2 ms·
> Thinking in general works very much like vision. Well, that's a sweeping statement. No reason to read any further. One simple argument against it: thinking a
by tgv 6y ago
> Thinking in general works very much like vision.
Well, that's a sweeping statement. No reason to read any further. One simple argument against it: thinking about problems is supported by much more memories, and of a different kind, than e.g. face recognition, and it's much slower, too.
- pbw 6y agoI believe how thinking works has many parallels with how visual perception works, because they both use the same underlying hardware. Chunking and hierarchy are super primitive and very likely are abilities that permeate the brain. Kurzweil (I know, I know) in his book How To Create A Mind writes about how we read: For example, to recognize a written word there might be several pattern recognizers for each different letter stroke: diagonal, horizontal, vertical or curved. The output of these recognizers would feed into higher level pattern recognizers, which look for the pattern of strokes which form a letter. Finally a word-level recognizer uses the output of the letter recognizers. All the while signals feed both "forward" and "backward". For example, if a letter is obscured, but the remaining letters strongly indicate a certain word, the word-level recognizer might suggest to the letter-recognizer which letter to look for, and the letter-level would suggest which strokes to look for. Kurzweil also discusses how listening to speech requires similar hierarchical pattern recognizers. https://en.wikipedia.org/wiki/How_to_Create_a_Mind https://en.wikipedia.org/wiki/How_to_Create_a_Mind