AI

Implemented — in the honest sense: AI demonstration examples written in ordinary Karkain arithmetic exist under examples/09-ai/ and run byte-identically on both engines.

There is no ``std.ai`` module and no AI/ML framework surface in the language. These examples show classic ML geometries (distance, weighted votes, linear decision boundaries) implemented by hand with plain arrays, maps and arithmetic — deliberately, so the language itself can be exercised without pretending a framework exists. See Implemented for how the status vocabulary is used.

examples/09-ai/

File

Demonstrates

01_nearest_neighbor.kark

Manhattan-distance nearest-neighbour classification by hand

02_linear_classifier.kark

Perceptron-style weight updates and a decision boundary

K-nearest neighbour (excerpt)

func distance(a, b) {
    return abs(a[0] - b[0]) + abs(a[1] - b[1])
}
...

Run:

$ karkain run examples/09-ai/01_nearest_neighbor.kark

See also

Machine Learning — the adjacent hand-rolled ML category. Planned — the not-yet-implemented AI framework roadmap.