Hard engineering problems are bottlenecked on insight, not mass computing power. Insight doesn't come from larger models trained on more data, it comes from structured reasoning: contradiction resolution, function transfer across domains, constraint reframing. This has been studied rigorously (Altshuller's TRIZ: 200,000 patents analyzed) but rarely systematized into a tool engineers can actually use.

Sparlo encodes innovation methodology into a usable system. Takes an engineering problem, decomposes it to functional primitives and core contradictions, maps against solution patterns from 40+ industries, evaluates feasibility against stated constraints, outputs structured analysis and recommendation. Thinks like both polymath innovator and pragmatic founder who has to ship.

Designed for mechanical, chemical, electromechanical problems where the constraint structure is real and the industry has settled into accepted tradeoffs that aren't physical limits. Not for software, not for strategy, not for problems where the bottleneck is execution.