If you have factual observations to make that's fine but can we stop with the "smells like vibe coding" attacks? It's like an AI version of an ad-hominem attack.