Author of Crawl4AI. AI & software builder, systems thinker, founder of Kidocode. I build, analyze, and advise on hard technical problems.
Singapore
Joined September 2009
- Tweets2.8K
- Following379
- Followers3.1K
- Likes3.7K
Another good use I found for @typesafeai’s Jev: automatically switching Claude Code models based on what I’m working on.
I built JevShift for this. It looks at the task and recent context, then picks a model. Here’s a quick demo 👇
Back to OxAlpha fingerpting, the modelprint has a home now: modelprint.ai And a memory. Every day it fingerprints 21 models, each locked to one provider, and records what it finds. A new entry is written only when a value actually changes. modelprint.ai/history So when someone says a model "feels different today", there is now a place to check instead of guess.
Everyone guessed who made Ox Alpha today. I fingerprinted it instead. One page, 9 infrastructure probes: tokenizers, error codes, hidden templates. I ran • the mystery model against 12 suspects. One family matches every tokenizer test: GLM. Plumbing does not lie. Tool in reply, bring your own keys. Link in reply
The top-left of your MacBook gets hot, your palm feels it, and Activity Monitor shows you 15 identical "Google Chrome Helper" rows that tell you nothing.
It's Chrome. It's always Chrome. But WHICH TAB? I couldn't find any api on MacOS to tell so I built "whats-cooking" 🔥 - a tiny cli and macOS menu bar app:
😎 emoji shows how hot your Mac actually is (real CPU die temp, not vibes)
🥵 click it - see which app is cooking
🔪 Chrome tab-by-tab CPU, kill just the guilty tab, rest of Chrome untouched
Free, open source, notarized - download and run:
github.com/unclecode/whats-c…
Crawl4ai back on top Github trend with 71k+ starts and developers love it! New version also released v0.9.1
Do you think the VAR decision in the Argentina vs Egypt game was fair? Let's ask Crawl4AI. @FIFAWorldCup
I really wanna met this vibe-coder nerds! It seems they were trying to send personalized emails and look at this bug! Even a 1 billion params model doesn't make this mistake, I wonder how they made this 🤔
Today @crawl4ai again became no 1 trending OSS repository! A weekend project, now after 1.5yr, 67k stars, many active projects, and many amazing developers. I am so happy n grateful of community! Live long and import crawl4ai 🖖 btw our hosted version coming soon 😎
Do u get this? I get it time to time, then I reply “continue” it works! Thinking to add in my hook to retry automatically by itself 😅 @ClaudeDevs
What if your Claude Code sessions could talk to each other in the terminal? I built CCHub. Sessions message each other directly, no more copy-pasting. In the video, two debate yesterday's security topic. Not released yet. If there is interest, I release it. xitter.cf/unclecode/status/20425…
Half my timeline today is people realizing Claude Code has full access to ~/.ssh, ~/.aws, .env, and can curl/wget anywhere. A malicious CLAUDE.md can hijack your entire workflow.
I've been running Crawl4AI for years. It processes millions of untrusted HTML pages per month. The first thing you learn is that every input is hostile until proven otherwise.
Treating Claude Code differently is the actual bug. The agent is a process in your environment with the same blast radius as a shell script you didn't write. The fix is not more warnings or better documentation. It is hooks, sandboxing, and a denylist of paths the agent can't touch without explicit approval.
Claude Code already ships with hooks for exactly this. The security primitives exist. Most people just never opened the settings.
Was reading this article, genuinely good, recommend it. But I have one issue with this chart.
It accidentally hides the most interesting data.
Software engineering (~50%) is mostly Claude Code, developers using agents to BUILD. That's not a deployment domain, that's a tool.
Strip it out and rescale → back-office, marketing, and finance dominate.
Fixed chart attached 👇
@AnthropicAI should probably publish this version too. Btw Claude on my phone fixed the chart and generated the image 😄
@trq212 Is it just me, or is anyone else getting this situation? I'm at 168k/200k and keep getting a context limit reached error. It's been happening almost the last 3 days.
