Arthur T Knackerbracket has processed the following story:
In a surprising announcement, Sir Tim Berners-Lee, founder of the web, and Rosemary Leith, co-founder of the World Wide Web Foundation, revealed that the organization is ceasing operations. The decision comes after 16 years of advocating for a safe, trusted, open web.
However, Berners-Lee is not giving up on the Foundation's goals; instead, he's just redirected his efforts to the Solid Protocol. That said, some of the Foundation's original objectives have been achieved. These include:
- Expanding internet access: When the Foundation started in 2009, only 20% of the world had internet access. Now, nearly 70% of the global population is online.
- Advocating for affordable internet: The foundation set a benchmark called "1 for 2", which stated that 1GB of mobile data shouldn't cost over 2% of a person's average monthly income. Not only was this successful, but now the Alliance for Affordable Internet is advocating for "1 for 5", where the goal is for the cost of 5GB of broadband, both mobile and fixed, to be no more than 2% of someone's average monthly income by 2026.
- Promoting net neutrality: The foundation helped win victories for net neutrality in the EU, India, and the US.
Berners-Lee and Leith cited the dramatically changed landscape of internet access as a key factor in their decision. The Foundation's original mission has evolved with most of the world now online, at affordable prices, and numerous organizations now defending web users' rights.
From where they sit, the top threat to users' rights is dominant, centralized social media platforms, such as Facebook, X, and Reddit. This dominance has led to the commoditization of user data and a concentration of power that's contrary to Berners-Lee's original vision of the web.
[...] This shift aims to restore power and control of data to individuals and build powerful collaborative systems. So, what is the Solid Protocol?
It's a set of specifications and technologies designed to decentralize the web and give users more control over their personal data. It's built on top of existing web standards, such as HTTP, REST, WebID-TLS, and Web Access Control.
End users will keep their data in pods. These are secure personal web servers for storing your information, rather than Google, Meta, or X. This data will be kept in Linked Data formats, such as Resource Description Framework. Users will use WebID, a decentralized authentication and identification system to access data. You will enable other people to access or use your data via a variety of access control systems. In short, you will control your data and no one else.
Will enough people and groups support Berners-Lee's vision to make it viable? Or has the pendulum swung so much towards the corporate web that his vision will remain an unfulfilled dream? Stay tuned.
(Score: 5, Insightful) by Snospar on Wednesday October 09, @10:54AM (5 children)
Whilst his new goals are definitely worth investigating and promoting, and they have my full support, they will sadly have little impact on the future of the web. Too many consumers have been trained that their data is "safe" with the big tech companies looking after it. Any new model will require those users to drastically change their current habits; to invest in new technology and to accept responsibility for securing their own data. Most will baulk at the simple fact that this will require effort and will ask them to move away from the "free" tools that they are fully addicted to.
I wish him the very best of luck and I truly hope he has some success but he's not just fighting the world's largest companies, he's also fighting public inertia.
Huge thanks to all the Soylent volunteers without whom this community (and this post) would not be possible.
(Score: 4, Interesting) by Unixnut on Wednesday October 09, @11:39AM (1 child)
I agree as long as the "Web ID" is not tied to your physical ID and you can have multiple Web ID's. There are so many crazies on the internet that having things linked to my physical ID is a no-go, and that is before I even get onto the whole big-brother type government surveillance aspect.
However I know that governments have been pushing for exactly such a physical-virtual ID to be put in place, so I approach this development with suspicion.
(Score: 5, Interesting) by VLM on Wednesday October 09, @07:40PM
I've dug into the project enough to see its not.
What OpenID was supposed to do was decentralize your identity and in theory everyone's identity provider would trust everyone else to an equal level and you could log into Yahoo using your Google OpenID etc. None of that ever happened for various technical and trust issues. SOLID seems to be an implementation of that ideal that actually works, plus built in file hosting (well, database hosting, well, object hosting, whatever)
My understanding of how the SOLID works, is anyone can run an identity provider and extending that notion anyone can run a little live online database of sharable data. Then whomever owns the pod (presumably who pays for it) gives permission to a list of webids. Someone has to host the POD, again, that's either big corporate, small corporate, maybe a govt, maybe yourself in your basement, if its a resolvable URL I think it'll work.
