Jonathan, an early user of Jev AI, a $40M-backed platform, shows how its retrieval-augmented generation cuts costs on processing huge datasets. Jev's three main uses: AI automation with clean APIs, real-time applications faster than human perception, and an AI-MapReduce approach that processes data up to 100 times cheaper.
Jev AI comes out, type-safe AI emerges from stealth with $40 million in funding with new model for composable AI. Jonathan, you actually have been playing around with Jev because you got early access. Any initial reactions, takes on using Jev yourself against our use cases?
First, I think this was exactly what I wanted to see in retrieval augmented generation tool. And for me, it's just exciting. Like we deal with a lot of transcripts, right? And we can ask a whole bunch of questions against a transcript corpus. In parallel, get answers back, and like, amazing. Like, you know, infrastructure setup, so much faster, so much cheaper. So just right off the bat, there's really practical use cases. When I requested early access, I said, hey, we build self-models. We have a probabilistic output for them. I want to see how Jev works in that use case. And so what I've been doing is breaking down our self-model versus world model simulator and having Jev predict like what the self-model will decide in a certain situation, given its belief structures. Basically what I'm doing is translating our self-models to a set of questions that are asked against a world model. That is in the— that carried in the state. And when you look at the TypeSafe docs, I really like just the abstraction that they bring to the API is really clean. When I talked about a transcript, that would be in the state variable. You can sort of chunk that up any way you want to as far as like a JSON structure. But then the important thing is like questions. How do you formulate the questions? And of course I'm using Codex to just like interact with their playground and Codex interact with their API. It's able to like just understand everything. Just dump the docs into your Codex. What Point it at the browser, pulls everything together. You have a state structure, you have questions, and then each of the questions has choices. So it's not an open question. You can't ask open questions. You can ask questions that have a set of choices associated with them. And there is some art to both how you structure questions and how you structure choices for each of the questions. I found it like works just as described perfectly out of the gate. And, you know, the, the probabilities it gives back, I'm getting more of a feel for as I go through the scenarios and like what it's good at and what it's not. But it's a really, really powerful tool.
Wanted to bring up, there's some example use cases in their documentation here. So AI automation software, interleave AI with reliable software in a way where you can run it a million times in the background without a human copilot. So for the first time, they're promising autopilot, true autopilot. Code owns control flow, not markdown files, while TypeSafe handles the semantic decisions and language understanding. Next use case for Jev: real-time applications. Frontier intelligence at real-time speeds means AI can make decisions faster than human perception. Fast and smart enough to be programmed to play games or embedded into a UI. AI MapReduce over big data. This is a cool big data use case. 100x cheaper means you can process gigantic datasets. Jonathan talked about transcripts there, like all of our soft models. We have snapshots of your digital twins from your marketing funnel, whatever emails they click, how far they scrolled, all those things, all the instrumentation you have on the mobile side. If you have a mobile app, all the different product analytics, where they click, how long they view a screen, those types of analytics. What we do at Clarity is, is we use all of those as episodes, normalize to your graph of that individual. So we have snapshots of who they are and what their probabilistic confidence scores of how they evidence their beliefs are and what their belief systems are, in particular with your ecosystem. But this has made me really happy because like Jonathan said, we have to process shitloads of data to be able to have this digital twin-like experience. And having it more real-time is possible now with this instead of just throwing money at the problem. This is really awesome. Like, you don't have to throw money at the problem and it's faster and it's more reliable from an engineering infrastructure perspective.
