01
AI·DesktopCommanderMCP is an MCP server for Claude that provides terminal control, file system search, and diff-based file editing capabilities.
AI Pulse · Hourly Refresh
The daily heat-map of what actually matters in AI.
Aggregating arXiv, OpenAI, Anthropic, HuggingFace, GitHub, and dozens of Chinese & English sources — auto-summarised, clustered, and importance-ranked by Claude.
- Live sources
- 20
- Articles · 24h
- 93
- On the board
- 50
- Last fetch
- 01:07 UTC
02
AI·Stitch-skills is a library of agent skills that follow the Agent Skills open standard for compatibility with coding agents such as Antigravity, Gemini CLI, Claude Code, and Cursor.
03
AI·Superpowers is an agentic skills framework and software development methodology that works.
04
AI·Terraform enables safe and predictable creation, change, and improvement of infrastructure through declarative configuration files shared as code.
05
AI·Claude-cookbooks is a collection of notebooks and recipes demonstrating fun and effective ways to use Claude.
06
AI·Claude-code-templates is a CLI tool for configuring and monitoring Claude Code.
07
AI·The blog post argues against constantly telling users to ask an LLM for help.
08
AI·Ant is a JavaScript runtime and ecosystem that includes a runtime with its own JavaScript engine, package manager, ants.land registry, deployment platform, and Ant Desktop for native desktop apps.
09
AI·Next-ai-draw-io is a Next.js web application that integrates AI capabilities with draw.io diagrams to allow creation, modification, and enhancement of diagrams via natural language commands and AI-assisted visualization.
10
AI·Article analyzes neoclouds CoreWeave and Nebius' massive GPU builds funded by circular financing: Nvidia's $4B equity stakes plus a $6.3B GPU purchase backstop, offset by $122B+ hyperscaler leases (Microsoft/Meta) that shift capex to opex and enable rapid Blackwell/Rubin deployment.
11
AI·SQLite strict tables enforce column types to prevent datatype errors like inserting text into integer columns, allowing only standard types or ANY for flexibility, but cannot be added to existing tables and require SQLite 3.37+.
12
AI·A female US rower has completed a historic solo journey from California to Hawaii.
13
AI·ClickHouse scaled PgBouncer throughput 4x by running a fleet of processes sized to cores, using so_reuseport to balance connections on one port and inter-process peering for query cancellation.
14
AI·This repository contains a curated collection of Codex plugin examples for OpenAI, each including a required plugin.json manifest and optional supporting files for skills, agents, and marketplaces.
15
AI·Site offers 80+ structured courses for building real systems including Redis, Git, databases, compilers, and kernels from scratch in Python, Go, Rust, and others.
16
AI·GitHub trending page for the Ansible repository describes Ansible as a simple IT automation platform for deploying and maintaining applications and systems.
17
AI·OpenAI integrates Codex capabilities into its new ChatGPT Work agent for general office tasks, rebrands models as Sol/Terra/Luna tiers, and makes GPT-5.6 the default for Microsoft 365 Copilot.
18
AI·Zig founder Andrew Kelley angrily criticized Jarred Sumner and Bun's 11-day rewrite of its 1 million lines of code from Zig to Rust using Claude Fable 5 and Claude Code; the move succeeded with 16.5K USD in API costs but left 27K lines of unsafe code and sparked debates on AI-generated code quality and community ethics.
19
AI·Apple filed a federal lawsuit on July 10, 2026, accusing OpenAI of trade-secret theft by poaching Apple engineers and exploiting a post-departure access bug; the suit targets OpenAI's hardware ambitions and a possible AI-powered iPhone rival.
20
AI·OpenAI released GPT-5.6 in three tiers—Sol (flagship), Terra, and Luna—boosting benchmarks in Agent's Last Exam (56.3 score), math, coding, security, and research while slashing token costs and adding Max/Ultra reasoning modes.
21
AI·OpenAI's GPT-5.6 limited preview, reportedly requiring White House approval, was abruptly released without permission after 10 days, with the government clarifying no approval is needed and labeling the cooperation 'voluntary'.
22
AI·A University of California San Diego team used two teleoperated Unitree G1 humanoid robots in preclinical trials to perform laparoscopic cholecystectomies on live pigs: one with a human surgeon as assistant, and one with two robots collaborating, marking the first use of general-purpose humanoids in live surgery.
23
AI·Open-source models attract more traffic and earn most of the revenue, causing Anthropic's revenue to drop from dominance to a small fraction of the market.
24
AI·Anthropic's J-space analysis of Claude reveals a small subset of internal representations forming a functional global workspace for flexible reasoning, but the paper explicitly states this has no bearing on AI consciousness and clarifies the core of AI alignment research.
25
AI·OpenAI folds Codex into the new ChatGPT desktop app as Work mode, shifting focus to usage-based billing and water-meter-style pricing for AI agents, effectively 'killing' the old standalone Codex product.
26
AI·Unitree G1 humanoid robot completed live cholecystectomy on two pigs using standard laparoscopic tools under remote doctor control, with the lead author being a 00-after PhD from UCSD; results published in Nature.
27
AI·Alibaba's Lingbo released six embodied AI models from July 7-10, 2026, introducing 'Embodied Native' — a new paradigm that trains foundation models from scratch on physical-world data, interactions, and causal physics rather than migrating internet LLMs, with LingBot-VA 2.0 as the flagship VLA model achieving real-time 150Hz inference and strong generalization.
28
AI·Zhipu AI founder Tang Jie releases internal letter after 'GLM Moment' success, announcing focus on long-horizon tasks, autonomous agents, and self-evolving AI beyond coding and reasoning.
