Connect Four Apprentice
It never sees the board. It picks one idea per move and learns which ideas work in which situations.
–Not enough games yet
- Lessons
- –
- Games
- –
- Players
- –
It started this season knowing nothing. Every game you play teaches it.
Adapt-1 is a learning engine from REI Labs. It is not a language model and it was never trained on these games. Here it plays people, learns from each result, and you can watch its brain change while it happens.
Watching is open to everyone. Sign in with X to play.
Mind Reader
Rock, paper, scissors. It seals its throw before you choose, using the habits it has seen in how people play.
–Not enough games yet
- Lessons
- –
- Games
- –
- Players
- –
How a season works
It starts empty
Each season gets a brand-new Adapt-1 brain with no memory of past seasons.
It chooses
Before each move it reads a few facts about the situation and picks what it believes works, sometimes trying something new.
You answer, it learns
The result becomes one lesson. Its beliefs update immediately, before the next move.
It becomes the champion
After a fixed number of lessons the brain freezes. You can still challenge it, but it no longer changes.
From its notebook
Patterns the Mind Reader has found in how people play rock, paper, scissors, written from its own rules.
No patterns yet. Play a few rounds of Mind Reader and they will appear here.