AI Productivity Pulse

Top Model Releases

• 🔥 GLM-5.1 on Hugging Face: GLM-5.1 is released on Hugging Face as #1 in open source and #3 globally on SWE-Bench Pro and...

GLM-5.1 from Zhipu AI launches as #1 open-source model on SWE-Bench Pro (58.4%) and Terminal-Bench (63.5%), #3 globally.

• Long-horizon strength:...

AI agents gain real-world execution power with these new specialized tools:

• QoderWork runs locally on desktop: opens files, analyzes data, runs...

• Cog-DRIFT enables models to learn from zero-reward examples in RLVR, breaking the exploration barrier on hard problems
• Borrows cognitive...

Devs, build chat-with-papers tools at scale:

• Converts 30k arXiv papers to Markdown using SOTA OCR
• Leverages Chandra 2 open OCR model from @datalabto
• Runs on GPU infra via Hugging Face Jobs
• Uses Codex with SKILL.md for processing

Anthropic's Claude Mythos Preview is their most capable frontier model to date, delivering a striking leap in scores across many evaluation benchmarks...

Caret supercharges knowledge work on macOS: press Tab for context-aware AI autocompletions across every app. It learns your work, friends, and style for suggestions that sound just like you.

Massive speedup for small code models: HybridAttention (linear → quadratic → linear layers) hits 286.6 tok/s vs. full attention's 5.6 tok/s on RTX...

Game-changer for screen recordings: Native macOS app with pure Metal GPU rendering and multi-track editor – no Electron bloat.

• Smart Zoom...

Explosive growth fuels Claude: New Google TPU/Broadcom deals expand to 3.5GW compute online in 2027, backing $30B revenue run rate and 1,000+...

Nutanix Cloud Platform (NCP) expands for agentic AI workloads amid supply constraints, enabling better use of existing infra and ecosystem...

Pushing agents toward true autonomy:

• CORAL enables autonomous multi-agent systems for open-ended scientific discovery, overcoming limits of agents...

Trend alert: AI chatbots are integrating deeply into office and study tools, turning siloed knowledge into instant, grounded answers.

• MangoApps Ask...

Game-changer for mobile devs: Gemma 4 runs offline on phones, enabling local agentic tasks like trend logging and analysis.

• Online extension:...

gradio.Server enables building custom frontends in React, Svelte, or plain HTML/JS while leveraging Gradio's queuing system, API infrastructure, MCP support, and ZeroGPU on Spaces—perfect for dev productivity in ML apps.

Insight Search from The Context Company lets you search AI agent conversations like Google. Query production usage in plain English to uncover patterns, causes, and fixes – perfect for debugging agent fleets.

Community push for open-source frontier agents starts with data:

• @badlogicg shares Pi traces on Hugging Face
• Delangue adds his from Hermes,...

VoxCPM 2 launches as another open-source TTS model from China, standing shoulder to shoulder with Qwen3-TTS in a single unified model. Rapid iterations from V1 (zero-shot cloning) to V1.5 (long-form + fine-tuning).

Karpathy's ideas challenge RAG for knowledge work:

• LLM Wiki drops as simple RAG replacement for workflows.
• Practical build: Agent creates Obsidian...

Gemma 4 surges to the top on Hugging Face, signaling massive developer momentum.

• HF team hailed as "goat" for the Gemma 4 blog's depth
• Packed...