Jev updates a voice order ~200ms after you speak
Kunal Abichandani says his voice coffee ordering system puts the drink on screen about 200ms after the word, while Everton Carneiro says 201 million tokens of Jev keyword judging cost him $8.04. Both numbers come from the builders, not a benchmark.
Built with it
Kunal Abichandani's voice coffee order lands on screen mid-sentence
Kunal Abichandani built a voice coffee ordering system on Jev where the drink and any changes to it render while he is still talking. He puts it at roughly 200ms from word to screen. He did not say how the audio path was measured or what the choice set looks like.
The drink and any updates show up while I’m still mid-sentence, ~200 ms from word to screen.
ILIAS ISM points an agent at a clone of his product and files the takedowns
After a customer asked for a refund on a clone of one of his products, ILIAS ISM says he pointed an AI agent at the clone's domain and let it run OSINT (open source intelligence gathering) and abuse reports in the terminal.
wild what one guy with an agent can do in 20 mins now tbh
Someone measured it
Everton Carneiro says 201 million tokens of Jev cost him $8.04
Everton Carneiro posted his running Jev bill: $8.04 for 201 million tokens. Jev is the model behind keyword judging in Sextant, his product. His argument is that you do not need a frontier model for every call in the stack.
You don't need a frontier model for everything.


