Skip to the video guides
← Back to videos
Previous video
Previous video
Why I couldn't build Jev at OpenAI — Diogo Almeida, TypeSafe Co-founder & CEO Latent Space
Next video
Next video
Jev: Typed AI Decisions in Python — Build a Classifier + Route Coding Standards with Claude Code Omar Kamal
JevMade field notes / Video guide
Jev AI just changed Trading Forever (way better than gpt-6 astra & Fable)
Moon Dev evaluates Jev for crypto trading pipelines, contrasting single forward-pass parallel classification with generative LLMs. He reviews latency, context, and calibration claims, highlights arithmetic limitations, and shows how to run Jev in shadow mode.
Original by Moon Dev Agent workflows Intermediate 13 min 11 sec Published 21 September 2026
Before you press play
What you’ll find in the video
Jev supports answering multiple structured classification questions in parallel within a single request forward pass. Jev cannot perform date math or division; applications must precompute arithmetic and feed finished numerical state. Shadow mode allows side-by-side evaluation of Jev against an existing execution engine without risking live capital. Worth knowing Fast classification does not equal market profitability, and claimed calibration figures reflect vendor benchmarks rather than verified trading edge.
▶
13:11
Go to the source
Watch it with its creator.
These are JevMade’s notes, not a transcript. The full video, examples, and discussion belong to Moon Dev.
Watch video ▶
Plays here in theatre mode. Use Close to return to your notes.
Jev AI just changed Trading Forever (way better than gpt-6 astra & Fable)
Close ×
Keep this for later
Sign in to bookmark experiments, guides and videos, and keep notes only you can see.
Continue to sign in
We’ll bring you back to this listing.