[Air-L] Migration to Mastodon

[Michael Ruigrok]

Be warned, these are the thoughts of a long-time Mastodon user,

Regarding Mastodon as an alternative, I think it would be a unique
opportunity if the academic community were to become involved in the
direction of the platform. As an open and community-driven technology,
Mastodon can improve directly from the input of its community. Points like
Emma's can have actual weight in shaping the possibilities of the platform,
especially considering that the platform's end goals correspond closely
with supporting vibrant, inclusive communities. I think it's critical that
these possible issues and improvements are discussed, both within this
community, and between us and the Mastodon community at large. Open-source
projects like Mastodon do have limited resources, so individual wishes can
be difficult to meet, and progress can be slow, but wider discussion and
support does have impact.

This openness applies equally to shaping the platform to support
opportunities for research. Given that the network of Mastodon is
predominantly run by people dedicating their personal time and money to the
mission of an ethical, open social technology, it's likely that they are
more than happy to collaborate closely with our goals. While the
decentralisation of Mastodon certainly presents a challenge for data
collection, sharing the load of infrastructure may provide an environment
where open data becomes a more sustainable, albeit smaller scale affair.

On a more idiosyncratic, but equally idealistic note, I'm fascinated at the
possibility of leveraging the platform for more direct HCI research. Since
Mastodon is part of a decentralised ecosystem, it would be possible to
create a Mastodon server integrated with the existing social network,
change one or a set of features, and get concrete metrics to how specific
changes influence behaviour in a real-world, yet controlled environment.
It's something I'd be interested in testing the feasibility of.

I do have some more specific thoughts on democratic discoverability on
Mastodon, but I think I'll save them for that Mastodon thread. I'd also be
keen to add anything relevant from these discussions on the Github feature
request thread: https://github.com/mastodon/mastodon/issues/14918.

Thank you for the compelling perspectives!

Michael



On Mon, Nov 7, 2022 at 8:17 PM Shulman, Stu via Air-L <
air-l at listserv.aoir.org> wrote:

> Andrew,
>
> I am solely responsible for the interpretation that firing the small
> academic Twitter team (along with other service and content oriented teams)
> is a precursor to ending the academic program. There is no reporting I have
> seen to that effect. It was based on private conversations. I would be
> pleased to be 100% wrong; the program has been amazing and I love working
> with academics on the data. However, the program has specific compute costs
> associated with it. Currently academics with credentials can pull undeleted
> and unprotected Tweets from the complete history of Twitter for free. That
> query and the data pull itself are not frictionless. Searching many
> billions of Tweets and then retrieving the matches when there are 60+
> fields of text and metadata and hundreds of millions of rows of data per
> day, over more than a decade of Twitter online, is costly. I run one rack,
> with one server, and one disk array, holding about one half billion Tweets,
> and the electricity bill is a real cost of making our tools free for
> academics. When Musk is done kicking people off Twitter for not labeling
> their jokes as parody (so much for free speech leadership) he may look at
> the cost of the free data access for academics, or anyone else, and choose
> to curtail or monetize it. Will the Twitter Search API remain free and or
> operational? Nobody knows. From what I have heard from folks still inside
> and recently departed, some of the critical functions of the platform are
> currently under- or unstaffed, not just the academic program. Certain key
> people who are bearers of institutional knowledge about what keeps Twitter
> servers running, insanely complicated and in parts aging tech, are now
> focused on planning a group trip to Disney with the buyout cash, which, not
> to go too far astray, appears to be one of the largest money laundering
> operations in history. My point was if you have the academic credentials,
> have not used them to the full extent, and you have a PhD thesis or
> scholarly publication dependent on that access, the program is unstaffed,
> the group in charge of Twitter is fretting about losing $4M US/day,
> advertisers are bailing out, and we are entering what some political
> science and history professors call a historically contingent moment with
> potential for a major ideological realignment or worse. Democracy in the US
> is under specific and well documented threats and some rightly say social
> media is an enabling factor for authoritarianism and dystopian politics. If
> we go into 6 weeks of civil unrest over election denial and another bigger
> and better violent insurrection is organized on Twitter, does anyone think
> Elon Musk will want academics or journalists fully empowered to document
> the role of weaponized Twitter functionalities in that? I am an election
> worker. People are making violent threats on Twitter about election
> workers. The folks in charge of regulating that "free speech" are now gone
> or have diminished resources. Public statements from the Trust & Safety
> team aside, on Twitter, you can call for the assasination of political
> leaders all over the world, the killing of vaccine advocates, mob violence
> against election workers, blatantly false election denial, and some other
> entirely anarchic, anti-Semitic, and racist stuff, and that was all before
> Elon fired everyone who was responsible for keeping a lid on such things.
> Twitter is a loaded weapon. There is no Board of Directors. Literally
> anything could happen with no check or balance. I remain of the view that
> Monday will be one of the strangest days in the history of the Internet.
> Truly, I hope I am wrong about all of it.
>
> Stu
>