Skip to main content


Truly an assault on the internet we've all helped build.

- "Dear Stack Overflow denizens, thanks for helping train OpenAI's billion-dollar LLMs" https://www.theregister.com/2024/05/07/stack_overflow_openai/
- "Reddit has reportedly signed over its content to train AI models": https://mashable.com/article/reddit-signs-ai-content-licensing-deal
- "WordPress and Tumblr will sell user data to train AI models: https://readwrite.com/tumblr-and-wordpress-to-sell-user-data-to-train-ai-models/

#IndieWeb #internet #TheWeb #ai #llms #capitalism #enshittificaton

reshared this

in reply to Stefan Bohacek

As #Bluesky has been deliberately structured to (in effect) route all its content feeds through large corporations, wonder if that is their long term business plan?

Feels like any company that has access to a large collection of user-made data is now looking at how to sell it to AI companies.

This entry was edited (3 weeks ago)
in reply to FediThing 🏳️‍🌈

I was told by a Bluesky developer that you can bypass the BGS if you want to. I don’t know if anyone has done that yet.
in reply to Chris Trottier

@atomicpoet
I don't particularly trust Bluesky to be totally honest about their long term plans. Bluesky is VC backed, that means they are already locked on the road to enshittification.
in reply to FediThing 🏳️‍🌈

There’s no need to trust them. If it happens, it happens. The great thing about an open source protocol is that if it works as designed, it should be outside the initial developers hands.

But at a certain point, this is the Internet. Entities will scrape and won’t ask for permission. At least the data on my server is mine and I can do what I want with it.

This entry was edited (3 weeks ago)
in reply to Chris Trottier

"The great thing about an open source protocol is that if it works as designed, it should be outside the developers hands. "

Not in the case of Bluesky's protocol. They, a VC-backed for-profit, have structured it right from to be reliant on large companies.

They are asking us to trust large tech corporations running the feeds. That's a bad idea given the bad track record of all VC and corporate tech corps in similar situations nowadays.

(It's different on the Fediverse or similar federated networks, no one is dependent on any large corporation.)

This entry was edited (3 weeks ago)
in reply to FediThing 🏳️‍🌈

Okay, but what I’m getting at here is it should be possible to bypass large companies by not connecting to a BGS. At least, in theory—eventually. If it works, trust doesn’t matter.
in reply to Chris Trottier

@atomicpoet
You were told that by a Bluesky engineer, we have no idea what the reality will be or what BS's ultimate plans are.

This is the problem, once you put VC money into it, it messes up the relationship between provider and end user. It sooner or later makes the provider see the end user as cattle to be exploited.

It makes absolutely no sense to go down that road when we already have a VC-free alternative in the Fediverse.

in reply to FediThing 🏳️‍🌈

You don’t have a VC-free alternative in the Fediverse. Multiple Mastodon servers are owned by entities financed by VC entities. Some of the biggest servers even.

To me, that doesn’t matter, though, because my server isn’t owned by a VC. It’s owned by me.

But I also don’t look at VCs as a metaphysical evil. I just look at them as banks. Maybe banks are evil, but I still use them. You probably do too.

Some people try to bypass the evil banks. And what does that get them? They’re either storing cash under their mattress or buying Bitcoin. And we know what kind of kookery that leads towards.

You can’t avoid VCs any more than you can avoid banks. Almost everything you use has been touched by them. That includes the development of ActivityPub itself—but I’m sure you know that.

in reply to Chris Trottier

"But I also don’t look at VCs as a metaphysical evil. I just look at them as banks. Maybe banks are evil, but I still use them. You probably do too. "

No, VCs aren't banks. Banks loan money but they don't buy control of the company. VCs take control of your company and force you to do growth-oriented things that enshittify you.

Every single online project backed by VCs has enshittified.

"You can’t avoid VCs any more than you can avoid banks. "

That just isn't true. Most companies and organisations aren't backed by VCs.

