Mr Maigo

Comment Threads

• Scott Sanford commented on What are you reading this summer?
  And apparently the U. S. Navy will be reading Walter Jon Williams. I assume in order to get an edge over the Army's antiquated Heinlein modules. Do they know that Heinlein was a Navy guy? grin...
• Catherine Taylor commented on What are you reading this summer?
  It's meta-meta-meta politics joke. If a response to an innocent: "Well, I thought the earlier works had more substance, and here's something else" is "GRAAAR FALL BEFORE OUR GIVEN TITLES AND REWARDS" with an added bit of snark dissing one of the authors as 'unworthy'... it's all just very silly. (And yes, Goldstein isn't exactly a WASP surname). There's also some bits of historical politics in there. CASBS was founded by the Ford Foundation (yes, that pro-Nazi Ford), and has had some interesting proteges (e.g. Harold Garfinkel, J Rawls, Shmuel Eisenstadt, Thomas Kuhn) many of whom would view Prospect Theory...
• RDSouth commented on What are you reading this summer?
  "That would fuck chemistry, and there wouldn't be any humans" Not necessarily. Maybe it only affects semiconductors. The difference between an insulator and a semiconductor is much smaller than the difference between a conductor and an insulator, rather than midway between, so a small shift (especially if it mainly affected stuff at that end, making almost insulators more insulating more than it makes conductors more insulating) might be enough to make semiconductors impractical without affecting chemistry too much. I speculate. And while my wonder "field" is at messing with atomic behavior, maybe oxidation could be a little slower outside of...
• Bill Arnold commented on What are you reading this summer?
  Re "Simple Heuristics That Make Us Smart" (Evolution and Cognition) Gerd, Todd 2000, just wanted to say thanks for that reference. The book is hitting the sweet spot as vacation mind-tech reading. I'll defer any pointed comments until finishing it. So far, it's - "Why am I reading ML(non-connectionist) and decision systems literature from an alternate universe?". If I'm recalling the time correctly, induction of parsimonious, human-runnable decision systems e.g. decision trees, was a busy topic at the time (2000-ish), and there was already a venerable (as these things go) literature on feature set pruning and feature selection (and frugal...
• Bill Arnold commented on What are you reading this summer?
  OK Brief summary of "Simple Heuristics That Make Us Smart". First, please remember that it was published in 2000 and covers earlier work. - Very simple and frugal algorithms can perform very well in the real world, including some algorithms that are frugal (for meat-minds) in both setup and evaluation and therefore biologically plausible. This aspect makes the book worth reading. - Eschew overfitting. Reinforced this point by showing that some more complex methods like regression (or NNs) overfit on a bunch of simple small test domains. (/snark) - It can be hard to determine through observation/experiment which, if any,...