[ImaCake]

Aside from helping those with disabilities, what is stopping authorities using this as a lie detector? I assume the tech isn’t quite there yet.

[FPGAhacker]

Not yet.

> Models were trained separately for every participant and are not generalizable across people.

[ImaCake]

Ah right, so the models are subject to some serious overfitting then. Good proof of concept, but not useful in practice yet.