Hacker News new | past | comments | ask | show | jobs | submit login

As someone who grew up alongside the internet, I wonder if this is just part of the bigger picture of where the internet has always been heading...

The bigger, more popular, and ubiquitous it became the more corporate and political powers were going to seek to rule it. Now were are at an age of internet giants with the GDP of small countries, and political elections being swayed by the bovine herds of Facebook and Twitter users (or useds as Stallman calls them!). The internet has come so far from my happy memories of the late 90s.

My prediction is that we will see multiple 'internets'. Whether for political reasons (e.g. China), or commercial (someone like Facebook or Google providing their version of internet to a 3rd world country).

Then of course we have things like dark web. I think many will stop seeing the darkweb as a place of CP and drug dealing, and more of an internet free from regulation.

It's an interesting point in history. . . . (I'm a bookwork so please share any recommendations on this topic!)




"I think many will stop seeing the darkweb as a place of CP and drug dealing, and more of an internet free from regulation."

I think it's more likely that most people will forget that the internet was once wide open and will accept the locked down state as normal.


But how can a normie get into it? Without becoming like a trend "top10 Scary things on dark web"?


Yes, the Internet caught many "corporate and political powers" by surprise. But now they're getting it locked down.

> I think many will stop seeing the darkweb as a place of CP and drug dealing, and more of an internet free from regulation.

You mean Tor, I presume. But Freenet is still around, although it's too risky to use it, except via Tor. And there are other overlay networks. Some basically just use VPN connections (such as tinc).

And they were always about being "free from regulation". It's just that hobbyists, activists and people into recreational drugs and CP were early adopters.

If you haven't read Vernor Vinge's True Names lately, I highly recommend it.


> But Freenet is still around, although it's too risky to use it, except via Tor.

Hi, I'm a Freenet developer of the past ~ 10 years, so I'd like to clarify upon this :) (The project is still active, there was a release just this week!)

While there technically were indeed lawsuits in the US, the situation is not as black and white as "it's dangerous".

It is an anonymizing peer-to-peer network as well! What is dangerous under certain circumstances is only one of the three modes to use it: 1) Opennet, where Freenet uses random strangers as peers. 2) Darknet, where you only connect to peers you manually select, e.g. your friends. 3) Opennet with some Darknet peers in addition (I'll call it "mixed mode").

So Opennet allows law enforcement to connect to your Freenet potentially and thus analyze your traffic. Still, this does not mean that your Freenet will plainly tell its peers what you are downloading! Traffic is always redirected across a random number of peers, none of which tells the others who requested it - which provides plausible deniability. All traffic is encrypted, only the recipient can decrypt it. So you cannot just watch traffic and filter out illegal JPEGs or whatever.

What LEA did then is to come up some math and then claim to deduct from it that there is a certain probability that the illegal downloads were requested by the people they claim it came from. Their math is known and discussed by the Freenet core team, it may be addressed eventually - but from watching the discussion (not the math) I can say it should be taken with a grain of salt. It's not absolute proof that the claimed downloaders were in fact the downloaders. It's just a probabilistic assumption, which may possibly be wrong because the way Freenet works is rather complex (>200 000 LOC).

So as Freenet stores content encrypted on random user's machines (which is the advantage over Tor, Freenet is completely decentralized!), it is imaginable that law enforcement accusses people who did not willingly download it, but just happened to store it.

But: You can use Freenet in Darknet or mixed mode to be reasonably safe: The more of your peers are not controlled by attackers, the lower the probability that a statistical attack can be conducted.

Further, the said legal cases only happened in the US to my knowledge, and I'd argue that the legal system of that country seems a bit flawed. Outside of the US you can just run Opennet and probably be at the same risk as some random non-exit Tor node. You transport traffic which you cannot look into (because its encrypted) and store files which you cannot look into (because they are encrypted), so what's illegal about it anyway?


> So Opennet allows law enforcement to connect to your Freenet potentially and thus analyze your traffic.

Further, it should be clarified that this is not a problem specific to Freenet:

ANY network which tries to be anonymous will suffer from the so-called "sybil" attack if it connects to random strangers:

If an attacker runs e.g. 100 000 machines on a network of only 1000 actual users then the probability that a single user only has connections to them is very high.

And anonymization must rely upon redirecting traffic across multiple peers - but it cannot if all peers belong to the attacker.

To my understanding Tor addresses this problem by heuristics, e.g. closely monitoring important, big machines in their network, trying to ensure they are in fact distinct entities - but that is really just guesswork, not hard mathematical security.

If Tor wanted to be truly secure it would have to add a darknet mode as well.


Thanks for the explanation of the LEA exploit.

I should have been clear that I was talking about opennet mode. If you want to use Freenet in darknet mode, among people who know each other well, and trust each other, it's at least safer than (say) using a private torrent tracker. I mean, torrent traffic is also encrypted, these days.

