Games & real time

Playing around with Jev and completely automated Minecraft in a couple hours. Token costs under $1 for hours of gameplay. What a time to be alive :) https://t.co/BJ7gn8T2DI

Jev同士のぷよぷよ対決も作ってみました 落ちゲーのAI対決は一生見てられるな... 多分改善すればもう少し連鎖も作れる気がする 今の所５連鎖を観測できてない https://t.co/YSaLWnfCbi

I let @typesafeai's Jev play SameGame. No search, no lookahead. Each move, code writes down what every legal move does and Jev picks one. About a third of a second a move, a tenth of a cent a game. Watch it clear this board, then press ✨ Jev: https://t.co/p6s1oO3fqH https://t.co/PyfsUtLI2Q

Let me get in on the Jev hype train. Since I have no early access I tried recreating Jev's O(1) decision only architecture using a standard Causal LM to run a real-time Snake Game ~120ms per decision (8 moves/sec) with zero token generation https://t.co/jPRrIxIoyp

I put Jev and GPT-6 Astra head-to-head in my new word game Yapello. Available here: https://t.co/di5kpYZEId Jev answered 10+x faster at a fraction of the price. Astra scored more. Both confidently submitted a word the game rejected. Final: Astra 105, Jev 72. Five decisions each: Jev 2.02s total API time; Astra 24.29s. API cost: Astra $0.32 · Jev <$0.01 (creator-provided billing figures). How this

TypeGPU + ruNNtime + Jev @typesafeai is a very fun combo :D ruNNtime gives me efficient local inference, TypeGPU lets inference and rendering share GPU resources directly with zero copy. That’s 3 separate NN inferences plus rendering, all happening in realtime Since we control the pipeline, Jev can just sit in the middle and add the semantic bit. camera + mic → Moonshine + YOLO26 + DepthART → Jev

Jev v1.13 dodges rockets with probability calculation 🚀 @typesafeai's new non-LLM model returns decisions instead of text so we had it calculate a safe tile every 330 ms while rockets fell and it survived 25 of 26 for under a cent Run Jev via API -> https://t.co/RbcCOIgVkj https://t.co/f2VmOEPFnO

I got early access to Jev, so obviously I gave it $1,000 and let it play blackjack by itself. It made 243 hands in ~191 seconds, briefly reached $1,080, then ultimately depleted the $1,000 bankroll to $5. Turns out Jev is fast. Just not a very good gambler. @typesafeai https://t.co/NarPF9o3jB

2/ Cua Driver + Jev playing 2048. In this measured 2048 run, Jev was 5x faster and about 1,000x cheaper than Astra. That result is specific to this run, not a general claim about either system. Video attachment: 30 seconds, silent, 1920 x 1080. The final frame shows Jev at 44.9 seconds and about $0.00108 API-equivalent cost, versus Astra at 294.9 seconds and about $1.45. Demo by @injaneity

So I did a thing with Jev. Saw a few posts about training Jev to play video games. I figured, why not try to get Jev to play a song? Found a piano version of Bella Ciao. Put it through a piano transcription model. That gave 1,353 notes with their timing and loudness. The timing and loudness became the fixed level layout. The pitch itself was hidden from Jev. The task: At each note, Jev had to pick

Can Jev play Mario? Made this quick demo to check it, and from 10 tries it was able to get to the finish line only once! Most probably the issue is within the prompt that I am sending, as it is not being able to properly calculate the statistics from the information given. Will polish it further....lets see!

In my demo, Jev can play 2048 but makes some beginner mistakes without an advanced harness. It's so fun to watch https://t.co/HBqjIPNxqX

made a Jev-like classifier model + pathfinder to control my enemy AIs in my capture the flag minigame built in @threejs https://t.co/GkZujHExNG

I gave an evolution simulation to Jev, a small model from @typesafeai that answers typed questions with odds instead of writing text. Two species with opposite DNA. 14 generations. An ice age. 432 creatures. Jev decided who survived, who mated, and what killed each one. https://t.co/Law8fyfy6d

Can Jev bluff itself? I spun up two Jev agents, gave them 1,000 chips each, hid their cards from each other and let them play heads-up poker until one went broke. 20 hands. 153 Jev decisions. Biggest pot: 1,760 chips. Both got caught bluffing. Red Jev took all 2,000. The entire match cost $0.02. @typesafeai

Jev only outputs probabilities. So I gave it the Snake board as text and asked one question every frame: up, down, left, or right? No training, no game code inside the model. It played fine for a while, then boxed itself with nowhere left to go. https://t.co/e6lGyVTrlk

Wow it’s fast 🤯 Here’s Jev, in real time, scoring guesses and categorizing questions inside a tiny “Hot or Cold” game I threw together https://t.co/tJ0nKVsMxJ

I used Jev to make an audience of 100 personalities to yap to. Each blob has its own personality and makes its own Jev call every time you talk to decide if it’s bored of you yet. Each round costs <$0.01 in credits. Really cool model from @typesafeai ! https://t.co/nGlKuxxfAA https://t.co/YgRdcS5pfs

This is absolutely crazy. I connected @typesafeai to my Icy Tower game and I’m shocked by how fast Jev reacts. Jev picks the landings, the game handles the jumps, and a live inspector shows its choices, probabilities, and inputs. Super excited to see how this evolves! https://t.co/wPWS9HrqxP

Have been sick all day so haven’t had much time to mess around with Jev yet. But I did add an update to my game. I realized I suck at it so added a training mode where you can design your training session. You can add rebound boards, cones, more balls and much more. Then just jump into practice. Actually really fun. Also learning tons of cool stuff you can do with @unity

PythonでJevにプレイさせるSnake Game作った（AIにつくらせた）リアルタイムで操作するのが素晴らしい。赤は従来のアルゴリズムNPC（盤面全部みて計算してるので有利） https://t.co/gOy4kq2YXl

Made a wikiracer using jev. Go try it out https://t.co/qUoye0IDVH https://t.co/5F4rvqXwOw

Jev by @typesafeai could be the pilot on your next flight. I've built a simulator where Jev has full control of the airplane. It takes off, follows waypoints, and lands safely. It has full access to flight telemetry, avoids mountains, etc. 🤯 https://t.co/G2xrcaYzGk

I gave 24 AI people a town. Then I flooded it. Each had their own memories, relationships, and limited sight. Jev chose their actions: escape, help a stranger, or look for family. 237 real API calls. Here’s what happened: https://t.co/xYcIn5e0RQ @typesafeai @CompleteSkeptic

I got early access to @typesafeai's new Jev model and built a demo of it playing tetris The generation speed is so fast that it's pushing blocks down This feels like another "this changes everything" moment https://t.co/ND9rVcsPiD

We have jev at home. The arch is good for RL with state-dependent actionsets, so here is the same arch playing doom and chess (shittily) with two different controllers. https://t.co/OdhBY0jF9U

JEV CAN DRIVE! 🤯 I hooked up Jev to the raw controls of a driving simulator, this video shows him driving in realtime. Realtime is the kicker here. Asking a model what to do every 50ms then rendering that into a "realtime" video is one thing, but in this simulation I wanted to give Jev the challenge of actually controlling a moving vehicle in a simulation that will never pause while he thinks.

got access to jev and instead of doing anything useful i made it play itself at 3D chess same model on both sides, slowly developing beef with itself you can watch jev vs jev or jump in and ruin its day https://t.co/VzDwchWqsI https://t.co/ly8d31pjSr