Shared my first experiences with Jev in my newsletter here.

Shared my first experiences with Jev in my newsletter here. Some interesting findings and a proof for the origins of the Jev name. https://www.linkedin.com/pulse/sandbox-side-quest-1-jev-system-one-model-ryan-bruins-rpesc/ I replicated the "Line-by-line Search" cookbook example and benchmarked it against an LLM, proving the claimed performance gains, but also showing the massive input token count for using this search approach with the API. I tried a binary search approach (Branch & Bound), and sadly that took even more input tokens than the brute-force line-by-line for my use case. Anyone know of a more efficient approach that would result in significantly less input tokens? I would be happy to learn. 😁