points by simonw 1 month ago Once you've seen the AI not make mistakes like that a few dozen times you start to trust it not to mess that up in the future.
discreteevent 1 month ago Is this really a rational strategy for something whose nature is to be right most of the time and then spectacularly wrong a much lesser amount of the time? simonw 1 month ago Yes, because the mistakes in code are easy to spot. I wouldn't use this to write me an unreviewed legal brief. discreteevent 1 month ago But you said that in this case you don't review the code because you trust the AI having seen it write that kind of code before. twister2920 1 month ago you literally just said you don't review this code!
simonw 1 month ago Yes, because the mistakes in code are easy to spot. I wouldn't use this to write me an unreviewed legal brief. discreteevent 1 month ago But you said that in this case you don't review the code because you trust the AI having seen it write that kind of code before. twister2920 1 month ago you literally just said you don't review this code!
discreteevent 1 month ago But you said that in this case you don't review the code because you trust the AI having seen it write that kind of code before.
Is this really a rational strategy for something whose nature is to be right most of the time and then spectacularly wrong a much lesser amount of the time?
Yes, because the mistakes in code are easy to spot. I wouldn't use this to write me an unreviewed legal brief.
But you said that in this case you don't review the code because you trust the AI having seen it write that kind of code before.
you literally just said you don't review this code!