It's true that you're relatively safe from adversaries, if you only use darknet mode. But there's always the possibility that one or more of your peers will get busted through some other exploit. And that they cooperate, and become informants.

But in darknet mode, you can only communicate with your peers, and can only access stuff that you and they have uploaded. If you want to communicate with the global opennet, and share stuff with it, at least one of your peers must have opennet peers. And that exposes them, at least, to adversaries.

If they get busted, and cooperate, others in the darknet are now at risk, because an adversary could use their client to probe its peers. They couldn't add other peers to the darknet, however, without some social engineering.

So anyway, it's whatever nodes that peer with the global opennet which are the main risk. And to do that safely, one can use anonymously leased throwaway VPS as gateways to the global opennet. You reach them via Tor. So if they go down, adversaries don't learn anything actionable about the darknet itself.


Sorry, I'm not familiar with Freenet. Why must you use Tor to browse it?


It's a ~20 year old P2P network that relies on traffic obfuscation for plausible deniability. But peers see each other's IP addresses. There's lots of child porn on it.

In recent years, investigators have been using customized clients to serve child porn, and track which peers receive it from them. For IPs in their jurisdiction, they get and execute search warrants.

Although there is arguably plausible deniability, most defendants lack the will and resources to fight. So they typically plea bargain.

Anyway, if you use Tor, they can't find you. But it's not as simple as that, really. Basically, you lease a VPS, working ~anonymously via Tor. You run a Freenet node on the VPS, and access the webGUI as a Tor onion service. They can take down the VPS, if they like, but won't know who was using it.


> It's a ~20 year old P2P network that relies on traffic obfuscation for plausible deniability. But peers see each other's IP addresses.

Traffic is not just obfuscated, it is encrypted. Sure, you see the IP of a peer which transfers stuff across your client - but you do not know what the stuff is as it is encrypted.

So the IP address is worthless unless you figure out a way to guess what the stuff is, and who requested is.

See my other reply in this thread for further details.


Sure, but some criminal investigators have figured out how to know "what the stuff is". They serve it themselves, and have hashes for the encrypted chunks.

As far as "who requested" the stuff, as you say in your other reply, they have some statistical arguments. I agree with the Freenet Project that they're very likely bullshit.

However, if you're facing criminal charges, you'd better have resources for expert testimony to discredit their arguments. And if you don't, accepting a plea bargain may be the best option. Even if you are truly innocent.


AFAIK in the cases of the convicted people, the Freenet stuff was only used to obtain warrants, at which points the cops captured computers and found accessible CP. I don't remember seeing anyone convicted based on the Freenet evidence alone.


Maybe so, but that ex Philly cop is still jailed on contempt, because he says that he's forgotten the disk password.

But generally, I'd rather avoid having a warrant served, and my stuff impounded. So advertising my IP address as a Freenet node seems like a dumb move.


Book recommendation: The Master Switch by Tim Wu. It's a few years old now but it lays out where we may be headed with the internet based on experience with previous information industries (radio, telephone, television, and film). https://www.goodreads.com/book/show/8201080-the-master-switc...


Thanks!


You can try any book of Evgeny Morozov, a opinion writer for Slate. Some of his articles: https://slate.com/author/evgeny-morozov


Thanks!


Very interesting.

If the dark web becomes popular though, won't the same people try to regulate it?

They could make .onion link illegal for ISPs to load and kill it overnight.


That's actually already been proposed in Germany. The parliamentary state secretary of the interior ministry just last month said:

"People who use the darknet usually are up to no good. This simple realization should be reflected in our legal system."

and

"I understand why the darknet can be useful in autocratic systems. But in a free and open democracy, in my opinion, there is no legitimate use."


The irony is quite delicious.


Really? Tastes bitter and a tad coppery to me.


I use Tor + VPN just to read Reddit these days :D


Isn't one of the main points of Tor that other than by knowing that an IP is an exit node, no one is able to tell you're using it? And .onion resolution is entirely client side, no?


The ISP can see your traffic going towards entry nodes if I'm not mistaken.


TLDR: You can use pluggable transports.

It's super difficult to build a low latency mix-net that covertly works inside an adversarial network.

In addition to the current models over which onion/garlic routing are based upon you would (at least) need to add to the core of your software traffic obfuscation, a series of covert channels, NAT bypass, ...

Tor and other mix networks simplify the problem assuming that there's a portion of the Internet that is free and introducing censorship circumvention mechanisms.

Of course, the assumption is increasingly untrue.


Perhaps we could crowdfund Tor node satellites?


Random thoughts: high latency, costs, vulnerability to spoofing and jamming, distribution of ODUs/IDUs, trust in the owner/manager of the link


What about (maybe) lasers + LEO, basically what Musk does?


ISPs never see .onion URLs. They just see encrypted traffic to and among Tor relays. But yes, they could block that.

