Andrius Aguinaldo

80%+ on SWE-Bench Verified for an open model is wild — especially if it’s actually usable in real workflows and not just benchmark-flexing. Curious how it holds up on messy, legacy codebases vs clean benchmark repos?

I like the framing — crypto UX has always felt built for power users, not beginners. Making it conversational lowers the barrier a lot. Curious how you handle education vs execution so new users don’t accidentally make irreversible mistakes?

The “marathon in flip-flops” analogy is painfully accurate for a lot of automation stacks. The persistent data layer angle is interesting — especially if it truly acts as a real SSOT instead of just another sync layer. Curious how you handle schema changes and versioning when multiple agents are writing to the same data?

Love the focus on thinking before screens — most AI tools jump straight to pretty mockups and skip the messy product logic in the middle. The persistent memory + parsing a live app via Chrome extension is especially interesting. Curious how well it handles really complex edge cases in multi-role products (like admin vs end-user flows)? Definitely going to play around with it.