Building a Codebase Classifier with Jev
A developer shares building a codebase classifier with Jev, suggesting it may solve overengineered code from agents, and asks what to test next.
Same task benchmark: LLM costs $0.013880 and takes 8.566s; Jev costs $0.000081 and takes 0.114s, about 0.6% of the cost. The author notes Jev cannot write prose, only typed answers and confidence, and asks whether you would trade prose for 444x cheaper.
Same task, measured: LLM — $0.013880, 8.566s Jev — $0.000081, 0.114s That's 0.6% of the cost. A million decisions: $13,880 → $81. The catch: Jev can't write. Typed answer + confidence, nothing else. Would you trade prose for 444x cheaper?