Stefan Streichsbier
@s_streichsbier
Building AI agents that keep software secure 🤖 ✨
The new definition of skill is how many agents you can command with meaningful objectives and utilization in parallel.
I really enjoy writing test-suites with claude code. It's really fun to built a rock-solid, fast, and comprehensive set of tests to continuously verify the state of your application. Speeds up the agentic coding loop.
Are there any OpenAI Teams users in Asia that got access to ChatGPT Agent already?
Based.
It's becoming more and more clear that Claude Code is the everything agent
I would love to get some credits and try it out.
Lovable just crossed $100M ARR in 8 months. Faster than OpenAI, Cursor, Wiz, and every other software company in history. Today we're launching a game-changing update, it reduces the error rates by 91%. Introducing Lovable Agent:
One thing that stands out is how fluently Opus uses the command line. Here, Claude queries the @kuzudb graph directly. Many commands are longer.
Claude Code with Opus 4 feels close to AGI for me.
.@Anthropic When can we finally get named sessions in claude code and the ability to fork them?
forking; when you do a `claude --continue` on an active session and end up with two child agents off that one point.
Claude code is like a TODO list that resolves itself.
Do we have claude code controllable from mobile phones yet?
Calling it now: Specialized tools for agents will be a big market.
The question becomes: How many CC Max plans can you use concurrently?