AI Disruption
Subscribe
Sign in
Home
Podcast
Chat
Chip
Meta
Paper
Qwen
Agent
Robot
Cursor
Google
OpenAI
AI Code
AI Video
Amazon
Microsoft
Windsurf
Anthropic
DeepSeek
AI Weekly
Elon Musk
AI Writing
AI Painting
AI Business
🎈 Guest Posts
AI Open Source
Chinese Outbound
Foundation Model
Archive
About
AI Code
Latest
Top
Discussions
Claude Code Released 17 Versions in Half a Month
Claude Code shipped 17 versions in half a month. Here are 8 changes worth using, from security fixes to cache savings and new plugins.
2 hrs ago
Â
•
Â
Meng Li
4
2
GPT-6 Astra: 7 Ways to Save Tokens
Learn 7 practical ways to save tokens with GPT-6 Astra, including low reasoning effort, prompt caching, deferred tool loading, and context compaction.
Sep 10
Â
•
Â
Meng Li
10
4
Claude Security Plugin: Scans Vulnerabilities, Writes Patches
Claude Security Plugin uses multi-agent AI to scan codebases, filter false positives, and generate reviewed patches. Scan commits or full repos and…
Sep 9
Â
•
Â
Meng Li
19
6
Claude Code Truly Understands: Voice Input in Action
Claude Code voice input review: real-time transcription, tap or hold modes, Chinese support, and tips for long prompts.
Sep 8
Â
•
Â
Meng Li
9
3
Claude Fable 5.1 New Command: Clear Old Prompts
Upgrade to Fable 5.1? Use Claude Code's prompt-audit to clean stale instructions, tool overprompts, and legacy scaffolding.
Sep 7
Â
•
Â
Meng Li
6
3
Feeling Overwhelmed by Claude Code's Many Features
Feeling lost with Claude Code's many features? This guide maps each tool to real problems so you know exactly what to use and when.
Sep 7
Â
•
Â
Meng Li
8
3
Don't Let AI Ask 200 Questions and Avoid 6 Grill Skill Traps
Stop AI grill failures. Learn 6 mistakes in fidelity, scope, model choice, and context to run smarter planning sessions.
Sep 6
Â
•
Â
Meng Li
9
3
Claude Code Prompts: Top 10, Just Copy
Stop improvising prompts. Top 10 Claude Code prompts from the official library—copy, paste, and ship better code faster.
Sep 4
Â
•
Â
Meng Li
19
4
Building a Self-Improving Agent Framework with Skills
Build a self-improving AI agent with Skills, human feedback, and automated loops. Practical guide from Warp's open-source implementation.
Sep 1
Â
•
Â
Meng Li
13
2
OpenClaw 2.0 Released
OpenClaw 2.0 launches with 933 contributors, 16,000 PRs, new memory, automation, security, and multi-agent features for a smarter AI runtime.
Aug 31
Â
•
Â
Meng Li
13
4
ego lite is the best browser agent
Ego Lite is the best browser agent for logged‑in web tasks, faster and simpler than Codex plugins.
Aug 30
Â
•
Â
Meng Li
12
4
I Now Completely Stop Reading AI-Written Code
How top developers stop reviewing AI code, build agent‑verifiable tasks, enforce rules via CI, and design projects so even average agents write correct…
Aug 28
Â
•
Â
Meng Li
16
5
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts