Tools & repos

How Jev Plays DOOM: The Real-Time Decision Demo Explained

What TypeSafe’s ten-queries-per-second demo proves, what it does not, and how to design a small decision-loop experiment.

2min read
1prompts to copy
5sections

The J video shows why a fast decision can matter more than a long answer. TypeSafe’s DOOM demo connects Jev to a running game, but understanding the input is essential to understanding the demonstration.

Watch the original demonstration

TypeSafe’s launch post and DOOM demo

The company says the demo makes ten queries per second and uses structured game-state text, not screenshots. The model chooses actions from that representation. TypeSafe reports roughly $7 per hour for that demo configuration and acknowledges that a conventional DOOM bot could play better. This demonstrates a particular decision loop, not visual game understanding or dominance over specialist bots.

Separate the demo from the applications

The robot and voice-assistant scenes in the reel are illustrative applications. They do not establish that Jev controls a real household robot or a production ordering service. Likewise, the speed and cost comparisons are company-reported results for selected tasks, not a guarantee against every model on every workload.

Read the broader Jev and TypeSafe AI guide

Free, for your email

Get the full setup steps + the copy-paste setup prompt

The rest of this guide (3 more sections, 1 copy-paste prompt) opens right here. The setup prompt, one paste into Claude that sets it all up, comes by email only. Use your real address.

Unlocks: Design a small decision loop · Measure the whole loop · Use the official API starting point

50,000+ follow where these guides come from. No spam, reply “unsubscribe” any time.

Read next