In this case, it helps if the evil US military actually does use Tor :)


Other than VPNs and Tor (which can be blocked), how would people circumvent the fact that the pipes are still owned by the same players? We can't physically lay out "another internet". Maybe there's potential in radio mesh networking, but the wires we have are the wires we get. As depressing as it feels right now, I think the only real option is to continue pushing in the democratic channels we have. Governments can stop anything if they care to; we have to work with them.


Even if some trillionaire decided to lay out all the wires for a new internet it wouldn't matter with respect to things like this. It would still be subject to the same laws. The same applies to radio mesh networking.


The postal service can send a box with say 10 * 6 TB hdd quite fast. Say 2 days for delivery.

480 000 000 000 000 bits / 86400 sec = 5555.56 mbit/s

And you can send as many boxes as you like in parallel.


High throughput but extremely high latency networks are significantly less useful than current networks.


Fidonet worked and many an offshoot were spawned.

I fondly remember conversations on AmiNet - an Amiga - dedicated FTN network.

That was all done over PSTN with 9600 - 19200 bps modems. Latency was days. All this didn't preclude massive amounts of collaboration over it.

I'd rather question if lower latency delivers any benefit.


Yet, notably, copyrighted material meant for human consumption is exactly the kind of content that aligns best with high throughput/extreme latency scenarios.


>you can send as many boxes as you like in parallel

I imagine that in practice the postal service may start to decline your custom somewhere around the quintillion-box mark, or perhaps even before...


Even if some trillionaire decided to lay out all the wires for a new internet

What if a billionaire launched a constellation of low earth satellites which provided Internet?


How are satellites different from the radio mesh networks they already mentioned?

They're not, regardless of medium, the same cops and IP lawyers can get on it and track down people to arrest and sue according to whatever laws are on the books, good or bad.


They're not, regardless of medium, the same cops and IP lawyers can get on it and track down people to arrest and sue according to whatever laws are on the books, good or bad.

What if this satellite Internet service is a multinational corporation, with the directly owning entities based in Russia and China? Or perhaps Sweden? Part of a conglomerate under the ultimate control of a corporation on Mars?

I think we'll definitely have crossed some threshold when we have our first extradition from a different gravity well.


I think one piece of context that your missing is that the people targeted by this law aren't the people running the wires, it's the people running the servers behind the wires.

Today, if people in Russia or China run the servers there isn't much the EU can do (see: scihub and the US, for a long time piratebay too). You don't need a massive indestructible satellite constellation for that.

I'd also add that a billionaire isn't the criteria you're looking for here even if it was a policy targeting the people running the wire. It's a foreign government with sufficient military power to deter the US from arresting you, and sufficient technological power to set up such a network. Maybe it's a Russian Billionaire who launches them in your hypothetical, but it's the Russian government who provides the security that allows him to do that.


I think one piece of context that your missing is that the people targeted by this law aren't the people running the wires, it's the people running the servers behind the wires.

I was imagining containerized server clusters in low earth orbit as well, with the ability to rapidly export the entire state of servers across super high bandwidth laser links. Everything would be done remotely, and the corporations running them would also be in Russia, China, Mars, etc.

(Containerized in the sense of hardware in a shipping container, not Docker, though that would play a role as well.)


What if, you had a worldwide mesh network of internet satellites, too many of them to take down...?


Countries could jam the signal or find you easily since you would also have to transmit data to the satellite.


laser uplink.


Arms race against state-level actors is not a good position to be in.


Right, but not unwinnable. George Soros vs. Bank of England comes to mind as an example, but you can have other options, if, say, you have a technological edge (easier for you to "attack" with data than for them to "defend") and a sympathetic population. Doublespeak could be popular under less than free regimes.


Most content doesn't really need low-latency. If we can develop a good protocol/UI for 'hours-to-days' latency or 'offline' networking, we could just pass thumb drives between ourselves and probably accomplish all of our data transfer needs. Swapping hard drives with a buddy twice a month can be more bandwidth than most mobile contracts.


That's effectively the network setup I have. A long range radio link to a farm in the area that then has an upstream link (again via radio) to a residence that has a fiber connection.

The latency can be huge when we have less-that-optimal conditions.

Since the throughput is low and the latency high, it's not streaming-video capabl. A buddy downloads a load of movies to an ssd and drops that off every now and then.

For real-time news I use an actual radio. There are some good news and music stations in the area.

I have an offline copy of the spanish language wikipedia.


Very cool. I would love it if there was better tooling for using a kbps connection for download/searching file lists and making a 'drop drive request', so you could use your low speed connection to decide what you want on the drive, then buddy could basically just get a list of files to curl, throw em on the drive with new indexes of available file directories.




> Swapping hard drives with a buddy twice a month can be more bandwidth than most mobile contracts.

Sounds like the 21st century version of Samizdat :)


Book recommendation: 'The Future of Money' by the co-architect of the Euro - Bernard Lietaer


The internet is already by definition a collection of separate networks...


I'm speaking more about the publics perception, rather than a technical definition that most on HN know.




Consider applying for YC's Summer 2025 batch! Applications are open till May 13

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: