Jacob Ibáñez 🔒 on Nostr: And note that I'm not talking about the wasting of resources or the border line ...

And note that I'm not talking about the wasting of resources or the border line outlaw methods to train it. I'm just questioning it purely from the logical side. It goes against EVERYTHING I've been trained as a programmer to do. How many times we didn't trust some code because the tests are green on the first pass? Or even better, how many times you've forced a test to fail to check that the test is indeed making the assertions correctly? Seriously, this is nuts.