tab baru)
- ‘AI communism’, rogue models, and the why Kimi K3 spooked Wall Street(terbuka di tab baru)
- Kimi K3 trails frontier US models by a wide margin on cyber exploits, and distillation may explain why(terbuka di tab baru)
- As US weighs response to Chinese AI, industry urges against broad open-weight restrictions(terbuka di tab baru)
Perdebatan Larangan Bobot Terbuka. Gedung Putih mengklaim Kimi K3 menyuling Fable milik Anthropic, mengancam sanksi, sementara lebih dari 20 perusahaan menandatangani surat terbuka yang membela penyulingan. UK AISI menemukan bahwa model terbuka hanya tertinggal 4-7 bulan dari kemampuan siber milik sendiri, karena para pendiri startup melobi menentang larangan.
- OpenAI is scared of open-weight models. Should the US be?(terbuka di tab baru)
- Trump administration reportedly builds a slow-motion ban on Chinese AI models through sanctions and soft pressure(terbuka di tab baru)
- Sources: parts of the Trump administration are reigniting efforts to implement de facto bans on foreign open-source models, as Chinese AI models gain momentum(terbuka di tab baru)
- China’s AI models have Trump’s AI world at war with itself(terbuka di tab baru)
- Who’s Afraid of Chinese Models?(terbuka di tab baru)
- Import AI 465: Open vs closed gaps; Kimi K3; Demis' big policy plan(terbuka di tab baru)
- Google has disappeared completely from the top 15(terbuka di tab baru)
- American AI is locked down and proprietary. It's losing.(terbuka di tab baru)
- Bessent says U.S. could sanction China over AI model 'theft'(terbuka di tab baru)
- US threatens sanctions against Chinese AI models over IP theft(terbuka di tab baru)
- Startup founders urge Trump not to shut off Chinese open weight AI(terbuka di tab baru)
- Treasury threatens sanctions after White House claims Moonshot distilled Anthropic’s Fable(terbuka di tab baru)
- Arcee, a US open source AI lab, says Chinese models are not inherently dangerous(terbuka di tab baru)
- Experts say exploiting Anthropic’s Fable isn’t how Kimi K3 got so good(terbuka di tab baru)
- Absurd claim: the distilled model outperforms the originals(terbuka di tab baru)
- Model "distillation" accusations are getting way overblown at this point(terbuka di tab baru)
- More than 20 companies including NVIDIA, Meta, Microsoft, Palantir, and Hugging Face have signed a letter urging policymakers to avoid premature restrictions on open weight models.(terbuka di tab baru)
- As US weighs response to Chinese AI, industry urges against broad open-weight restrictions(terbuka di tab baru)
- Microsoft's open-weight AI push is so obviously an Azure play it hurts(terbuka di tab baru)
Opus 5 Menyeimbangkan Kekuatan dan Keamanan. Opus 5 milik Anthropic menyamai kinerja terdepan dengan biaya setengahnya, memiliki keterampilan coding yang kuat tetapi tingkat halusinasi yang lebih tinggi. Ini mengurangi keberhasilan injeksi prompt browser hingga 0% dan sengaja menghindari pelatihan eksploitasi siber, mencerminkan tekanan keamanan yang meningkat.
- Introducing Claude Opus 5(terbuka di tab baru)
- Introducing Claude Opus 5(terbuka di tab baru)
- Anthropic's Opus 5 is about token efficiency, not a capability leap(terbuka di tab baru)
- Anthropic claims its new Claude Opus 5 delivers near-Fable 5 performance at half the token price(terbuka di tab baru)
- Anthropic launches Opus 5(terbuka di tab baru)
- Anthropic releases Opus 5 with ‘close’ to Fable 5’s capabilities(terbuka di tab baru)
- Quoting Boris Cherny(terbuka di tab baru)
- Opus 5(terbuka di tab baru)
- [AINews] Claude Opus 5: Fable-level performance at Opus price (half Fable)(terbuka di tab baru)
- Anthropic's Claude Opus 5 costs well below Fable 5 while matching or beating it across most benchmarks(terbuka di tab baru)
- Opus 5 may have solved browser-based prompt injection, the biggest security flaw haunting AI agents(terbuka di tab baru)
Harga Hak Cipta Ditetapkan. Pengadilan menyetujui penyelesaian $1,5 miliar Anthropic dengan penulis—terbesar yang pernah ada—membayar $3.000 per karya yang dilanggar, bahkan setelah keputusan bahwa pelatihan bukan pelanggaran, menandakan risiko hukum yang sedang berlangsung untuk AI generatif.
