Open FPGAs would be disruptive to the status quo, that's why they are closed. Right now many technologies like video cards are wandering in the desert, adding

Open FPGAs would be disruptive to the status quo, that's why they are closed. Right now many technologies like video cards are wandering in the desert, adding on more and more opaque layers rather than providing general purpose computing. So that's set simulation back at least a decade or two and hindered progress in fields like AI. I just want to address one of the main complaints about FPGAs - that they require more chip area to route logic. This is not as big of a deal as it sounds because today 3/4 of a chip's area or more is often dedicated to mundane things like cache. Also since there has been little progress in breaking the 3 GHz barrier since the early 2000s, everything is moving towards higher transistor counts. So the added cost of layout goes down over time and after about 3 years is on par with the previous generation. So we have a chicken and egg problem where true general purpose parallel computing can't get off the ground because it's perceived as too expensive, but it's too expensive because it hasn't gotten off the ground yet. Breaking that chicken and egg cycle by opening FPGAs could trigger an overnight adoption of them, even greater than when Bitcoin triggered renewed interest in ASICs.