They replied to me they did not see any legal problem with notarization.
We need a DMA 2.0 which address the oligopoly of dominant Operating Systems, including the freedom to install alternative OSes (no more signed bootloaders, proper hardware documentation, etc...).
"Purely based on the MX-records, we learn that 72% of Belgian municipalities run Microsoft mail servers and 60% of the Dutch municipalities. For Scandinavia, it’s 64% in Norway and 57% in Sweden. In Finland, it’s a whopping 77% if the cities that are being served by Microsoft."
It doesn't need to be thousands of developers. Whoever owns the project can use their machine for builds, or spend a few bucks a month on a VPS, or find a sponsor to pay a few bucks for them or provide their own machine (should be easy if the project is any size at all).
My theory is that we waste lots of resources on CI because it's invisible. If you can hear your fans spin up you might try to optimize more, even if the financial cost is still negligible. Back when every developer built the software and ran the tests on their own machine, slow-building projects were the exception not the norm. There was also a much faster feedback cycle for changes to the build and test processes.
Tooling around a self-hosted CI workflow really needs some work to make it as convenient as even the very difficult GitHub Actions.
> make a WASM slave and thousands of developers can give their CPU/Memory/Disk for build slaves
WASM isn't a magic bullet for sandboxing. CI environments assume a full Linux. So you need to either ran a VM (with the attack surface that implies) or a write an x86 emulator in WASM (which would be very slow).
You also need anti-abuse to stop bitcoin miners from using your system. GitHub probably have full-time employees working on it.
> Like bitcoin mining, there could be some competition between 3 parallel builds to pick the winner if the output is the same.
It's a lot more complicated because many builds are not deterministic, you need to store artefacts, build secrets, etc.
Companies like golem.network or iex.ec have been working on this problem for a decade and they are still not easy to use.
to be fair there are some degree of "hand curation" of the data so while "it is the internet", the actual trained data is a derivation of that.
in a mild but productive analogy:
I could actually hand a K&R book C programming book + lots of specs to say "this is the linux source code" (the raw data that were all observations were made, aka "the internet")
...or just send them the "kernel the source code" (the refined training data, after a LOT of manual stuff)
... that your compiler consumes to generate the kernel. (the Open Weights model, what they actually shared)
Mildly related rant: honestly its a bit shit to say "open source model" in a "open weights" model, its like saying World of Warcraft is opensource because they gave you an executable of the game. (you can still change it, but in more restricted ways)
There are other countries in the world.
And there are also places where there is no electromagnetic policies (think about over the oceans).
reply