[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

![GregKamradt Avatar](https://lunarcrush.com/gi/w:24/cr:twitter::240008974.png) Greg Kamradt [@GregKamradt](/creator/twitter/GregKamradt) on x 41.7K followers
Created: 2025-07-25 15:48:19 UTC

X day update w/ ARC-AGI-3 Preview

* As expected, humans have found the optimal run for each game. This helps us set the lower bound of # actions needed. Though, keep in mind this is after many iterations of playing, so it's useful only as a lower bound of actions (not learning)

We'll release another X games in Aug (per the competition) and they'll figure those out too

* Agent progress is either 1) human hard coded knowledge or 2) random agent w/ 60K steps.

The goal was learning and we're learning

![](https://pbs.twimg.com/media/GwttYf3bgAIkh40.jpg)

XXXXX engagements

![Engagements Line Chart](https://lunarcrush.com/gi/w:600/p:tweet::1948772309768311144/c:line.svg)

**Related Topics**
[playing](/topic/playing)
[bound](/topic/bound)

[Post Link](https://x.com/GregKamradt/status/1948772309768311144)