We have a sneaking suspicion that the robots that write LLM code are going to like SQLite too. We think what coding agents like Phoenix.new want is a way to try out code on live data, screw it up, and then rollback both the code and the state.
Prescient.
Agents would of course work well if they can go back in time to checkpoints and branch from there, exploring solutions parallely as needed.
Anyone who has experience with building workflows (Amazon SWF, Temporal, and the like) knows how difficult it is to maintain determinism in face of retries & re-drives in multi-tier setups (especially, those involving databases).
Replit recently announced their Agent's integration with Neon's time travel feature [0] for exactly the purpose outlined in TFA. Unlike Fly.io though, Replit is built on GCP and other 3p providers like Neon and it is unclear if both GCP & Databricks won't go all Oracle on them.
Comments
Agents would of course work well if they can go back in time to checkpoints and branch from there, exploring solutions parallely as needed.
Anyone who has experience with building workflows (Amazon SWF, Temporal, and the like) knows how difficult it is to maintain determinism in face of retries & re-drives in multi-tier setups (especially, those involving databases).
Replit recently announced their Agent's integration with Neon's time travel feature [0] for exactly the purpose outlined in TFA. Unlike Fly.io though, Replit is built on GCP and other 3p providers like Neon and it is unclear if both GCP & Databricks won't go all Oracle on them.
[0] https://blog.replit.com/safe-vibe-coding