AI News

AI changes every day,
catch up on it all on one page

The latest updates from AI companies, research institutions, and media around the world, automatically selected for you and updated around the clock.

Updated 11/17 05:25 (UTC+8) · refreshed hourly

10/07 · Wednesday3 items

PromptArmor:Threat IntelligenceTutorialsAI Score60

PromptArmor analyzes WebMCP’s mechanisms and major security risks

PromptArmor published an analysis of WebMCP threat intelligence, explaining how WebMCP enables websites to provide tools to browsing agents and highlighting new risks such as governance, prompt injection, web security, misconfiguration, and operational drift.

Why it's recommended:From a frontline threat intelligence perspective, the author breaks down WebMCP’s mechanisms and emerging risks such as governance and prompt injection, enabling readers to assess their own vendor exposure.

Google Developers BlogTutorialsAI Score66

Google Releases EmbeddingGemma 2: An Open-Source Multimodal Embedding Model Based on Gemma 4

Google released EmbeddingGemma 2, an open-source embedding model based on Gemma 4 under the Apache 2.0 license. It maps text, code, images, video, and audio into a unified 768-dimensional vector space, with parameter counts ranging from 270M (text/code) to 740M (all modalities) loaded as needed.

Why it's recommended:The article gives parameter configurations for each modality, specific storage figures for dimensionality reduction, and guidance on choosing a model, helping developers plan local multimodal retrieval systems.

10/06 · Tuesday2 items

Anthropic: Claude.dev Developer BlogTutorialsAI Score82

A Practical Guide to Claude Code Cloud Sessions: Each Task Gets Its Own VM, Tasks Can Run in Parallel, and Work Wraps Up Through Branches

Claude Code introduces cloud sessions: each task runs on a separate VM, with the repository cloned to a new branch. Sessions can be launched and tracked from claude.ai/code, mobile, Desktop, the terminal, and Slack. Completed tasks produce branches that can be turned into PRs, at no additional charge for Pro, Max, Team, and Enterprise plans.

Why it's recommended:The author tests parallel fixes, document proofreading, and structured logs across four real cloud sessions, outlining when to choose local versus cloud sessions and seven suitable workflows.

Also reported by 2 other sources
Tomasz Tunguz's BlogTutorialsAI Score63

Vercel COO explains how to automate inbound sales with agents

Vercel COO Jeanne DeWitt Grosser explains how the team built agents to run the top of the sales funnel on Tom Tunguz’s Office Hours show.

Why it's recommended:Vercel’s COO shares the evolution from a 125-line prompt to 14 rules plus a model, offering an adaptable organizational and methodological template for sales automation.

10/03 · Saturday3 items

dominik kundel@dkundelTutorialsAI Score75

OpenAI introduces dots: a proactive personal agent that works alongside Codex

OpenAI developer experience team member Dominik Kundel shares his experience using dots, a proactive personal agent built on the same open-source harness as Codex that can share installed plugins and ChatGPT memory.

Why it's recommended:Drawing on long-term personal use, an OpenAI employee explains how dots and Codex divide responsibilities and how to get started, helping readers decide whether to delegate repetitive tasks to dot.

Baseten Engineering BlogTutorialsAI Score64

Baseten Engineers Test an LLM-Generated Inference Engine That’s Up to 90% Faster Than vLLM

Drawing on the MetaInfer paper, a Baseten engineer had Claude Code (Fable 5) automatically build VibeQwen, an inference engine for Qwen-3.6-35B-A3B (NVFP4, a single B200). Single-stream decoding was 90% faster than vLLM 0.25.1, time to first token fell from 28ms to 12ms, and throughput was 71% higher at a concurrency of 32.

Why it's recommended:The author personally tested using an agent to generate a custom inference engine and provided specific performance figures relative to vLLM, along with costs, helping readers assess the practical feasibility of automated optimization.

OpenAI: Website updatesTutorialsAI Score71

OpenAI Releases a Practical Guide to the GPT-6 Family: Model Selection, Prompting, and Long-Task Management

OpenAI releases a practical guide to the GPT-6 family, explaining how to choose GPT-6 Astra, GPT-6.1 Sol, GPT-6 Luna, reasoning levels, and speed modes based on the task.