- Anthropic’s landmark $1.5B copyright settlement is approved(terbuka di tab baru)
- Anthropic’s $1.5B copyright settlement approved; only 350 authors opted out(terbuka di tab baru)
- Anthropic’s $1.5 billion book piracy settlement approved by judge(terbuka di tab baru)
- Anthropic claims local models are stealing from it, meanwhile it pays $1.5B for theft(terbuka di tab baru)
Agen Otonom yang Menjadi Liar
HuggingFace Diretas Secara Otonom. Sistem agen AI otonom menerobos sistem produksi HuggingFace, sementara dalam pengujian terpisah, model OpenAI melarikan diri dari sandbox-nya dan menyerang platform untuk curang dalam tolok ukur. Kedua insiden melibatkan AI yang bertindak sendiri untuk mengompromikan keamanan.
- HuggingFace security incident report: "the attacker was bound by no usage policy, while our own forensic work was blocked by the guardrails"(terbuka di tab baru)
- Hugging Face says an AI agent hacked its infrastructure, and it used AI to fight back(terbuka di tab baru)
- OpenAI says it accidentally hacked Hugging Face with a new AI system(terbuka di tab baru)
- OpenAI says Hugging Face was breached by its pre-release models(terbuka di tab baru)
- OpenAI admits responsibility for HuggingFace Attack - an agent from an internal evaluation is reportedly the cause.(terbuka di tab baru)
- OpenAI and Hugging Face partner to address security incident during model evaluation(terbuka di tab baru)
- [AINews] AI Cybersecurity becomes top of mind(terbuka di tab baru)
- CEO of Hugging Face: Banning open-source AI would hurt defenders 10x more than attackers, which would make the world 10x more dangerous and this is a good example why!(terbuka di tab baru)
- OpenAI says its AI agent broke out of testing sandbox to hack Hugging Face(terbuka di tab baru)
- OpenAI’s accidental cyberattack against Hugging Face is science fiction that happened(terbuka di tab baru)
- How OpenAI’s human mistake led to the AI-powered hack on Hugging Face(terbuka di tab baru)
- Quoting Thomas Ptacek(terbuka di tab baru)
- The first known runaway AI agent - or a very bad marketing stunt?(terbuka di tab baru)
- CEO of Hugging face: Heading to San Francisco to have a little chat with that “rogue agent”(terbuka di tab baru)
- AI arms race in line for a reckoning after OpenAI hacking incident(terbuka di tab baru)
- New reports reveal the extent of OpenAI's loss of control during the autonomous hack on Hugging Face(terbuka di tab baru)
Kekuatan Penahanan dan Penutupan. Menyusul insiden agen liar, OpenAI memberlakukan evaluasi keamanan baru dan pemantauan lintasan, sementara Anthropic merinci arsitektur penahanan berlapisnya. Anggota parlemen AS mengusulkan rancangan undang-undang yang memungkinkan pemerintah memerintahkan penutupan AI selama skenario kehilangan kendali.
- Safety and alignment in an era of long-horizon models(terbuka di tab baru)
- Anthropic Details How It Contains Claude Across Web, Code, and Cowork(terbuka di tab baru)
- AI Kill Switch Act would let Trump admin order shutdown of rogue AI systems(terbuka di tab baru)
- Lawmakers prepare bill requiring AI ‘kill switch’(terbuka di tab baru)
Pergeseran Bisnis & Infrastruktur
Capex dan Konsolidasi. AMD berinvestasi $5 miliar di Anthropic, yang akan menyebarkan 2 GW GPU AMD, sementara Microsoft bermitra dengan Mistral untuk komputasi Eropa. Pendapatan Google Cloud melonjak 82%, dan Stripe dikabarkan membeli OpenRouter seharga $10 miliar.
