News for hackers. Stuff that matters.
Re:Peter Norvig critically reviews AlphaCode's code quality by AnimalMuppet on Friday December 16, 2022 @06:08PM (#34021869)
Let's say that I have an encyclopedia of philosophical arguments, with an index and table of contents. Let's suppose I have a philosophical question, and I get the book, use the table of contents to find a relevant article, read the article, and am impressed by the reasoning.

Do I decide that the paper the book is printed on has good reasoning? That the table of contents does? Or that the author does?

What you linked is a regurgitation of a digest of the reasoning of others. It's not reasoning. (Of course, you could argue the same about much of human "reasoning", too...)

Re:Peter Norvig critically reviews AlphaCode's code quality by fossuser on Friday December 16, 2022 @06:14PM (#34021928)
> “ Of course, you could argue the same about much of human "reasoning", too...”

This my point - you can argue the exact thing for humans. I thing Turing was actually more prophetic with his test than people give him credit for.

Of course humans aren’t looking up and reporting back the exact text of an article, but neither is gpt.