Fresh stories
Google introduces Gemini Intelligence on Android with browser use, AppFunctions, and Rambler
Google unveiled Gemini Intelligence at the Android Show with cross-app task automation, Gemini in Chrome, Rambler voice cleanup, custom widgets, and AppFunctions. The rollout moves Gemini into core Android workflows on Pixel and Galaxy devices this summer.
agent-tui 0.2.0 adds markdown rendering, tool approvals, and local Gemma 4 support
agent-tui shipped v0.2.0 with markdown rendering, tool approvals, configurable reasoning views, and an AI SDK-only dependency chain. The demo also showed Gemma 4 31B running locally, so the terminal UI now covers hosted and on-device models.
Sentence Transformers 5.5.0 adds train-sentence-transformers skill with one-shot 0.8856 NDCG@10
Sentence Transformers 5.5.0 adds an agent skill for fine-tuning embeddings, rerankers, and sparse encoders from Claude Code, Codex, Cursor, and Gemini CLI. The author reports a one-shot German embedding run rising from 0.6720 to 0.8856 NDCG@10 on a local PC.
Researchers report Mini Shai-Hulud hits OpenSearch, Guardrails, and RubyGems after TanStack
Researchers tied Mini Shai-Hulud to OpenSearch, Guardrails, and a RubyGems incident after TanStack's npm postmortem. Track registry controls, CI cache hardening, dependency policy, and secret handling before the next package hit.
Google introduces Gemini Intelligence on Android with browser use, AppFunctions, and Rambler
Google unveiled Gemini Intelligence at the Android Show with cross-app task automation, Gemini in Chrome, Rambler voice cleanup, custom widgets, and AppFunctions. The rollout moves Gemini into core Android workflows on Pixel and Galaxy devices this summer.
Perceptron releases Mk1 with 2 FPS video reasoning, 32K context, and $0.15 per 1M input
Perceptron launched Mk1, a multimodal model for video and embodied reasoning with native 2 FPS video, 32K context, and structured spatial outputs. OpenRouter access and the low input price make it usable for deployment, not just demos.
Claude Opus 4.7 opens fast mode with ~2.5x speed as Cursor, v0, Droid, and OpenRouter add support
Anthropic rolled fast mode for Opus 4.7 into Claude Code and tools including Cursor, v0, Droid, Conductor, and OpenRouter. Use it where latency matters, but watch pricing: Cursor disclosed a 6x multiplier and others treat it as premium.
agent-tui 0.2.0 adds markdown rendering, tool approvals, and local Gemma 4 support
OpenAI Codex supports background computer use with Mac app control and Telegram BotFather setup
Claude Code 2.1.140 fixes /goal hangs and adds case-insensitive subagent matching
Sentence Transformers 5.5.0 adds train-sentence-transformers skill with one-shot 0.8856 NDCG@10
SophontAI releases Medmarks v1.0 with 30 medical benchmarks and 61 models
Diffusers 0.38.0 adds Ace-Step 1.5 pipelines and Flash Attention 4 support
Perplexity benchmarks Qwen3 235B on GB200 NVL72: NVLS latency drops from 586 µs to 313 µs
Google DeepMind tests Gemini pointer demos in AI Studio with PDF bullets and recipe doubling
Top storiesthis week
Thinking Machines introduces interaction models with 200 ms full-duplex audio, video, and tool use
Thinking Machines previewed interaction models that process audio, video, and text in 200 ms micro-turns, letting the system listen, speak, and react at the same time. The demos matter because the interaction loop is trained into the model instead of stitched together from separate speech and tool layers.
TanStack reports npm supply-chain attack across 42 packages with credential-stealing payload
TanStack disclosed a supply-chain attack that pushed two malicious npm versions across 42 packages in a 10-minute window. The payload targeted cloud keys, GitHub tokens, npm credentials, and SSH material, so teams should audit installs and rotate secrets.
OpenAI launches Daybreak with GPT-5.5-Cyber, Codex workflows, and repo scanning
OpenAI launched Daybreak, combining GPT-5.5, Codex workflows, repo scanning, threat modeling, and patch generation for cyber-defense teams. It packages frontier models into a continuous secure-software workflow, so teams can test whether it fits their response pipeline.
Developers launch Agent FM, Mate, and ntm for multi-session Claude Code and Codex control
Independent developers shipped new control-plane tools for long-running coding agents, including Agent FM audio monitoring, Mate phone-first remote control, and ntm for provider-agnostic multi-agent workflows. It matters because teams running many Claude Code and Codex sessions still need better visibility, handoff, and checkpointing than a single built-in session list provides.
OpenAI launches Deployment Company with $4B backing and 150 forward-deployed engineers
OpenAI launched the OpenAI Deployment Company and tied it to Tomoro’s acquisition, giving the unit 150 forward-deployed engineers and $4 billion in initial backing from 19 partners. It matters because OpenAI is packaging services, deployment help, and organizational integration as part of the product stack instead of leaving enterprise rollout to outside consultancies.