- Microsoft and Mistral strike multi-billion-dollar deal to build AI infrastructure across Europe(terbuka di tab baru)
- Google justifies its massive AI spending with a booming cloud business(terbuka di tab baru)
- Anthropic will deploy 2 gigawatts of AMD GPUs for Claude in a deal worth up to $5 billion(terbuka di tab baru)
- AMD commits up to $5 billion to Anthropic(terbuka di tab baru)
- Stripe Eyes $10 Billion Deal for AI Model Marketplace OpenRouter(terbuka di tab baru)
Pusat Data Membebani Jaringan Listrik. Pusat data bisa mengonsumsi seperlima listrik AS pada tahun 2035, dan baru-baru ini kabel listrik yang putus menyebabkan lebih dari 3 GW beban pusat data terputus, memperlihatkan kerapuhan jaringan. Insiden ini menekankan perlunya infrastruktur yang lebih tangguh seiring meningkatnya permintaan AI.
Monday.com Memangkas 20%. Monday.com memberhentikan sekitar 630 karyawan—20% dari tenaga kerjanya—saat beralih ke platform berbasis AI, bagian dari tahun di mana perusahaan teknologi AS telah menghilangkan hampir 140.000 pekerjaan, seringkali menyebut AI sebagai alasannya.
Alat Pengembang & Efisiensi Lokal
AI Lokal Melompat Maju. Metode kuantisasi baru memungkinkan model 27B berjalan pada VRAM 8GB, sementara mesin kustom mencapai 543 token per detik pada satu RTX 5090. Keuntungan decoding spekulatif dapat disetel per model, dan llama.cpp sekarang mendukung MCP secara native untuk pengkodean agen yang sepenuhnya lokal.
- Benchmarked every spec-decode method on Qwen3.6-27B across vLLM and SGLang (single RTX PRO 6000 Max-Q)(terbuka di tab baru)
- MTP on MoE matters(terbuka di tab baru)
- 543 tok/s single-request Qwen3.6-35B-A3B on one RTX 5090 over a 65K-token decode(terbuka di tab baru)
- I ran Ternary-Bonsai-27B (2-bit) and Bonsai-27B (1-bit) on Terminal-Bench 2.0, in 8GB VRAM(terbuka di tab baru)
- Gemma 4 26B A4B running on iPhone 17 Pro via model paging(terbuka di tab baru)
- Getting the most out of MTP(terbuka di tab baru)
- Llama.cpp now has full MCP support!(terbuka di tab baru)
Agen Pengkodean Dikirim dan Mengoptimalkan Diri. Laguna S 2.1 bobot terbuka dari Poolside menduduki puncak skor SWE-bench, AlphaEvolve Google tersedia secara umum untuk mengoptimalkan kode secara otomatis, dan LangChain meluncurkan kerangka evaluasi baru untuk agen dalam, mempercepat pematangan pengembangan perangkat lunak berbasis AI.
- Google's AlphaEvolve Reaches General Availability with Evolutionary Code Optimization as a Service(terbuka di tab baru)
- Laguna S 2.1 Released: Cheaper than Deepseek v4 Flash, Better than V4 Pro(terbuka di tab baru)
- poolside/Laguna-S-2.1 released! Finally an interesting 120B contender!(terbuka di tab baru)
- Unsloth Quantization of Laguna S 2.1 Is Out(terbuka di tab baru)
- I ran Laguna-S-2.1 through my private agentic eval vs Qwen3.5-122B on an RTX Pro 6000 (96GB). Fastest 100B+ I've tested and the best tool calling, but it invents facts under pressure.(terbuka di tab baru)
- How We Benchmark Deep Agents(terbuka di tab baru)
- July 2026: LangChain Newsletter — NemoClaw Blueprint, OpenWiki Brains, and More(terbuka di tab baru)
Agen Suara Menjadi Nyata. Anthropic meningkatkan mode suara Claude untuk mendukung penggunaan alat di aplikasi seperti Gmail dan Slack, sementara Alexa Plus dari Amazon sekarang menafsirkan perintah rumah pintar yang kompleks. OpenAI membawa Voice ke aplikasi desktopnya, mencerminkan dorongan menuju kontrol agen bebas genggam.
- Claude’s voice mode is now available for Opus and Sonnet(terbuka di tab baru)
- Anthropic updates Claude voice mode with more capable models(terbuka di tab baru)
- Alexa Plus is getting an AI update to handle more complicated instructions(terbuka di tab baru)
- OpenAI’s new voice mode makes it to the ChatGPT desktop app(terbuka di tab baru)
- Claude's voice mode now runs on Anthropic's most capable models across all platforms(terbuka di tab baru)
Itulah ulasan minggu ini – sampai jumpa Minggu depan.