2 ms·

I don't think the sceptics have tried it. It's true it gets things wrong sometimes but that's a different problem to synthesising novel solutions.