You need to be harness-maxing, spawning 5 subagents to specify a PRD into 100-200 individual sub-features with full XML dependency graph and full Contract specification for each system component.
Spawn 3 orchestrator agents minimum and have them each manage 10 subagents to
Conspiracy theory: do models get dumber on purpose if they think that you like correcting them, and do they become smarter if you indicate that you trust them and want them to make all the decisions for you?
For how good Claude is at explaining math and other technical concepts, its surprisingly bad at data analysis and validating its own claims and methodology
