2026-01-17 13:42:18 <veltas      > KipIngram: What you're talking about works for embedded but isn't competitive (and I don't think it could be) for performance workloads
2026-01-17 13:42:45 <veltas      > So I can excuse Intel, ARM et al for making them that complicated
2026-01-17 13:43:33 <veltas      > There are a ton of important areas where to be competitive you need the best acceleration you can get, and sometimes that comes from the CPU
2026-01-17 13:43:55 <veltas      > Is it abused to hell and back by consumer market and lazy bloated software ... yeah definitely
2026-01-17 13:44:24 <veltas      > But it's also an amazing luxury that this sophisticated incredible city-sized transistor-scape is available to consumers, for personal computing needs
2026-01-17 13:44:51 <veltas      > You all have the option to go use the simpler processors, tpbsd does an awful lot with those simpler processors
2026-01-17 13:45:12 <veltas      > I like using them too, and I'm amazed at just how few applications really need that performance
2026-01-17 13:46:07 <veltas      > I love seeing how simple and small you can go, and thinking and discussing about where the right tradeoffs are
2026-01-17 13:46:40 <veltas      > It's why I engage in conversations about the least-equipped RISC-V cores and whether their weird bit ordering in instruction encoding is justified by the handful of transistors they can leave out
2026-01-17 13:47:09 <veltas      > We've got a lot of luxury in this realm too really with all the advances in RISC-V out there, and the range of RISC-V cores to choose from