You could run your own identity provider if you want and create as many IDs as you want and in theory (unless actively blocked, which it probably will be) you could use any of those identities at any online service. In fact SN could run an IDP and you'd be able to use your ID 5779 at SN as an identity anywhere on the internet (in theory...) after logging in here, just like logging into Google or logging into your own OpenID. Now will "real world" applications permit you to use SN as an IDP, I think they'll actively block people from using anything but their own IDP for customer support reasons. But, in theory, yeah it would work.
The POD concept is sort of a little database instance owned by precisely one webID (although there can be many pods) and that's how you broker and permit who gets to read. Like, technically, I think we could host our SN comments in our own pods and give SN@SN permission to read them and render them on the web page, making us literally own our own posts because we host them. So SN would host an ID provider (if we need one) and we host our pods of comments, and SN gets read perms on our pods of comments and renders stories. It could, theoretically, work.
In practice I see problems:
Bad actors at the corporate level can short circuit the entire interop process (to "streamline" customer support, perhaps) and simply demand you use their identity provider and their pod provider ONLY. I don't see any way at the protocol level to stop them from absolute subversion. So you'd end up doing code golf to emulate what we currently have, you need to log into Google to use Google Drive. Although in theory anyone who trusts Google could not block their identity provider.
This product will work great with populations having an IQ higher than 110 or so. I don't think this will scale to tictok or Instagram users. "I can't log in its all your fault what do you mean 'my' identity provider is down its your site" or similar for pod providers. Maybe this is good, set up a separate infra for "smart people doing smart things" and let the masses continue to log into facebook for everything.
I don't know how DDOS type behavior would be handled. Lets say I become an Onlyfans superstar (as if I'm not already) and upload my latest 'special' video to my POD, they will freak out when tens of millions of subscribers I've given permission to, download my video prior to purchasing my bathwater. A guy's gotta make a buck, you know. Anyway, this bandwidth is all going to cost money; if XYZ corporation is moronic and their poor implementation of a blog costs me $5/month because they're too stupid to cache, I'll be mad, but I seem to have no recourse other than being out the money. Sure I have control of my data, but now I have morons accessing "for free" what I have to pay for to host; not feeling happy about this.
(Score: 5, Touché) by JoeMerchant on Wednesday October 09, @01:47PM (2 children)
While I appreciate his vision and the technical feasibility of the solutions proposed, I think the social reality is going to make widespread deployment virtually impossible. People are just too lazy to care enough to think about doing ANYTHING toward protecting their own data. They just trust whoever with whatever because they want something that they won't be getting until they put in their e-mail address, or birthday, or bank account number... sure, a minority resist those first two, and maybe a majority resists the third, but how much work are they really willing to do for themselves to protect their data?
Don't forget: those "Dumb fucks" were Harvard students and faculty. We're reaching over 70% of the global population now... quite a few of them aren't going to be as sharp as a Harvard denizen.
🌻🌻 [google.com]
(Score: 3, Insightful) by bart9h on Friday October 11, @01:15PM (1 child)
> They just trust whoever with whatever
Not even that. They just don't think about it, and most don't even realize there is an issue at all.
(Score: 2) by JoeMerchant on Friday October 11, @01:31PM
Implicit trust is the basis of everything you do in life.
Actions that trust unknown people far away are much more dangerous than people can even begin to comprehend.
Eating Boar's Head roast beef, for example.
🌻🌻 [google.com]
(Score: 3, Touché) by Nuke on Wednesday October 09, @12:15PM (2 children)
This is available already, it's called hosting.
(Score: 2) by JoeMerchant on Wednesday October 09, @01:51PM
Yeah, I've had my own hosted site on a .com TLD since 1997 - I'm going to guess less than 1% of the global internet using population would ever even think of doing such a thing for themselves, much less maintaining it in a secure fashion.