VCs are looking for rapid unsustainable growth so they can cash out a large profit.

This entry was edited (3 weeks ago)
in reply to FediThing 🏳️‍🌈

Chris, it is pretty nasty to imply that those who criticise VCs are "kooks". There are serious long term problems with the structure of VC projects due to their unsustainable growth-chasing. It will kill the planet amongst other things.

It is reasonable to criticise and condemn VCs, it's not "kookery".

"You don’t have a VC-free alternative in the Fediverse. Multiple Mastodon servers are owned by entities financed by VC entities."

Vast majority of servers aren't involved with VCs. The standards aren't created by VCs.

And it's possible to defederate the tiny number of VCs on here.

This entry was edited (3 weeks ago)
in reply to FediThing 🏳️‍🌈

Here's a textbook example of how Venture Capital gradually enshittified and destroyed a social network that claimed to be an ad-free privacy-friendly alternative:

https://waxy.org/2024/01/the-quiet-death-of-ellos-big-dreams/

"The Ello founders (say they) won’t sell your data or show you ads."

"If they were independently-funded and run as some sort of co-op, bootstrapped until profitable, maybe that’s plausible. Hard, but possible."

"But VCs don’t give money out of goodwill, and taking VC funding — even seed funding — creates outside pressures that shape the inevitable direction of a company."

VC backing means enshittification.

This entry was edited (3 weeks ago)
in reply to FediThing 🏳️‍🌈

This entry was edited (3 weeks ago)
in reply to FediThing 🏳️‍🌈

"At least the data on my server is mine and I can do what I want with it."

We don't know that in the case of Bluesky. There may be some legal sophistry saying that data which passes through a feed can be legally sold to AI companies. And the people in charge of the network are financially incentivised to do this if they can.

This entry was edited (3 weeks ago)
in reply to FediThing 🏳️‍🌈

We already know that once content is published via ActivityPub, it’s out there. No take-backsies.

What I mean is whatever else is on my server, that’s mine. This important because many servers (like mine) support local-only posts.

in reply to Chris Trottier

Scraping is different from selling the content.

The original post was about companies that run public sites on the web which could easily be scraped by competitors.

But they are still selling the user-generated content on their sites, so there is obviously substantial demand from AI companies for non-scraped content.

This entry was edited (3 weeks ago)
in reply to FediThing 🏳️‍🌈

Oh yeah, there’s something distasteful about the fact that I spent 15 years on Reddit making content for that site, and now Reddit is selling my work to A.I. companies. I’m not seeing a single dime from that content! Although, in fairness to Reddit, they did offer me stock. But since I’m in Canada, I can’t vest any of it.

So yeah, I’m pretty bitter about all of that which is why I’m thankful about owning my server now.

in reply to Chris Trottier

Here’s a scary thought, though. How much data might some 3rd party Mastodon clients be gathering from us, and re-selling to AI companies? Without disclosure, we don’t know.
in reply to Stefan Bohacek

But I mean, still. BGS creates an additional layer of complexity. Only big corporations will be able to run one anyway.
in reply to FediThing 🏳️‍🌈

@FediThing Well, we already know that they will rely on ads (https://stefanbohacek.online/@stefan/110254051709008742), but this is a good point, definitely wouldn't be surprising.
in reply to Stefan Bohacek

It's not too late though. Start a blog, make a website for your hobby, join a forum. Stick with the fediverse!

Roni Laukkarinen reshared this.

in reply to Stefan Bohacek

Yeah, you can actually do it yourself! It's much cheaper and easier than people realise 👍

Managed hosting makes creating your own site about as difficult as setting up a webmail account, you don't need server level skills or command lines etc.

in reply to FediThing 🏳️‍🌈

@FediThing I recently made https://simplepagebuilder.app, just to see if making a website, for free, can be made even more approachable. (I'll leave it to others to judge.)
in reply to Stefan Bohacek

If you start a blog again, won't they just scrape that, too?