The danger here is control: total and absolute control of who will be able to see what, at any given time. Once upon a time this was a hard problem, not anymore.
I don't see any proposals yet that will actually work for what they say they'll do.
By all means, if you have an adversarial relationship with the state, fight every expansion of its power tooth and nail. But conversely, if you think that everything is going basically fine and the police are reasonable folks overall, then giving them more power doesn't look like a slippery slope straight to dystopia.
The control bit is the dangerous one, sure.
Kids CAN absolutely bypass a lot still. Jump on Nostr and nobody can stop anyone -- a server can be run on any device (mine is on my phone), and it'll even run on non-TCP/IP networks (FIPS is in early days and still largely runs on existing infrastructure but absolutely doesn't rely on it staying there). Give it another few months with some LoRA build out (and enough motivation for the kids) and even the ISP gateway stops being a filter point. Not that ISP level controls have ever really been the most effective anyway, given the existence of VPN's and other tunneling methods (to say nothing of public wifi networks).
We see this happening in real time as India tries to stamp out Bitchat. The codebase has appeared on more and more repositories, and the most recent version of Bitchat is capable of transmitting itself for install to other devies without relying on any server or Internet at all.
So should we push back? Of course. At some point we would need to worry that they'll say enough is enough and make the hardware itself inaccessible. We're a long way off from that, and it'd be a surefire way to get a LOT more pushback.
But in the meantime, we shouldn't despair too much. The cypherpunks that gave us private and secure access on the Internet used by everyone today despite being once illegal to export didn't go anywhere. They're still here, kicking down every wall that gets put up, or at least pointing out that there are multiple ways around (or through) it.
https://news.ycombinator.com/item?id=49080929
There is discussion about LLMs, models, CSAM, undressing people with AI, and posting those images to places.
No handy wavy stuff, makes this problem go away due to filtering at creation. That's because, open models are a thing. So you either ban all open models, or maybe make it a crime to publish uncensored models, or you punish what individuals do on the internet.
Slander and libel and defamation are a thing, and have been a thing forever, millennia. Suddenly, on the open and free internet, for barely an eyeblink, anyone could say anything with zero liability and responsibility.
Then, 2010s came. Then, doxing of people we didn't like, or didn't agree with, became OK. So suddenly, it was OK to "out" a person, but not to just force everyone to be de-anonymoized. In other words, unpopular opinions? Yeah, that's OK to remove the cloak.
And with AI, everyone and their dog will be able to render images of anyone they want naked. CSAM. Political lies, made up videos, slander, libel, literal wors in people's mouths, on and on.
What's my point?
Well, just as 2000 brought a new world to people, 2025 is doing the same.
This is only going to get worse. And I fully agree, the dangers for "control" are going to exist. But the dangers of people posting all of this stuff, has hit the point where a lot of people are uncomfortable with it. And if you think Grok creating CSAM is bad, then you must think anyone creating CSAM is bad. Or anything. So how to you prevent the harm?
Well really, you either heavily restrict software. Open LLMs, or?
You make people posting this stuff liable.
The world is at a fork. Everything will change. Because banning software requires immense control over hosting software, but so does deanonymizing the internet. And really, this is inline with how the political and societal landscape changes.
It's only been 20 years since a lot of people were on the Internet, and only 10 years since I'd label it 'mainstream'. As in, even grandma uses it. And so, now the laws are coming, as legislation and law and change takes time for politicians, and political bodies.
We're here. Right now. This is the moment.
So I 100% guarantee that absolutely nothing will stop the path to liability for slander, libel, and more that is coming. Nothing. Whatever it takes, you can be sure it will happen.
So how will it happen?
Well, by knowing who to sue.
Like it or not, that's the truth. And no, I don't like it myself.
It is trivial to demonstrate that this is incorrect (despite sounding plausible). There's plenty of incredibly toxic behavior to be found all over facebook.
Among other considerations it's worth noting that most of the usual social feedback mechanisms are missing from typical online interactions.
> The cost of running LLM botnets is going down.
Agreed that this is an existential threat to anonymous discourse as it is typically implemented.
And yes, anonymity should be the default because otherwise the alternative is that you need to identify. If not, most users would stay anonymous because it is simply the smarter decision with the exception that you seek exposure.
You can do that already, Facebook is pretty close to exactly what you're talking about. They do a lot of work to identify real people and de-prioritize spam/bots. You can configure your account to only be able to interact with real people if you want. Many people like anonymity though, it's important for speaking truth to power and discussing topics which are outside the Overton window.
There's a good reason that many of the United States' founding fathers used pseudonyms when writing. Anonymity and real freedom of speech are foundational to democracy. If you can put someone in jail for saying something that the government doesn't like, then the government will be able to steer discourse towards outcomes it deems acceptable, even if those outcomes subvert the will of the people.
