The public parts of my notebook.
Concepts that are too simple are sometimes not powerful enough solve harder problems. For example, Copernicus tried to explain the motion of the planets with circles. What was needed was a slightly more sophisticated building block: ellipses.
Humans are copers, which means we’re not built to throw away a solution that isn’t working, but, instead, we’re built to patch it to keep it working.
He draws an analogy between hierarchical structures and systems of gears. The latter seize up after about 1000 cogs. In contrast, the internet is like a biological organism with a huge number of decentralised cells. The internet has never broken, despite the fact that every node in it has been replaced at least twice.
He says that it’s much better for the invention arms of companies to work on ten year time scales. This avoids people trying to incrementally get to big inventions, which would mean those things never get invented. At XEROX PARC, they spent the first three years inventing, then the next five years polishing the inventions into things that were sellable.