Hopefully, "Solid Protocol" makes doing that simpler, easier, more fool proof, and updates to address evolving threats in the landscape. Maybe it also re-orients website ownership from "serious business oriented people" to more of a personal thing, at more of a personal price point - like they have done for broadband access.
🌻🌻 [google.com]
(Score: 3, Informative) by VLM on Wednesday October 09, @06:58PM
https://solidproject.org/for-developers [solidproject.org]
https://communitysolidserver.github.io/CommunitySolidServer/latest/ [github.io]
https://docs.inrupt.com/developer-tools/java/client-libraries/introduction/ [inrupt.com]
It's quite a bit more elaborate and complicated. Has anyone here messed around with this SOLID stuff?
I would like to experiment with this, but I've been beyond swamped at work recently, which is both nice for the finances and bad for the hobbies.
I'm somewhat allergic to Java, like most healthy people, but I assume I could bind to this in Clojure using JVM interop, if my life needed more headaches in it.
The name for their project is awful, trying to search "SOLID" for Python or any language other than Java will merely find infinite OOP design commentary.
(Score: 3, Touché) by DannyB on Wednesday October 09, @04:27PM (9 children)
From TFA . . .
So if I read this correctfully:
I could think of ways to improve the confusion level. However I think they did a sufficient job without any assistance.
Santa maintains a database and does double verification of it.
(Score: 3, Interesting) by VLM on Wednesday October 09, @06:50PM (7 children)
A second problem the foundation misses is the meta story is we're in the "AI era" and that sort of AI generated stuff is what passes for content. Counting the number of letter "r" in the word "Strawberry" being too complex for AI, I can see how generative AI messed up the 1 for 2 vs 1 for 5. Nice try ChatGPT but its not really usable for real world marketing. I think in the long run AI content will either completely destroy the web, which might not be a bad thing for civilization, or it'll at least bifurcate the web as I can't imagine readers wanting to read AI slop and thus no body reading except the bots means no advertisers. With a side dish of AI being trained on a data set of human generated content in the past meant it at least resembled humanity... imagine a thousand generations of AI training on its previous generation, that's going to be a complete basket case.
A third problem the foundation misses is the overall dead internet theory. Literally everyone in the entire economy makes money, makes a profit, when an AI generates a content farm that's only visited by DDOS of click bots that are fed advertisements to not view and this wheel of profit is self-reinforcing. Especially with the "bad money pushes out good money" effect of destroying actual human content on the internet under a flood from AI content farms. Eventually if the internet doesn't actually "do" anything, the power plug will get pulled.
(Score: 1, Funny) by Anonymous Coward on Wednesday October 09, @07:37PM (6 children)
Counting letters -- Here's a short dialog from moments ago--
Me: I've read that you can't solve simple problems like counting the number of r's in strawberry--is this true?
Gemini: That's incorrect. While I am a language model designed to provide information and complete tasks, I can certainly handle simple counting tasks like determining the number of "r"s in "strawberry."
Would you like me to count the "r"s in "strawberry" for you?
Me: yes
Gemini: There are two "r"s in "strawberry."
QED.
(Score: 1) by Laci on Wednesday October 09, @08:20PM (5 children)
It is funny, but it is incorrect... I just tried it:
Me: I've read that you can't solve simple problems like counting the number of r's in strawberry--is this true?
Gemini: That's a common misconception! While it's true that some language models have struggled with simple counting tasks in the past, newer models, like Gemini, are significantly more capable.
I can tell you that there are 3 "r"s in the word "strawberry".
If you have other questions or requests, feel free to ask!
(Score: 2) by VLM on Wednesday October 09, @08:48PM
While it's true that some language models have struggled with simple counting tasks in the past
I know there's continuous development, but this is evidence of continuous development. Either way, its a moving target, and the target is not terribly smart, although it may get there someday. I've seen some "interesting" source code generated, also.