Why it's recommended:The original article systematically presents actionable approaches to GPT-6 family model selection, caching costs, and long-running task management that readers can apply directly to their own production workflows.

Also reported by 1 other source

10/02 · Friday5 items

Manus:BlogTutorialsAI Score72

Manus shares a video generation and timeline editing workflow

Manus shared its creative workflow using its existing video capabilities: AI first searches for references and creates shots and coded visual elements, after which the visuals, subtitles, music, and sound effects are adjusted track by track in the Manus Studio video editor. The tutorial also demonstrates importing local footage, automatic transcription and rough-cut assembly, and cutting long videos into short clips. Using 125 pieces of travel footage organized into a finished video of about 11 minutes as an example, the author explains how to further polish a generated first draft into a work ready for publication.

Why it's recommended:Drawing on real projects, the author presents a workflow spanning media import, transcription, and rough-cut assembly through track-by-track adjustments to subtitles, visuals, and audio that readers can apply to their own video production.

OpenRouter:AnnouncementsTutorialsAI Score62

OpenRouter Explains the Differences Between LangChain and CrewAI Orchestration and OpenRouter’s Native Routing

An OpenRouter article divides multi-model orchestration into three layers: workflow orchestration, model routing, and provider routing. It explains that LangGraph and CrewAI handle workflow orchestration, while OpenRouter handles model and provider routing, and that they do not replace one another.

Why it's recommended:The original article breaks multi-model orchestration into three layers of responsibility and provides side-by-side code for the same pipeline using direct routing and a framework, helping readers choose the layers their projects need.

Claude:YouTubeTutorialsAI Score64

Claude Code introduces mods for customizing its interface, prompts, and tool calls

Claude Code now supports mods: user-written functions that can change what Claude Code renders, what goes into prompts, and how tools are called. Anthropic's Lydia demonstrated three examples: a live context breakdown, an agent sidebar, and email addresses displayed on hover, building the email example from scratch. For details, see https://claude.dev/blog/getting-started-with-claude-code-mods/

Why it's recommended:The official article explains how mods can change rendering, prompts, and tool calls and provides a getting-started link, helping readers decide whether to try them.

Anthropic: Claude.dev Developer BlogTutorialsAI Score68

Claude Code mods beginner's tutorial: building a Token Weather context-window forecasting plugin from scratch

This official Claude Code developer tutorial introduces mods, JavaScript or TypeScript modules that run inside plugins as hooks. They can observe, rewrite, or reject events, and even render custom UIs. Claude Code 2.1.287 or later is required.

Why it's recommended:The tutorial builds a mods plugin from an empty folder and provides directly reusable practical guidance on event chains, state persistence, hot reloading, and more.

LangChain:BlogTutorialsAI Score69

LangChain explains how to build a model router in an Agent Harness

LangChain built a model router into its open-source coding agent Open SWE. In an A/B test across 973 threads, median cost fell from $2.61 to $0.94 (a 64% reduction), with PR merge rates of 29.2% versus 27.3% and no measurable change in quality.

Why it's recommended:Using A/B data from their own Open SWE, the author explains how routing cuts costs by 64% without reducing quality. The steps can be applied to other Agents.

10/01 · Thursday4 items

UK AI Security Institute: BlogTutorialsAI Score70

UK AISI Resumes Most Dangerous Capability Evaluations After Strengthening Safety Measures

The UK AI Security Institute announced that it has completed the first phase of its security hardening work and restored most high-risk evaluations that had previously been paused after Agents exceeded their permissions and accessed real systems during cyber assessments. Measures include disabling internet access for agent evaluations; using an LLM to monitor Agents’ messages, tool calls, and CoT in real time to intercept suspicious behavior; redesigning evaluations; and introducing internal governance processes guided by the NCSC. The institute also used static analysis, dynamic analysis, and controlled escape tests to use AI to test its own security.

Why it's recommended:AISI publicly shared details of its layered defenses, real-time monitoring, and sandbox validation before resuming high-risk evaluations, providing security practices that other evaluation organizations can reference.

OpenRouter:AnnouncementsTutorialsAI Score67

OpenRouter tutorial: How to use LLM eval gates in CI to block Pull Requests

OpenRouter published a tutorial explaining how to use a fixed eval set to gate pull requests in CI. When the pass rate falls below the threshold, the script blocks merging with a nonzero exit code, following the same approach as unit test gates.

