Meta won't federate anyone. They don't want to host illegal content on their servers, that would be an absolute PR disaster for them. And this is what will happen if they federate with a random instance. Even clean instances will want to play tricks on Meta if federated.
They are the prototype of the mono instance federation. They want control. They want to attract the people leaving Twitter. I don't think they care about us, what they want to avoid is that our instances become too big and start to offer an alternative.
Be patient, if you want your content deleted you will have an efficient way to do it soon. Don't count on reddit to do this for you. It seems we have enough access to the API to do it ourselves.
What’s so bad about giving AI models something to learn on?
From a user point of view? A lot. So far the AI has made itself the champion of the creation of fake. Fake news, fake pictures, fake videos, fake history, fake identity. Do you think that the AI will be used for your own good? Do you think that your private data are farmed for you own good? I don't.
I posted an example about fake identities and fake posters on Twitter. This is the end goal. This is where the money generated by the AI will come from.
That way you could detect and address rogue scrubbers while still working with LLM creators who are open to an honest training integration. And if your company can’t really detect the difference between users and LLM crawlers after implementing something like this, well, then those crawlers don’t really affect the company as much as the CEOs would like to pretend.
Twitter and Reddit probably want to be their own LLM creators. They don't want to leave this market to another LLM. Also it doesn't take a lot of API calls to generate the content that will astroturf your product.
Anyway the cat is out of the bag and this data will be harvested. The brands will astroturf their products using AI processes. People are not stupid and will realize the trick played on them. We are probably heading toward platforms using full authenticated access.
For reddit and twitter it's also induced by the threat of AI. Twitter and reddit host a lot of content, organized, sorted, coherent. It's invaluable for training an AI and these companies don't want to let it go for free. They want control over it, therefore they are making it very hard for AI companies to farm their content. The fact that it's happening now is because AI companies are probably rushing to copy as much data as possible before laws are voted to put a limit over them.
It will be the same for the fediverse, our content will be scanned by AI's. Our content is freely visible, organized, sorted and scored. We should be careful about that. If you are not a professional publisher or a public person then you should probably think about rotating your username as often as possible.
edit: But also, with the rise of tiktok, a lot of countries are now suspicious about the soft power of those apps, and are ready to legislate against them. The EU already did, they did vote fines against them and are regularly getting money out of them. The taboo is gone, you can attack those companies, it works. They were supposed to be out of reach, but they are not.
Also there is no genius in Twitter, as far as I know they have no patent over anything. If someone manages to become more popular than them on the same principle then twitter is done. Gravity will do the rest and users will move to a different platform. People are using it because people are using it. So the model is fragile and the value is questionable.
If Reddit can bot comments then Reddit can also bot moderators. Come on, don't be lazy Reddit! Show us your leadership capabilities and come up with a solution!
"How to sabotage your community: 101"
Red Hat is taking notes.
The person speaking was Laura Nestler, here is her bio from REddit:
Laura Nestler, Reddit's VP of Community, is a global leader with a 15-year track record of building strategic, high-impact teams and scalable community systems at growth-stage startups. Nestler leads Reddit’s Community Operations team where she is responsible for defining our international community strategy, driving key initiatives for community development, evolving Reddit’s community governance model, and transitioning the team into a global organization. Prior to Reddit, she served as Global Head of Community at Duolingo, working across product, marketing, and strategy to develop community products and programs.
She is a global leader, guys, with high-impact teams! She will solve the crisis in no time, you'll see! Is there anyone among you who can claim to be a "global leader" ? No one?
They wanted out anyway, Microsoft wants control and they were using Reddit just like DJI is using reddit: they were a bit forced to follow. This is a perfect opportunity to leave and make the support happen on their own platform. Other big brands will be inspired and will leave reddit too.
