From prompt writing to flow engineering: what AlphaCodium teaches about problem-solving with models
kashifmanzoor · x · 2026-09-05
The author argues the better question isn't "how do I write a better prompt?" but "how do I design a better problem-solving process around the model?" He's reading up on Flow Engineering — notably the work from the AlphaCodium project — and finds its contrast with standard prompt engineering fascinating.
More from coding & agent
- HKUDS open-sources Vibe-Trading, a personal trading agent for backtests, hits 32.5k GitHub stars — tom_doerr · 2026-09-05
- 106 MCP servers measured: tool-schema context cost spans 1,700x, GitHub's official server eats 27% of a 200K window — aksh_stocks · 2026-09-05
- Agent builds a rocket in Blender via computer use in under 2 minutes for under $0.30 — cneuralnetwork · 2026-09-05
- Authority as a first-class primitive: why agents need portable, verifiable permission objects — DogInteresting8548 · 2026-09-05
- Codex usage telemetry gotchas: keep explicit zeros, as OpenBench adapter nulls cache_write_input_tokens — Due_Emu_8229 · 2026-09-05
- If your agent can pay automatically, it can get scammed: GateKeep402 guards x402 wallets — Wild_Expression_5772 · 2026-09-05