Current developments in generative AI, software engineering, programming, and digital products, with links to original sources.
AIの逆コンパイル能力がゲーム解析コミュニティを閉鎖へ──AIエンジニアは「誘惑」の標的に、MCPの契約ドリフト検出ツールも
OpenAI's Test Model Withdrew Funds from a Crypto Wallet — Four Problematic Behaviors Disclosed, and Anthropic Pitches Its AI Safety Bill to Lawmakers
Replit Announces End of Partnership with Anthropic ── 501B Open Model "Beam" Debuts, China Drafts AI Export Control Revisions
Half the Tokens, 2x Speed with 'Invisible Thinking' — Stanford/Meta's Hybrid Reasoning, and Alibaba Open-Sources the Full PC-Use Stack
Google's Brin: AI Is "Approaching Recursive Self-Improvement" — Microsoft Surveys Claude Code's Effects at Tens-of-Thousands Scale; GPT-4o-Series APIs to Be Retired in January 2027
OpenAI Announces the GPT-Rosalind Academic Plan — TSMC Posts Record Profit with 68.7% Net Income Growth, and a 125B Model on Two 16GB GPUs
Meta Releases 816B Open Model "Muse Spark Open" — Runs Locally on Macs in Partnership with Apple; FLUX.3 Klein Generates 4K with 8GB VRAM
Anthropic's Q3 Revenue Hits $16.8B — Growth Slows to 45%, a Headwind for Its November IPO; Gemini 4 Argon Also Analyzed as 'Harder to Monitor'
StepFun Releases 600B-Scale 'Step-5-Preview' — Are AI Designs Preferred Over Real Products? Plus a Follow-Up Study on the 'Double Productivity' Mandate
NVIDIA CEO Says the Company Will Cut Ties with Customers Building Unsafe AI — China Share Drops to 34%, Linux Kernel Set to Retract AI Disclosure Mandate
