News for hackers. Stuff that matters.

A comment on Intelligence as a Halting Oracle.

Re:Intelligence as a Halting Oracle by empath75 on Monday November 05, 2018 @04:58PM (#18386341)
> If we only have a partial halting oracle, we still get a speed up by eliminating the non-halting programs that can be identified by our partial oracle.

This is wrong on so many levels that the entire rest of the essay becomes nonsense.

[ Parent | on HN ]
Re:Intelligence as a Halting Oracle by tree_of_item on Monday November 05, 2018 @05:03PM (#18386382)
Why is it wrong? If the oracle can only identify a certain class of nonterminating programs, you still get a speed up by doing so.
[ Parent | on HN ]
Re:Intelligence as a Halting Oracle by empath75 on Monday November 05, 2018 @07:45PM (#18387388)
If you told me you wanted to count through all the natural numbers but we’re going to skip the evens, how much time would you save.
[ Parent | on HN ]
Re:Intelligence as a Halting Oracle by yters on Wednesday November 07, 2018 @08:56AM (#18399087)
In the article's scenario there are a finite number of Turing machines to check in order to find one that matches the specification. So, eliminating some of them will lead to a speed up.
[ Parent | on HN ]
Re:Intelligence as a Halting Oracle by yters on Monday November 05, 2018 @05:19PM (#18386514)
You may want to check out the proof for further information.

https://news.ycombinator.com/item?id=18377525

[ Parent | on HN ]
Re:Intelligence as a Halting Oracle by yters on Tuesday November 06, 2018 @11:41AM (#18392557)
Direct link: https://www.am-nat.org/site/law-of-information-non-growth/
[ Parent | on HN ]