3 ms·

When I think about treating laws and recipes as code, I think of the difference between absolute machine computation vs a process that collaborates with human j

When I think about treating laws and recipes as code, I think of the difference between absolute machine computation vs a process that collaborates with human judgment. The "stupidity" of computers' literal-mindedness is something that non-programmers just don't seem to understand. Some recommended things to consider: - Aristotle: laws are for the general case, judges for the particular. This gives the saying, "Hard cases make bad law." It's good to have rules, but keep people in control. - Michael Polanyi: not all knowledge is "formalizable". Actually recipes are an example of his: two cooks following the same recipe may give drastically different results. - Roger Penrose: If the human mind were a computer, why does it seem not bound by the constraints Gödel's theorem implies for a formal axiomatic system? I wonder if the author's book engages with any of these people?

tekacs 8y ago
The first quote makes sense, the latter two are at best mistaken. > Michael Polanyi: not all knowledge is "formalizable". Actually recipes are an example of his: two cooks following the same recipe may give drastically different results. ... or rather, our formalizations are lossy for the purpose of their usage, lossily interpreted by their users, or simply not good. Recipes are a good example of a lossy formalization, lossily interpreted (very much on purpose, to keep the recipe a reasonable size and to put only a reasonable amount of effort/time into cooking for the circumstance). They are in no way indicative of our inability to formalize (and indeed changes due to environment, whether they be ingredients, humidity, temperature, etc. are in fact encoded in some more sophisticated recipes, or more commonly in baking). > Roger Penrose: If the human mind were a computer, why does it seem not bound by the constraints Gödel's theorem implies for a formal axiomatic system? It's 'extremely hard' to show that the mind would somehow escape being bound by this (a very incomplete reading as to why can be found at [1]). This is like suggesting that the mind isn't subject to complexity theory by observing that we can factor 'large numbers' in reasonable time. [1]: https://thefutureprimaeval.net/godel-vs-the-human-brain/ https://thefutureprimaeval.net/godel-vs-the-human-brain/

tekacs 8y ago

The first quote makes sense, the latter two are at best mistaken. > Michael Polanyi: not all knowledge is "formalizable". Actually recipes are an example of his: two cooks following the same recipe may give drastically different results. ... or rather, our formalizations are lossy for the purpose of their usage, lossily interpreted by their users, or simply not good. Recipes are a good example of a lossy formalization, lossily interpreted (very much on purpose, to keep the recipe a reasonable size and to put only a reasonable amount of effort/time into cooking for the circumstance). They are in no way indicative of our inability to formalize (and indeed changes due to environment, whether they be ingredients, humidity, temperature, etc. are in fact encoded in some more sophisticated recipes, or more commonly in baking). > Roger Penrose: If the human mind were a computer, why does it seem not bound by the constraints Gödel's theorem implies for a formal axiomatic system? It's 'extremely hard' to show that the mind would somehow escape being bound by this (a very incomplete reading as to why can be found at [1]). This is like suggesting that the mind isn't subject to complexity theory by observing that we can factor 'large numbers' in reasonable time. [1]: https://thefutureprimaeval.net/godel-vs-the-human-brain/ https://thefutureprimaeval.net/godel-vs-the-human-brain/