It's a practical approach, I used vagrant many years ago mostly successfully. I also explored the docker-in-docker situation recently while working on my own agentic devcontainer[0]- the tradeoffs are quite serious if you are building a secure sandbox! Data exfil is what worries me most, so I spent quite some time figuring out a decent self-contained interactive firewall. From a DX perspective, devcontainer-integrated IDEs are quite a convenient workflow, though docker has its frustrating behaviours
Looks like the M5 base has LPDDR5x-9600, which works out to 153.6 from base M4's 120GB/s DDR5x-7500. The Pro/Max versions have more memory controllers, 16, 24 and 32 channels accordingly. The 32 channel M5 top-end version will have 614GB/s by my calculations.
It would take 48 channels of DDR5x-9600 to match a 3090's memory bandwidth, so the situation is unlikely to change for a couple of years when DDR6 arrives I guess
Ventura got a security update last month. Sequoia will get updates for at least another 3 years. These glaring issues will get resolved eventually, even if it is the 'Frosted Glass' update.
I had fantastic results with lxqt some years on an HTPC. System used less resources and seemed more stable with Qt. Perhaps GTK is better these days, but at the time lxqt was a clear winner for that kind of scenario.
For a daily drive DE though, it may be too minimal?
Alternatively, add the multi-row tabs CSS hack for Firefox; it has a the side-effect of half-breaking tab re-ordering, forcing you to CMD/CTRL+click thereby solving the trails problem. 5 or 6 rows of tabs during a research session is not uncommon, conveniently chronological.
That figures. It was fun though to pretend it would scale across processors. Were you tempted to fork the typescript compiler in this project to add some optimizations or something crazier like gpu acceleration for rendering or parallel workloads?
Reminds me of the guy on Youtube obsessed with optimizing Mario64 code, he found really interesting things like how removing optimizations made the game faster - presumably because the very basic tooling the original developers had made it hard to measure things across the board, and also not making the most of the powerful hardware architecture.
Fascinating. Just for fun, comparable compute (by cpumarks) to your 13900k for 30 fps would be 39,600+ 96-core AMD EPYC 9655Ps. Bravo on the project success!