EPYCREST / BLOG
Analysis, field experience and methods for building software with artificial intelligence.
A large portion of AI quality failures trace back not to the model itself, but to how humans communicate with it. The provocative question: should we flip the table and start structuring humans as instruments that feed AI — rather than the reverse?
Read article →
When AI agents can generate code faster than you can type it, the developer's role shifts from execution to orchestration. This isn't about coding less—it's about thinking differently, architecting at higher levels, and mastering the art of steering artificial collaborators toward quality outcomes.
Read article →
Definitive guide to AI models for VSCode, Terminal, and CLI: benchmarks, price comparisons, and recommendations by use case. GPT-5.3-Codex dominates with 77.3% on Terminal-Bench, Claude Opus 4.5 reaches 81.6% on SWE-bench.
Read article →