2026-09-20 · America/Los_Angeles · 社区动态 · #9

The famous "Car Wash" question on Jev

But then how should we evaluate its “intelligence”? Does Jev have benchmarks comparable to the usual LLM benchmarks, or is the comparison fundamentally different? Since it doesn't give text output, is it basically the same as forcing a regular model to return structured output, with Jev's main advantage being price and speed? And if so, how do we know whether Jev is actually “intelligent enough” to make the right decisions? For example, say I want to replace the router in my current personal project with Jev. How do I know Jev will make the right routing decisions, rather than just being…

热度 46.0 / 100；排名与评分保留该期记录。