レポート 2329

Interesting. The AI totally fails the cognitive reflection test. It gives the intuitive but incorrect answer every time.

This is despite doing fine with similar word math problems. It is theoretically satisfying, since the neural network is most certainly non-reflective. It’s an association machine, which is basically how intuitive reasoning works.

So this is even more interesting. With a little bit of help, it eventually worked its way to the correct answer, even if the reasoning is a little odd!

Am loving the discussion but will have to mute this thread now. If you want to play further, try Stanislaw’s trick of appending "Think about it step by step and only then produce a solution." to the problem.