There's no real need to have the government force all websites to always authenticate all humans. The only real argument for such a system is to move us toward dystopian authoritarianism.
Teens who want to see porn will be able to see porns, but they will learn somethings along the way to do so
This is caricatural of course, but it s just so dumb as law as enforcing age verification "anonymously" is just like a don quichotte battle at AI era
Kids these days have it too easy, we need stronger protections or the next generation will make pathetic hackers.
How will teens overcome the Android SafetyNet?
"Stop killing the internet" unfortunately means nothing at all to the layman, they will still be using it pretty much exactly as they were before, and governments will just claim "See, internet is still working".
And these authoritarian politicians will destroy the internet with these actions. Maybe they will even change society as we know it, and not for the better.
The problem is thinking such initiatives will achieve anything in a bureaucratic system devoid of honest debate and actual democracy.
These citizen's initiatives must at least be tried but simultaneously we must create a new internet forever out of politician reach.
I think it's time that "please let 1800 analytics firms monitor your every browsing habit in exchange for this content" was seen in the same light as "please let 1800 private foreign espionage and propaganda agencies spy on you".
Like, yeah, it was an accurate statement, but total shit if your goal is to get people up off the couch and into voting booth. Way too abstract and high-minded
Is the Pirate Party supporting Age Verification now?
Not without either being able to trivially let others use your proofs OR ending open computing. There is no zero knowledge proof about the slab of meat sitting in front of the screen.
"Stop overcontrolling and overregulating every aspect of our lives!"
Edit: the initiatives themselves also have shockingly low numbers of signatures. A few thousand. A million are required for the EC to take notice. This system seems a bit shitty. But maybe I'm misunderstanding it.
1) buy cigars and pipe tobacco (but not cigarettes!)
2) Buy "CBD" by the pound
3) Buy ammunition, a barrel and the rest of gun parts for a glock, except a frame/receiver that can be 3d printed in a few hours to complete it.
4) Order and have delivered wine
But there are people arguing to skip all that and go to straight to ID for social media. It isn't about the children. It's about keeping tabs on who says what.
Are you guys really in favor of giving governments and police corps more control and more power so that you are observed, monitored, restricted because "the internet is dead", "think of the children" and "you need to know who to sue"?
You guys are joking, right? Right?
NO. It really shouldn't. I will not comply. Porn sites can require this shit, but not the rest of the internet. Social media, dating websites, that's up to the users. If kids are on there, that's on the parents. Do your job, be the parent. I'm not going to let lazy good for nothing parents ruin the internet for everyone else.
Not sure how I feel about that, sounds kind of problematic in retrospect. Especially since there's no mechanism for triggering EU-wide referendums to help this AFAIK.
[0] population of the EU: 452 million, unique signatures required for this: 1 million
There is a real way out of it -- the governments can have their cake and eat it too. I published several academic papers to prove it, and I'm open to building and deploying it. Would anyone want to work on this together:
Only massive data gatherers should be able to track and know everything about people
i.e. drone swarms replacing ISPs?
I'm reasonably confident that in the relatively near future, the Internet will be so overrun with bots and other worthless "content" that people will flock to platforms that confirm IRL identity.
Government driving changes to laws, imposing fines on platforms which create EULAs resulting in technological controls (or alarm flags)—whether you can bypass it or not—leading to irreversible bans and prosecutions. Growing disenfranchised. Those who do not materially support the system are left out, and the laws prevent new players. /imagined dystopian future.
and yeah, LLMs kill it faster.
One of my (probably dumb) ideas is:
- Create a legal presence in a country that does not require being dystopian, relaxed business laws, less regulation, lower taxes, move most of the corporate side of the business there. Take the tax revenue with you. It does not have to be all at once. Make the move a slow boil so they don't see it until it's too late.
Which countries, provinces, regions, states have the least regulation, least dystopian behavior and the least taxes? There has to be a incentive for business leaders and board members to move things around. Bonus if the saved money pays for a Gulfstream G650 ER or two.
"No age verification" is even deader-on-arrival. Parents are literally screaming at politicians to protect "the children" (as always), but pretty much everything that has been tried so far to silence those screams has failed (remember PICS, the Platform for Internet Content Selection: that was the last serious contender, and it has been dead for, like, 20 years). Again: demand a way to implement this securely and anonymously, don't take an untenable abolitionist stance or resort to hand-waving in the vague general direction of "parental responsibility": that has been tried, and it's not working anymore.
And, please, stop vilifying the EU in these matters. The UK left the EU, ostensibly so they could make "more sensible laws" in, specifically, respects like these, and they're now leading the assault on basic freedoms with even more fervor. And no, that's not just big-bad-Starmer's fault: it's a preview of what you have to argue against and propose a compromise for instead.