Jev is a new kind of AI model built around taking actions instead of generating a stream of tokens. It is extremely fast and cheap in the right harness—but what happens when I give it an open-ended game like Minecraft? I compare Jev’s strengths in simple action spaces, computer use, robotics, and a Minecraft setup built with Astra.
HUGE THANKS to Crusoe for sponsoring today’s video: Crusoe Intelligence Foundry lets you call supported open models through an API and experiment with serverless fine-tuning without managing the underlying GPU infrastructure. New Crusoe accounts receive $5 in starting credits; use the official link below, or the tracked sponsor link supplied in the campaign brief if it differs.
▼ Link(s) From Today’s Video:
Jev on OpenRouter:
Typesafe System One / Jev concept docs:
Jev on Vercel AI Gateway:
JevplaysMinecraft:
Subway Surfers reference:
Community and demo posts referenced in the video:
MattVidPro Discord — share your experiments:
Follow Me on Twitter:
Buy me a Coffee!
▼ Extra Links of Interest:
General AI Playlist:
Instagram: instagram.com/mattvidpro
Tiktok: tiktok.com/@mattvidpro
Gaming & Extras Channel:
Let’s work together!
– For brand & sponsorship inquiries:
– For all other business inquiries: [email protected]
Thanks for watching MattVideoProductions! I make all sorts of videos here on Youtube! Technology, Tutorials, and Reviews! Enjoy Your stay here.
All Suggestions, Thoughts And Comments Are Greatly Appreciated
▼ Chapters:
0:00 Jev arrives: an AI built for action
0:11 What Jev is good at—and where the harness matters
2:25 Crusoe Intelligence Foundry sponsor
3:48 Jack Cheng’s real-time computer-use demo
4:55 Why Jev needs a harness
6:50 Jev controls a robot arm
8:25 Jev + Astra play Minecraft together
9:06 My solo Jev Minecraft test
11:35 How to access Jev
12:14 Final verdict: fast, limited, and promising in tandem
source
