@AmpCode however, amp is great at supporting you to come up with this context, so you can easily grow your agent documentation as you move on. The only drawback I haven't solved yet: avoiding context window creep.

@AmpCode in the process I learned some new things about how amp & models work. Context is still king, and you really need to slice and dice your codebase from time to time.

Just shipped an idea with @AmpCode I had in my mind for quite some time. A lightweight JavaScript library for building adaptive user interfaces that grow with your users. This is the first release and still a work in progress:

Most important member in the list was missing: Angular!.

Just launched Frametide: a production-ready MCP server bridging Figma → code! 🚀 Extract components, tokens & styles, detect design–code drift, and generate framework-agnostic data for React, Vue & Svelte. 👉

Also can’t really beat this office view :)

Currently stuck in Swiss traffic waiting for Gotthardtunnel to reopen, prototyping something with @AmpCode .Never has being stuck in traffic been this exciting

I've just tested @AmpCode for the first time and it is absolute mind-boggling how well it works. The only thing I'm wondering is why I didn't follow @thorstenball 's lead earlier. I've never seen a coding agent work this well without manual action.

Have you tried integrating AI into your workflow? What’s been your biggest challenge — or your biggest win?.

The future of software development teams is hybrid. Human creativity combined with AI scalability is the key to building faster, adapting instantly, and scaling effortlessly. What’s your take on this? 🤔.

AI is only as effective as the systems you build around it. What works for me: .✅ A well-documented design system .✅ Reusable component libraries .✅Clear, structured workflows . These systems make AI a seamless part of the team.

The success of AI integration starts with the briefing. Here’s my framework:. Context: Point to existing components and documentation as a reference.Goal: Define the exact output you need.Constraints: Add rules and coding styles to follow. Better briefing = better results.

Treat AI like a junior developer. - It’s not a magic wand. - It’s a team member that needs clear instructions and structure. Like any junior, AI performs best with:. Clear goals.- Well-defined rules.- Context to understand the system.

When working with design systems and component libraries, I constantly struggled with balancing speed, consistency, and scalability. Then I discovered how AI could be more than just a tool—it could act as an extension of my team. Here’s what I’ve learned.

What if you could scale your development team infinitely—without hiring a single person? Here’s how I turned AI into an essential junior developer for my projects. A thread 🧵👇.

RT @henrythe9ths: There's a shocking fact about AI that nobody tells you: You can catch up to the public AI research frontier in just 2 wee….

RT @iavins: Collection of insane and fun facts about SQLite. Let's go!. SQLite is the most deployed and most used database. There are over….

RT @ManfredSteyer: I proudly present 🕵️‍♂️ Detective: Visualize and Analyze your TypeScript-based Architecture! 🧵

interesting way to combat AI junk content causing degrading model quality.

RT @ThorstenHans: Less than an hour to go 🎉 . Come and see the new Command Trigger for @spinframework .Let's unlock Jobs and CronJobs on t….