My goal on X: 10,000 real connections:
I want to meet::
1. AI agent builders
2. iOS and indie devs
3. Solo founders shipping with AI
4. Engineers who read every AI diff
5. People building in public
Starting from 6 followers. If that's you, reply with what you're building..
Let's connect π
If you're building with Claude Code, Codex, Pi, or running multiple AI agents in tmux, we're probably into the same things.
Always happy to connect with fellow builders. I follow back active builders.
What are you building right now?
I don't want another timeline full of AI hype.
I want to see:
β Agents that actually work
β LLM experiments that fail
β Projects built from scratch
β Real code, not just demos
Show me what you're building, not what you're reposting.
Let's connect build in public
@dummerspast39 The AI race is moving faster than ever. Nobody stays on top forever. But counting OpenAI out after one bad week might be premature. The real winners will be those who keep shipping.
@euboid The problem isn't Claude's visuals. It's the lack of creative direction.
AI can generate everything, but without human taste, every video starts feeling like a copy of the last one.
Originality still matters.
@JebrilOsawaru The real shift is from AI that only answers to AI that actually executes.
Chatbots give you instructions. Agents can take action.
The future isn't just better prompts. It's giving AI the tools to get things done.
@notede_1manen Japan's numbers are surprising, especially Claude at just 4%
I'm curious how different the US numbers are. Adoption doesn't always reflect which model is actually the most capable.
I think the real question is: are people choosing the best AI or simply the one they're first get
@nikks_techie Yes, but only partially.
Pydantic validates structure, data types, and defined constraints. It won't automatically know whether the business data is factually correct.
Pydantic checks whether the data is valid, not whether it's true.