29
AI·Reports of Meta selling excess AI compute and Anthropic signing massive data-center deals prove the notion of AI compute oversupply is a myth; actual deliverable frontier capacity remains scarce.
30
AI·Meta plans to mass-produce its own Iris AI chip by September, DeepSeek and Zhipu AI are developing custom inference chips, and OpenAI unveiled Jalapeño, signaling a shift from NVIDIA GPU reliance to full-stack model + chip + cloud strategies.
31
AI·2025 saw multiple AI companionship app closures including Woebot (150M users, clinical tool shut down due to unsustainable economics) and Dot AI (vision divergence, funds exhausted), with global closures in at least five products and domestic in eight; survivors like Replika (post-data deletion), Character.AI (Google talent acquisition), and China's Xin Bing (ice) lost users amid regulatory scrutiny and high inference costs, revealing unsustainable business models lacking self-reinforcing revenue.
32
AI·Power semiconductor manufacturers are rapidly expanding capacity amid high demand from EVs, AI data centers, and power grids, while low-end supply exceeds demand.
33
AI·Upstart PhD founder of Eaglewing Zhiwei completes third funding round in three months for a bio-inspired flapping-wing robot that learns to understand and drive fluid dynamics via self-developed simulation engine.
34
AI·Nubia is set to launch the second-generation 'Doubao Phone' at WAIC 2026 as a mass-market flagship with improved AI agent OS, memory, and execution capabilities.
35
AI·SpaceX filed FCC for 100,000 Gen 3 Starlink satellites with 10x bandwidth upgrades, explicitly linking the constellation to AI workloads, data centers, autonomous vehicles, and humanoid robots for global compute connectivity.
36
AI·AI is tested in real disaster simulations like fires, evacuations, and crowds to assess decision-making, collaboration, and survival ability rather than human-like appearance.
37
AI·Momenta listed on HKEX and soared over 6% on day one, securing global capital, while Mushroom Car Network's ToG Robobus model lags due to heavy hardware investment and cashflow pressure despite earlier physical AI claims.
38
AI·Quantum Computing China has completed its first 100,000-card cluster, achieving national full-stack computing power support for the ten-thousand-card era and successfully running over 300 applications.
39
AI·OpenAI attributes a complete proof of the 50-year-old Cycle Double Cover Conjecture (every bridgeless graph has a cycle double cover) to GPT-5.6 Sol Ultra, which completed the 3-page PDF in under an hour using 64 parallel sub-agents guided by a 700-character public prompt.
40
AI·1X Technologies released an update for NEO robot's 25-DOF hand with tendon-driven actuators, high-resolution tactile skin, and IP68 safety rating, enabling complex tasks like bulb tightening and grape picking.
41
AI·MaaS shifts Chinese cloud providers from CPU to Token/GPU metrics, with token consumption exploding 16x at Volcano Engine and 8x client growth at Alibaba Cloud, but overall public cloud growth slowed to 8-11%.
42
AI·YouTube CEO Neil Mohan seeks to balance AI-generated content flood with human creativity safeguards, including labeling AI videos, while launching virtual avatar tools and noting 17.5% of shorts may be AI-generated.
43
AI·Cloudwise, marketed as Hong Kong's first AGI stock, saw a 41% single-day plunge on unlock day; frequent equity raises and project-based revenue reveal ongoing losses, long receivables, and lack of self-sustaining API/token model.
44
AI·LLM-as-judge scores fluctuate across evaluators even with fixed candidate responses, treating the issue as measurement validity. Across four datasets, scaling Qwen3 from 1.7B to 32B parameters yields only limited adjacent gains, while MiniMax M2-to-M2.7 API upgrades show none; stronger judges reduce but do not eliminate position and verbosity biases, with repeated juries offering little benefit under correlated errors.
45
AI·SK Hynix filed to list ADS on Nasdaq, raising $2.65 billion for Korean expansion while cutting the 'Korea discount' through improved disclosures and global investor access. The move aims to reprice its storage assets ahead of AI-driven demand.
46
AI·Viral stories of AI-crafted mRNA vaccines saving pets or spotting cancers mirror the Theranos blood-testing fraud, with exaggerated claims by non-experts creating hype while AI merely integrates existing tools rather than innovating breakthroughs.
47
AI·This position paper reviews LLM-driven formal theorem provers and argues that current systems function mainly as solvers for well-defined problems, not as research agents capable of discovering new theorems or resolving open conjectures at the frontier. It identifies key limitations in datasets, exploration, tools, and collaboration and proposes a roadmap for AI4Math systems to support genuine mathematical research.
48
AI·Hidden Decoding expands each token into n independent streams during continued pretraining of a fixed Transformer backbone, using separate embedding tables and retaining intermediate KV caches for latent computation; Stream-Factorized Attention limits cross-stream mixing to reduce costs. Experiments show frontier 80B and 617B MoE models improve on all benchmarks over matched baselines.
49
AI·LLM-based ASR in regulated domains like banking is limited by privacy and real-speech collection costs; synthetic TTS data is a cost-effective substitute, but acoustic mismatch hinders supervised fine-tuning (SFT). Group Relative Policy Optimization (GRPO) applied solely to synthetic speech reduces WER by 40% relative to SFT (36.71% to 22.09%) and by 45% in the SFT-then-GRPO sequence, by improving behavioral calibration and audio attention rather than representations.
50
AI·Across five instruction-tuned LLMs, ordinary answer-confidence separates correct from wrong answers on answerable questions but fails to distinguish unanswerable questions (e.g., false-premise ones in CREPE), while a linear probe on hidden states performs the opposite, revealing two distinct abstention axes that remain separable even at 14B scale.