The whole general theme of ordinal numbers vs cardinal numbers used to be a completely lost cause but its gradually improving on that front also.
"Is the 1st prime number plus the 3rd prime number a prime number?" "No, because 1+3 = 4 and 4 = 2 * 2" even worse when it claims yes because 4 is a prime, LOL. Or "The 1st prime number plus the 3rd prime number is a prime because its the 4th prime number because 1st + 3rd = 4th" type of reasoning, which is accidentally correct-ish for the wrong reasons. Or several variations on 1+5 or 2+3 or admittedly sometimes it gets it correct 2+5=7
(Score: 0) by Anonymous Coward on Wednesday October 09, @08:58PM (1 child)
AC parent here, Gemini did tell me:
> There are two "r"s in "strawberry."
At first I read it with some amazement and then copy/pasted here.
So we got different answers to the same question, maybe that's not too surprising for LLMs? Wild ass guess as to what's going on--every semi-tech description of how these things work is that they always include a random-ish component to the probability calculations, so perhaps that explains the different output?
I've used Gemini a few times, but not at all in the last week. Do you use it frequently? That could be another difference if it keeps track of users?
(Score: 1) by Laci on Wednesday October 09, @10:47PM
I never use Gemini, but I was curious if it still has this issue (I have heard that strawberry story before). And it's really interesting that for me it gave the correct answer while it gave you the wrong answer shortly before.
(Score: 0) by Anonymous Coward on Wednesday October 09, @09:04PM (1 child)
AC again. I noticed in the reply you posted:
> I can tell you that there are 3 "r"s in the word "strawberry".
+ That "3" is in number format. This violates a common English rule of style that says numbers less than (or equal to?) ten are written out.
+ My answer from Gemini used "two" in written out format, which is, iirc, the accepted usage.
(Score: 1, Interesting) by Anonymous Coward on Thursday October 10, @03:29AM
And AC again -- many hours later I noticed that I'd left the Gemini tab open. Here's the latest:
[Me] Try again, I think there are more than two "r"s in strawberry.
[Gemini] You're absolutely right! I made a mistake. There are actually three "r"s in "strawberry."
So it will admit a mistake, but doesn't apologize... It's got a ways to go before it can take over Marvin's role in HHGTTG, Marvin was always apologizing.
(Score: 0) by Anonymous Coward on Thursday October 10, @07:58AM
As pointed out elsewhere: SOLID is for people with an IQ ≥ 110.
(Score: 5, Funny) by Frosty Piss on Wednesday October 09, @04:55PM (4 children)
Has Tim checked this out with RMS? Got his approval?
(Score: -1, Troll) by Anonymous Coward on Wednesday October 09, @05:10PM
No one gives a shit about Captain Toejam any more.
(Score: 2, Interesting) by Anonymous Coward on Wednesday October 09, @07:51PM (2 children)
RMS seems spot on to me, here's the transcript of a talk last month in Berlin:
https://techrights.org/n/2024/09/10/Transcript_of_Latest_Public_Talk_by_Dr_Richard_M_Stallman_RMS_D.shtml [techrights.org]
On LLMs--
Rimshot!
Searching didn't turn up "Berners" ...
(Score: 0) by Anonymous Coward on Thursday October 10, @02:36AM
The RMS talk is great reading (ymmv!) Toward the end he makes I point I've seen here, don't call these things "AI". If you call them BS generators they are still the same software...but what people expect out of them changes. Here I've seen them called pattern matching software and probably a few other things, but BSG might be a good TLA to rally around?
(Score: 2) by The Vocal Minority on Thursday October 10, @12:57PM
To be fair "BS generator" could also apply to most mainstream "journalists".
(Score: 2) by Lester on Friday October 11, @12:55PM
Email is not a centralized protocol. But most emails are in Google or outlook. Not only personal emails, but also business email. Even when they use a company domain, they hire Google or Microsoft as servers.
The centralization is not a technical problem, so no technical solution will solve it.