1.0.0.1 is also down.
Many-many people remarked that this is the wrong approach in todays world, where CPUs are good at doing dynamic dispatch prediction, but the cache sizes (esp. L1, and instr cache) is very limited, for most code (with the exception of very hot tight loops), fetching code into cache is going to be the bottleneck.
Not to mention, for a systems programming language, I'd expect a degree of neatness of the generated machine code (e.g. no crazy name mangling, having the same generic method appear 30 places in assembly etc.)
That is I don't want google/amazon/samsung/apple to control my house. Most border routers are also connect to our smart home system. (there are exceptions but it isn't clear if they are better)
For what you're looking to do in principle you don't really need any of this after the initial commissioning. So long as the radio waves can reach the devices they will be able to talk to each other.
Although we don't know how 5050 will perform, 50 series have roughly ~same perf in render as models from 40 series at the same tier. 40 series in turn are only a mild bump over 30 at the same tier. And 30-series was a reasonable improvement over 20, but mostly in perf/$ measure and not raw perf. Extrapolating, 5050 is likely not going to give much of a boost if any, and spending money on a 8GB card in 2025 is just throwing money away at this point as software is now increasingly expecting to be able to work with >8GB of VRAM.
(That said I'm all for more dystopian stories so we can get past this AI-replacing-coders fad)
As another data point, the electricity price was already ~16 cents/kWh 10 years ago and ~12c/kWh in 2009. High despite trading with Russia being a thing back then.