Firmware build times can be comically bad. Most firmware builds are heavily optimized and you often build the same codebase dozens of times with each build optimized for one particular hardware configuration of many. You then get to deploy to a legion of test devices (also managed by machines you remoting into) and see which special snowflake you managed to break this time (it somehow is never the one you sanity checked your changes on during development :-D).
It's common to throw (a lot) of compute at the problem, and at some point it's way nicer if the 100+ cores you're cranking to 100% CPU for a few min are somewhere you're not sitting next to. It's also nice that those resources can be shared among many devs so they don't sit idle most of the time.
So, you end up with the thin client pattern. It's usually practical to build locally as well for a few particular targets, which might be a random dev board at your desk. But if I have a remote build system with a few hundred cores cores and a few dozen TB of ram, why would I not just use it instead of using my laptop?
Athough firmware and hardware is such a huge, varied field many people likely have different experiences. Brining up a big custom SoC is way different than bringing up a board with an FPGA and all off-the-shelf stuff.
What's "far more than ssh to some remote server" mean to you? It's always fascinating how different everyone's experiences are.
Every time I'm forced to use MacOS again, its worse than the last time. Everything feels like a bolted on afterthought in an OS that forces Apple's opinions on everything. My productivtiy is ruined trying to do things the Apple way, instead of the way that works best for me.
No thanks, I'll stick with Linux, where I can tinker to have the OS work the way that is best for me, instead of what Apple thinks is best for me.
It is accomplishing something. When 40 different applicants are equally able to do the job, the only selector you have is "culture fit", which is where bias starts to easily kick in (race, age, whatever), and that is a legal risk.
The leetcode hoops exist to provide a provably objective measure for hiring, even though that measure is unrelated to job performance. It's purely a lawsuit avoidance mechanism.
Seasoned site reliabilty engineer (SRE) with over a decade of experience creating tools & automation targeted at reducing toil, auto-remediation of faults/failures and delivering actionable insights to provide measurable reliability improvements at scale. Seeking out new & challenging problems to solve via code & collaboration.
I've thrived in startup-like environments, where new initiatives have created a strong demand to create infrastructure & tools where none existed before. Additionally, I've made a lot of impact focusing on pre-existing, more mature infrastructure by focusing on process and design improvements to achieve better performance, reliability and insights.
whatever elon says it is. or more accurately its inversely proportional to how oversold the region is. for much of the west coast, its very not good, especially at peak times.
