Tag
#architecture
Why Retrieval and Reasoning Should Be Separate
Retrieval-augmented generation collapses two very different failure modes into one system when retrieval and reasoning aren't treated as distinct stages with distinct contracts.
Designing Deterministic Boundaries Around Probabilistic AI
LLMs are probabilistic components. Reliable systems are built by deciding, explicitly, which parts of the system are allowed to be probabilistic and wrapping everything else in code that behaves the same way every time.