Patrick Wunderlich
Anil DashYou must picture Sam Altman holding a knife to Tim Berner-Lee's throat. We are in a truly existential level of danger when it comes to the survival of the open web, across every front. This year is when it all comes to a head. https://anildash.com/2026/03/27/endgame-open-web/
rogue_corq@anildash If Altman touches a HAIR on Tim Berner-Lee's head, I will peronally hunt Altman.
elle
gmoke
Sriram "sri" Ramkrishna - š¼@anildash a lot of this is a complete lack of regulation of AI by world govts. But eventually the content trade consortiums like MPAA and the RIAA will be putting in lawsuits. Because it is far too easy to make content trained on copyrighted material
zl2todTim Berners-Lee interviewed on Radio New Zealand yesterday:
When he launched the World Wide Web in December 1990, how did its inventor, Tim Berners-Lee, think it would all go?
His aim was to create a democratic network built on good principles - which was free to use - where anyone with a computer and internet connection could freely share ideas.
But decades later, the World Wide Web has taken on a new life - and has been optimised in harmful ways.
...
2qx š»Of the organizations listed, only The Internet Archive is able to list a plain pay-to-public key hash style bitcoin address.
WHY? Because someone put a gun to everyone's head and told them they could not accept money.
If they can't accept internet money, then yes, the internet is lost.
Anyone that wants the internet and money should put up a donation address or accept that the endgame is lost.
An "Error 402" is one way to deal with AI bot abuse. It could go a long way toward fixing content and saving journalism too. It could simplify supporting good organizations and open source projects as well.
But if people aren't going to implement those solutions because they're scared, then the open internet never really was what people claimed it to be, and we were only building tools for people to be oppressed.
Alfredo James Pacino
Toni Aittoniemi@anildash Western people had it too good for too long. From the inside it looked like #power was not what mattered
#Technology companies talked in the language of co-operation, while the power that enabled our ivory tower was directed towards the Global South
It was a ruse as old as time: Oppressors donāt take over strong systems built by egalitarians by force, they take over them by #deceit. They adapt the culture & speech of cooperation, then twist it to appear they were always there
v@anildash agree with your diagnosis but not the prescription. Conditions that brought about the seemingly impossible open web do not exist now that it has matured. I dont know what collective action could even be mobilized toward specifically, vis-a-vis preserving what's been, but even that is largely moot imo because the userbase of the modern internet is such a different set of cats to herd. (1/3)
@anildash My expectation is a new, less centralized internet will form. Possibly not http or even tcp/ip, almost midway between internet and sneakernet. More reliant on human networking to establish collaborations - making this slop spamming obsolete - and more dependent on geography for content discovery - making scraping, malicious or otherwise, largely obsolete at least in the way its been done before. In short: less open, but more trusting and more human
Edit: I am picturing meshtastic, but it could be something similar or a combination of things
(2/3)
@anildash the start of the open web wasnt something I was around for but I imagine it was a core of smart nerds inspired to be part of something new and different who created an inertial snowball rooted in openness, eventually pulling in a far larger number of people and hustlers until the latter blew it up and now we're a scattered mass of billions of people at the bottom of the hill, and the only scenario where we have an inertial openness core again is to find a new hill to roll it down
(3/3)
(3/3)
degenerating degenerate> Open source software projects, which power the vast majority of the internet's infrastructure, are now beleaguered by constant slop code submissions being made by automated AI code agents.
I think the era of trash submissions in order to try squeeze some prize money some of the time, has gone.
AI-powered vuln reports I get on my FOSS project for the last months have been accurate, scary and valuable. It will vastly improve security, if you survive the next months un0wned.
DÅŗwiedziu@hopeless
Curl shut down its bounty program, because of those slop squeeze reports.