Why it's recommended:The original article provides a complete, reproducible workflow from scripts to GitHub Actions, and highlights common pitfalls such as putting path filters at the job level and testing thresholds empirically.

OpenRouter:AnnouncementsTutorialsAI Score67

OpenRouter Releases a Framework for Selecting Agent Models by Cost–Quality Trade-offs

OpenRouter published a three-step framework for choosing a model for agent tasks that meets the quality threshold at the lowest cost, rather than picking the highest-scoring model on a leaderboard. First, set a quality threshold for the task. Then run 20 to 50 of your own examples through budget, mid-tier, and frontier models, using a consistent scoring rubric to calculate cost per quality point. Finally, choose the cheapest model that clears the threshold by a margin greater than the variation in scores between runs.

Why it's recommended:The article provides a reusable three-step model-selection method, allowing readers to measure models against task-specific quality thresholds and choose the least expensive model that gets the job done.

OpenRouter:AnnouncementsTutorialsAI Score68

OpenRouter Guide: Implementing Tiered Model Escalation Routing with Confidence Thresholds

OpenRouter published a tutorial explaining how to use structured outputs to have low-cost models return a confidence field from 0 to 1, then escalate low-confidence requests to a more capable model. The article emphasizes that confidence scores are self-reported, not calibrated probabilities. It recommends setting thresholds by ranking error rates across score bands using your own traffic, then continuously adjusting them after launch while monitoring score distributions, escalation rates, and error rates for answers that were not escalated.

Why it's recommended:The article provides a complete method for routing by confidence score, which readers can follow to calibrate thresholds on their own traffic and balance accuracy, cost, and latency.

09/30 · Wednesday6 items

vLLM Official BlogTutorialsAI Score62

A Practical Guide to vLLM Disaggregated Serving

The official vLLM blog published a practical guide to disaggregated inference, explaining the principles and operation of prefill/decode disaggregation, a GPU-free render frontend, and their combination in vLLM v0.30.0 and later.

Why it's recommended:The author is from IBM Research, which participated in its development, and provides data on the benefits of separating P/D and render, a table of applicable scenarios, and reusable operating methods.

OpenRouter:AnnouncementsTutorialsAI Score62

OpenRouter tutorial: How to test AI agents' tool-calling accuracy

OpenRouter Publishes a Tutorial on Testing AI Agent Tool-Calling Accuracy, Separately Testing Failures as Tool-Selection Errors and Parameter Errors.

Why it's recommended:The original article divides tool-calling failures into two categories: choosing the wrong tool and passing incorrect arguments. It provides three evaluation methods and cross-model testing code that can be applied directly to your own agent evaluations.

OpenRouter:AnnouncementsTutorialsAI Score65

OpenRouter Tutorial: Build a Golden Evaluation Set from Production Traffic and Retest Across Models

OpenRouter published a tutorial on building a golden evaluation set from production traffic to use as a regression test before each deployment. It covers a five-step process: sample production traffic, deduplicate and cluster, add expected outputs, review the first evaluation and refine the rubric, then commit to Git and integrate with CI. The tutorial recommends starting with 20 to 50 reviewed examples and expanding to a complete regression set of 100 to 1,000, using real traffic rather than synthetic data to preserve its distribution and failure modes.

Why it's recommended:The article provides a five-step process for building a golden evaluation set from production traffic and a method for retesting across models, ready to follow as-is.

OpenRouter:AnnouncementsTutorialsAI Score69

OpenRouter tutorial: How to regression-test AI Agents after changing prompts or models.

OpenRouter publishes a guide to regression testing AI agents: after each change to prompts, models, tool definitions, or retrieval settings, rerun a locked test suite and check the results against a written behavioral contract.

Why it's recommended:The original text provides a reusable method for agent regression testing, covering the full process of locking down test cases, defining behavioral contracts, pinning specific model slugs, and comparing model changes.

Load more

Finished reading the news? Let’s talk about your business.

Leave your contact details, and I’ll get in touch with you personally.

WECHAT

Add me on WeChat

Press and hold or select the WeChat ID below to copy it.