The way it works in France is that money goes to a company that collects it on behalf of all copyright holders. Its website does not offer any documentation as to how copyright holders can claim their share.
That sounds pretty shady. There's also the problem that most media generated globally is not French. Do they pretend to distribute the spoils globally?
In reality the system in these countries is pure corruption. The beneficiaries are large corporations who see it as an extra revenue stream and that's it.
Not completely. I know some french musicians who are great artists, but are not mainstream enough to sell enough records - and they do get state money to continue their art (progressive/psychedelic music, nothing tame).
> There are research models out there which are trained on only permissively licensed data
Models whose authors tried to train only on permissively licensed data.
For example https://huggingface.co/bigcode/starcoder2-15b tried to be a permissively licensed dataset, but it filtered only on repository-level license, not file-level. So when searching for "under the terms of the GNU General Public License" on https://huggingface.co/spaces/bigcode/search-v2 back when it was working, you would find it was trained on many files with a GPL header.
Links 1 and 2 have not had updates in 10 and 8 years respectively, they probably don't even compile anymore. They implement OTRv3 which was published in about 2005 and uses 1536-bits primes. As far as I know, neither the protocol nor the implementations were audited (and especially not audited recently). This is not good encryption at all.
Additionally, OTRv3 does not allow multiple clients per account, which makes it unusable for anyone who wants to chat from two devices.
I use link [1] all the time. It comes pre-compiled for many Linux distributions but not installed by default. And yeah like I said it needs cipher updates like was recently performed in OpenSSH. HN has a handful of cryptographic nerds that could update OTR in their sleep if they so desired maybe even rewrite in Rust but being cryptographic nerds they probably have no need. If the same is true with cryptographers as is with car mechanics and plumbers they probably only use plain text as mechanics have broken down cars in their yards and some plumbers have old leaky pipes due to burn out.
As a mechanic-minded person, all the broken down junk i plan to fix someday has no bearing on the state of the tools i actually use day to day
(In my case, all the old broken guitar pedals and vintage computers littering my house have no bearing on the state of my workstations and gigging setup)
> which has lead to the flagship instance being far bigger than its moderation team can handle, leading to a situation where it's a major source of abuse
Is that still true? As the admin of a small instance, I find the abuse coming from mastodon.social has been really low for a few years. There is the occasional spammer, but they often deal with it as quickly as I do.
> the plausible, strange, not-too-distant future in which AI models are autonomously running things in the real economy.
A plot line in Ray Naylers great book The Mountain in the Sea that plays in a plausible, strange, not-too-distant future, is that giant fish trawler fleet are run by AI connected to the global markets, fully autonomously. They relentlessly rip every last fish from the ocean, driven entirely by the goal of maximising profits at any cost.
reply