So your experience is not an universal one.
https://daniel.haxx.se/blog/2026/01/26/the-end-of-the-curl-bug-bounty/
The end of the curl bug-bounty
tldr: an attempt to reduce the terror reporting. There is no longer a curl bug-bounty program. It officially stops on January 31, 2026. After having had a few half-baked previous takes, in April 2019 we kicked off the first real curl bug-bounty with the help of Hackerone, and while it stumbled a bit at first ā¦ Continue reading The end of the curl bug-bounty ā
Put this another way: projects that monetize bug reports find it attracts flies. When they stopped waving money at the problem, the problem went away. No money on offer in the first place? No flies.
Also the reason I say that era is over, is that the AI tools were markedly less good 6 months ago, during the time of the bug spam. Now, they are producing on-point, terrifyingly accurate results. Maintainers receiving these are too busy shitting themselves to complain.
@dzwiedziu @anildash Could you calm down a bit, mate?
At least some of the "security researchers" handing me red hot potatoes are paid for by a FAANG company. They're not just doing it for me because I'm so likeable, or because they are either, but I would assume for anyone producing code they're are shipping. And they ship a lot of liberally licensed FOSS.
The people handing the bombs out are humans... they seem to have invested in their own pipeline using commercial AI and are doing well.
@hopeless
Okay āmateā, if you're trying me to pin me as someone not being calm when asked for data I will just walk away and assume that there is not a *broad* stream of *valid* slop-generated bug reports.
(So we're clear, because I'm feeling that I have to: that does not deny that there is a broad stream of *invalid* and bad-faith slop-generated reports.)
Bro... I certainly don't care what you believe... you are very welcome to continue to believe whatever makes you happy.
Of course if that's unrelated to what is actually happening in the world, it makes it pointless to talk to you.
Have a nice rest of your day.
Sergio R.-SolĆs@anildash @srsolis quoting the linked article: āTim Berners-Lee is no billionaire, but none of those guys with the hundreds of billions of dollars would have all of their riches without him. And the thanks he gets from them is that they're trying to kill the beautiful gift that he gave to the world, and replace it with a tedious, extortive slop mall.ā
Bad News Nobody@anildash Sorry but there's no "good AI".
SamuelJohnson@dukeboitans @anildash You're wrong.
F4GRX SĆĀ©bastien@samueljohnson
You're wrong.
@dzwiedziu @dukeboitans @anildash I'm not wrong. Any assertion that there is no good AI is balderdash.
https://www.nature.com/articles/d41586-024-03214-7
Just about all technology has the potential for misuse. Regulating it is the appropriate response not wishing it away.
@samueljohnson
The most misuse here is the usage of the term āAIā, which lumps useful technology, like deep learning neural networks with deliberately overhyped definition of āAIā, meaning generational āAIā of LLMs and diffusers.
So āno good Ā«AIĀ»ā is a reasonable answer to people who want to eat the world for āAIā datacentres.
@dukeboitans
Butā¦ butā¦ what about the mob AI from the OG DOOM? ššš„ŗ
JK, only good āAIā is āoff AIā (also for me being able to finish any DOOM).
Mateusz š³ļøāš@dukeboitans It depends what do you mean by āgood AIā. AI is a marketing term and I keep avoiding it. Maybe itās time for me to start explicitly explaining why people should stop saying āAIā.
You have to be precise: do you think thereās no good LLMs? No good applications that use LLMs?
Also, please explain how do you assess if itās good or not.
hansup@anildash Is it so black and white? Things are definitively changing in a big way. But isn't there a large playing field between full evil AI and no AI?
The AI hype is definitively super cringe, hilarious and dangerous in many ways. But radical opposition doesn't seem to be very helpful. In real life many of us do not have the privilige to say simply NO to AI.
Maybe not an endgame, but another level where we should try to keep on playing and deal with & tame any kind of new monster.
Kat S@anildash I read all of that, nodding along, only to be hit with the punchline of "good AI"?
Dash
Scott D. Strader š@anildash They treat the open web like they treat the environment. Theirs to conquer and destroy and then define the rules of access for others.
Paul@anildash
In such picture I'm legally allowed to shoot Sam Altman and will do so as soon I have a clear shot.
And then Zombieland rule two: double tap.
leftofcentre@anildash calm your tits down..
Den of Earth