Re: Meta Unspools Threads

Welcome, Ben!

(It will take some effort to ensure others keep feeling welcome too.)

> What can Meta do to support the fediverse?

I'll give it a try. In the short term, Meta Platforms Inc. can support 
the diversity and health of the fediverse with financial contributions 
showing a long-term commitment while actively giving up influence. To do 
so it needs to cooperate with something already existing on its own.

Just some examples come to mind:

1) Software developer support. Donate 100 M$ upfront to NLnet to run a 
10-year program similar to what they've been doing in this space, 
without any requirement other than FLOSS licensing (and perhaps 
non-profit status for projects/instances receiving support for 
operational costs). It's easier to federate with thousands of nodes if 
you're more confident about their cybersecurity.

https://nlnet.nl/foundation/ANBI.html

2) Legal support. If the userbase explodes, there will be a deluge of 
lawsuits and compliance issues for thousands of fediverse admins. This 
is tricky to say the least, but maybe donate up to 10 % of the existing 
budget to orgs active in legal support (like EFF, ACLU, Digital Freedom 
Fund). Find a research centre à la create.ac.uk able to willing to 
produce or commission legal handbooks and practical templates for 
fediverse admins in 100 jurisdictions and cover the cost.

3) Education. Ask the HEERF how to match their support to HBCUs, which 
provide much-needed diversity in STEM in the USA.
https://en.wikipedia.org/?oldid=1161896053#cite_ref-87

> How can we ensure our entry to
> this ecosystem is a positive thing that helps grow the community?

No idea but you're right that moderation plays a big role in the puzzle.

Has Meta Platforms considered running more than one node? For example 
one by country, with jurisdiction-specific rules. Or even one by 
county/province in the bigger jurisdictions.*

> How can
> we support this standard?

While not necessary to be standard-compliant, a commitment to opensource 
in all the implementations would help. It's much easier to understand 
how a standard works in the wild, and to evolve it properly, when the 
implementations can be studied.

Even if the main application source code can't be shared at the 
beginning, it would still be possible to open up individual components, 
especially those which have spillover effects on federating nodes. For 
example you could start from whatever is going to be the actual 
ActivityPub endpoint and with the classifiers used internally for 
moderation etc. (That could include working with the industry and 
researchers to publish PhotoDNA, GIFCT etc. datasets in a suitable 
privacy-preserving way that allows them to be studied in the open and 
potentially used by self-hosted pure-FLOSS software.)

Cheers,
 Federico

P.s.: My first post here. (Second try after subscribing.)

It was supposed to be a short reply but here we are. The list archives 
state that "Anyone may read or write to this list." Please let me know 
in private if it's inappropriate.

Cf. https://fosstodon.org/@blake/110698633595942360

(*) Apparently that would mean up to 700k users in California and DC. 
That's a bit less than the current biggest fediverse nodes, though far 
larger than I've personally considered advisable so far.
https://en.wikipedia.org/wiki/County_(United_States)#Statistics

Received on Wednesday, 12 July 2023 08:00:35 UTC