Week 21, 2026 AI roundup

Models Build a Complete Langfuse Observability and Evaluation Pipeline for Tracing, Prompt Management, Scoring, and Experiments MarkTechPost Creative StepFun Releases StepAudio 2.5 Realtime: An End-to-End Voice Model with Roleplay-Specific RLHF and Paralinguistic Comprehension MarkTechPost Security Everyone is navigating AI security in real time — even Google TechCrunch AI Agents Workspace agents in ChatGPT: Admin and builder controls OpenAI Models AI agents are quietly generating chaos engineering failures enterprises don’t track yet VentureBeat AI Dev Tools I tried Amazon’s Bee wearable and am both intrigued and slightly creeped out TechCrunch AI Models ByteDance study finds that asking LMMs questions beats making it transcribe text for long document training The Decoder Models Deepmind's Hassabis sees humanity "in the foothills of the singularity" while LeCun says current AI isn't intelligent The Decoder Dev Tools The Sequence Radar #865: Last Week in AI: Last Week in AI: Karpathy, Google, Colossus, and the Coming IPO Wave The Sequence Models Why you shouldn't leave model selection on default in Copilot, Gemini and other AI tools The Decoder Agents Microsoft Research Releases Webwright: A Terminal-Native Web Agent Framework That Scores 60.1% on Odysseys, Up from Base GPT-5.4’s 33.5% MarkTechPost Models Anthropic may keep supplying Claude to the NSA despite being flagged as a supply chain risk by the Pentagon The Decoder