Infrastructure evolves layer by layer, new L1s emerge, faster L2s arrive, but the user ceiling remains unchanged. @anoma's "upward" thesis is truly thought-provoking: shift innovation to the application layer, treat blockchains as interchangeable services, and let intent drive the "what" while solvers handle the "how." ⬆️ I read the article carefully and double-checked the key takeaway: the interface specification is essentially composed of two parts: transaction functions (writes) and projection functions (reads), so applications don't need to be hardwired to a single blockchain to achieve their desired results. The roadmap reinforces this: starting with the EVM and expanding outward; Anoma isn't "another L1 layer," but an operating system built on top of it. --- What this unlocks in practice: ❯ Cross-chain application interface (no need to rewrite every chain). ❯ Intent → Solver combinatorial routing → Balanced transactions (goal-first, not action-first). ❯ Faster delivery: define transactions/projections, choose a provider, deploy to where users are (starting with the EVM). --- Monolithic architectures continue to pursue marginal TPS; this model decouples protocol, compute, and data, enabling applications to be composed across systems while respecting privacy, logic, and user agency. It reads less like marketing and more like a clean operating system abstraction. If you're building, design your intent model now and deliver based on transactions + projections; think of chains as configurations, not architectures. I'll call the "upward" model a win when we see a public solver marketplace and real applications shortening delivery times across multiple chains—those are the outcomes I'll be tracking next.