Hi Christian, Thank you, yes I don't disagree with anything you said. There can be no "one size fits all" and customizing performance tuning will always be important but I will argue 1. There can be a "one size fits most" at least for desktop client environments ("general optimization") 2. It may be surprising what "general optimization" entails given the general lack of consideration to the psychological experience of the user. Apple discovered this as a historical accident a few decades ago. I am no fan of them, though, because they don't allow that customization. My wife is dyslexic and uses a Mac for work that does not allow her to install and use OpenDyslexic fonts in the OS because Apple has already "determined what is best." To ground the 2nd point above, I would give a hypothetical: trading 25ms of latency/jitter for a 10% gain in throughput might seem like a no-brainer from a benchmarking perspective. But when user psychology is factored in as well as allowing for adequate default performance for the widest use cases available, the tradeoff quickly becomes unacceptable. The "relatively large" 10% throughput has very little relevance outside of benchmarking whereas the 25ms of latency/jitter can make or break entire workflows and usage scenarios from a user perspective covering a broad set of scenarios. The only danger there is that people will compare "out of the box benchmark performance" and say "this system is slower than that system!" But I agree, now it's time for more discussion and input (including data) from others. I'm glad that this discussion is occurring! I don't think I have anything more to offer at this point. ethan On Thu, Jan 11, 2024 at 11:20 PM Christian Ehrhardt  <