A week after Jev's launch, 1,142 GitHub repos catalogued into 23 categories
Someone sorted the 1,142 GitHub repos built around Jev into 23 categories; the largest are evaluation and observability, agent tools, games and sims, and security guardrails, while voice chat has only 5.
someone just catalogued 1,142 GitHub repos built around Jev, sorted into 23 categories the model had been out for about a week and look at where people put it. not in the writing. in the if statements this Ruby gem is the cleanest example: > if Hunch.almost_certain?("this order is fraudulent", given: order.attributes) > order.hold! > end http://github.com/carldaws/hunch no prompt. no parsing a reply. a calibrated probability goes straight into control flow the rest of the list is the same move in different spots: > pi-jev (157 stars) your coding agent wants to run a command or edit a file. Jev checks it first: can this break something? leak data? if yes, it stops and asks you http://github.com/y0usaf/pi-jev > jev-cli you give it a claim and the evidence. if the claim isn't backed up, it returns an error code and your CI build fails http://github.com/Nasrallah-AL/jev-cli > Jev Browser an agent that browses websites, and Jev picks every click. Coffee -> Espresso on Wikipedia in ~4 seconds for $0.0016 http://github.com/jkudish/jev-browser the biggest categories in the list: 118 evaluation and observability 109 agent tools 97 games and sims 57 security and guardrails and speech and dialogue? 5 repos most of this is early. test it before you trust it a week in, people aren't building chatbots on Jev they're building brakes
This post quotes another post: x.com/asteri_eth/status/2104901781725598103