Let's knock down social media's walled gardens – Tim Berners-Lee
102 comments
·March 17, 2025porcoda
I'm all for this more open internet without the walled gardens, but skeptical that we'll get there simply by creating open alternatives. Not only is there a massive amount of inertia behind the walled garden networks that exist now, but you need people to build these complex things. Given how focused most engineers are on compensation (e.g., the obsession with FAANG as the pinnacle), I don't see how open things that likely don't have such a gigantic mountain of money behind it can compete for the necessary resources to execute the vision. Compensation will probably always pull the people who build stuff to the walled garden companies and the alternative will have trouble providing a competitive alternative. Plus, it seems inevitable that if some entity DOES start to build an open alternative to the walled gardens, at some point they'll feel the itch to grow, and slowly but surely I expect little tiny walls will begin to subtly sprout. Email is a bit of a bad comparison since the protocols existed BEFORE the big companies we see today, and that sort of meant they all interoperated from the start. Had we not had email in 1990 and different companies had to come up with their own messaging systems, we'd probably not see the clean interoperability we see today.
lukev
That’s why though I like ideas like Solid or the Fediverse, I’m skeptical they can be more than aspirational. The tech here isn’t the hard part, it’s the social factors.
That’s why I’m currently most excited about ATProto (Bluesky)… bootstrapping a protocol as a centralized service that has already hit critical mass, and that can now be fully decentralized on an opt-in basis… feels promising.
pfraze
What pcorda is describing is actually the premise of Bluesky, which is why it was initially funded through grants and established as a Public Benefit Corporation. The logic has been to find the appropriate hybrid of commercial and non-commercial interest which will support the work.
The initial team came from a bunch of activist projects (typically p2p) and the thing we concluded was that activism requires a theory of change. How is your new technology going to get adopted across the world? The software world still largely operates by markets. Startups are the vehicle for distribution. So we need to behave like a startup while still accomplishing the technical/mission work.
The incredible gift was the initial funding and awareness, which Dorsey deserves a lot of credit for^1. That gave us about 10 months to do protocol work prior to commercialization, which was less than we wanted but just enough to get it done.
ActivityPub deserves a lot of credit too for its progress under an entirely different model, which I tend to describe as communitarian. What both projects did correctly was connect the protocol to an initial application that drives that distribution. Whether either get to "mass" adoption (greater than centralized services) is now the big question.
^1 Apologies for teeing up the inevitable "Dorsey left" conversations.
desireco42
I looked in at ATProto and it is really well thought out and promising. Thank you for posting this.
It isn't easy to create alternative, but it should not be created as alternative to Facebook for example, but an alternative space for people to connect and express themselves. That will work more then just create another copy.
realfeel78
> I looked in at ATProto and it is really well thought out and promising.
No it isn't.
realfeel78
The tech is also hard and Bluesky and Mastodon are shit.
BrenBarn
I think you're at least partly right, which is why the solution is heavy financial penalties against large companies. The problem isn't tech, it's mountains of money. Flatten those mountains.
realfeel78
Well that isn't happening so…
bxhx7
They will be turned into public utilities. Especially messaging. One country will start. Others will follow.
After which majority that doesnt need the levels of constant Attention the social media addicted free content manufacturing minority does will leave. And social media will face major issues.
Cuz when ever growing Content chases finite non growing eyeballs you cant keep increasing Ad prices every quarter and report profits for ever. This is zuckanomics which is utter shit.
Analysts are already seeing those limits being hit but since they havr seen time and again when limits get hit corporate robots like zuck then do layoffs, buy the competition, capture regulators, avoid tax, sell more personal data, inject more ads etc etc they still see scope for parasitic growth.
But parasitic growth comes with predictable and unpredictable cost. And they will build with time. Just like Jurassic Park. And we know how that story ends.
Human Attention cannot be exploited at population scale or there will be chaos.
dkh
> Human Attention cannot be exploited at population scale or there will be chaos.
Related and very relevant paper: https://hdsr.mitpress.mit.edu/pub/ujvharkk/release/1
Attention is being exploited at scale. It's continuing, and evolving, and might soon manifest as other things (like intention).
I am not sure your predictions are accurate, but I at least would like a couple of them to be. I have thought a lot about the idea of certain ubiquitous and permeating things like social media as utilities, or at least non-profits with a public oversight and standards/requirements to comply with that are substantial, limiting, and entirely designed around what is good for the public rather than what is good for a business. Not sure this will ever actually happen, as time goes on, I only find more arguments for why it should, not less.
JumpCrisscross
> will be turned into public utilities. Especially messaging.
Honestly, yes. It's working for China. With America doing its thing right now, it makes sense for e.g. the EU and Canada to look at whether handing communications and social media to American companies makes sense.
realfeel78
No they won't.
rglover
It's interesting that the proposed solution to this problem is more often than not "add another layer."
The solution as far as I see it is to just go back to the zero algorithm version of social media. You follow people, you see their stuff in your feed. At best, you get recommendations of who to follow, but there's zero compulsion to follow or look at anything. No AI, no fancy algorithms, just good ol' fashioned CRUD.
Wrap the whole thing up in a nice UI/UX and, most importantly, take a 100% neutral stance on politics and other "faction building" modes of thought.
The chaos really didn't start until the algorithms (and by extension, the low-quality content they elevate) and politics came in. All of the alternatives so far have implemented some of the above ideas, but the juiciness of taking a political/ideological stance has been too attractive.
standardly
Yep. I agree. And, the last thing we need are more echo chambers... The current trend is really dangerous.
slowtrek
"One way to do this is to mandate that social network platforms follow new standards. Another is to quietly build an alternative world using these standards and let people realize that it is better — just as they did when leaving restricted online services AOL and Prodigy for the freedom of the web."
We would all have to take on the role of the evangelist in a serious way. I really love that he wrote this article.
readthenotes1
"Changing the algorithm to make it benevolent rather than polarising is easy for those who control the platforms."
What bubble does the author live in that tells hen that "benevolence" is easy?
The values of the cultures most of us have regular contact with are very different.
dkh
It is easy in a physical and technical sense, and it also probably be easy in the other senses if the power to make that decision was held by a different group of people than those who currently do.
Nobody is saying that it is "easy" to mean that there's no friction preventing it and that it's astounding that it hasn't been done. On the contrary--the fact that certain beneficial things are very easy in a logistical sense but incredibly difficult in a power/cultural/social structure sense is the core of the problem. And it's worth reminding people about.
BrenBarn
There are lots of things that are easy but still aren't done.
sapphicsnail
Am I missing something? He wants to make an app where you put all of your data and then an AI to act on it? Can someone smarter than me explain why this is a good idea?
charcircuit
>When the web started, you could make your own website so long as you had a computer and an internet connection (admittedly back then this was a big proviso). You could get a domain name like abc.com and put whatever you liked there. You could blog, and link to other blogs. You were part of an incredibly valuable thing from which you seemed to contribute a tiny bit and gain a great deal.
>That feeling of personal empowerment we sometimes call digital sovereignty has since been lost.
It hasn't been lost with the rise of social media. If someone wants to blog their thoughts they still only need a computer with an internet connection, but now it's even easier than before. They can register an X account, for example, for free compared to spending double $xx per year for a domain. Instead of having to build a whole website and finding a server to host it X makes it easy for you. For blogging X is a more attractive platform than the web is for the average person.
Just because there are aspects that are decentralized, that doesn't mean it's better. It's not like you are free to talk about anything on the web. In practice DNS is centralized. The server's upstream is centralized. Instead of making reports to X, they make reports to registrars and ISPs. If you want to fix the danger of being cut off you need to do it via the legal system as it's not something decentralization can entirely fix.
>This is the system that we already have for email — that’s why you can have an email group that includes people using different providers such as Gmail, Outlook and Yahoo.
Sure email exists, but in practice consumers prefer to chat using centralized platforms. The federated nature of email made it stagnant which made it fail to compete with other messaging platforms.
tbrownaw
> Changing the algorithm to make it benevolent rather than polarising is easy for those who control the platforms.
I probably don't want a feed where someone else decides what's good for me to see.
> When the web started, you could make your own website so long as you had a computer and an internet connection (admittedly back then this was a big proviso).
Substack, Medium, GitHub pages, Blogger, several WordPress places, etc. Or a $2.50/mo VPS and a $10/yr domain name, or some of the big cloud providers probably have things they scale down even further.
> We did not insist that we could share our Facebook photos with our LinkedIn colleagues, for example. Nor did we insist that we could use the same identity and transfer the same friend list from Instagram to X and then to Reddit.
I specifically do not want all sites to be views onto a single global space with global identities and relationships.
> Another is to quietly build an alternative world using these standards and let people realise that it is better
Has ActivityPub managed to eat the world yet?
vdqtp3
> I probably don't want a feed where someone else decides what's good for me to see.
I agree, but today you have a feed where someone else has decided what's good for THEM for you to see. Neither are good.
ChrisArchitect
Thing is, for awhile now, "social media" is the RSS feed. The social networks are the feed to access the content in the various gardens beyond the wall. As in the POSSE approach. Happy for more ppl to publish on their own sites but we need the social networks to provide the launching point, the syndication point to get to that content. And as along as the networks offer that "following" view so I can see most of the content from those I choose to follow, I'm also ok with some algorithm showing me other stuff related/beyond/curated. The network effect is a bonus and keeps the fresh content and viewpoints beyond what's immediately in front of me and that's healthy and useful somewhat.
jsemrau
If only Google would stop indexing social media and start linking to blogs and sources of real people the world would be better off. Maybe we need a humanity score for the search results and not an engagement score. Do I really need to see Reddit, Quora, Twitter here?
jijji
The most ironic thing about this story is that 1) it is a story about the problems with walled gardens, and 2) it is hosted on a site that actually is a walled garden and will not let anyone read the article until you pay them
https://archive.is/4Vvms
Related Tim Berners-Lee Wants the Internet Back (24 points, 4 months ago, 10 comments) https://news.ycombinator.com/item?id=42171823