Everyone is building what they think is the next Agentic IDE.
I'm building mine too. But I think most people are missing the mark. My version, I believe, is better. No false modesty here — I think there's a real difference.
I got an agent to generate a landing page based on a brief description of how I think the future should work. Then I got other agents to prototype it. Finally, an agent created a video showing the concept.
Is it finished? No. Far from it. But it's interesting because there's so much development happening — multiple agents working on different parts, coordinating, building something coherent from a shared vision.
That's what the IDE should enable. Not just writing code. But orchestrating agents to work together. Defining roles and responsibilities. Letting them collaborate toward a larger goal.
Most IDEs I'm seeing are just refactored versions of existing tools with AI bolted on. Better prompts in a sidebar. Autocomplete that's slightly smarter.
But if you're designing for agents, not humans, the entire paradigm changes. The interface isn't for visual inspection. The workflow isn't about reading. It's about delegation and trust and verification.
That's a different product entirely.
Part of the #100DaysToOffload documenting agentic development in 2026