Maybe the internet as a whole is intractable. But perhaps we can get a meaningful digest of snapshots of the important stuff that can work over slow links, intermittent links, and one-way links? Perhaps we can mechanize the selection of important content.
The real question is how you do this in a way that doesn't encourage one-sided propaganda, sybil attacks, etc.
I'm reminded of systems like Secure Scuttlebutt and the Othernet satellite transport. https://othernet.is/
I know the internet is fundamentally two-way. It seems there's always a few competent users who know how to get information out of a conflict zone or disaster area... getting information in beyond a few choke points seems to be the hard part.
I think Othernet is really cool. Don't even need a big dish. In 2kbps, they fit in RSS news feeds (BBC, Al Jazeera, Boing Boing, and others), APRS messages, WikiNews, global winds report, a daily Voice of America news report.
Currently only works in Europe and North America, but nice to know there's a middle-of-no-where service that doesn't require a sub that gives you some barebones info from the outside world.
It's a social one, because social means are what are used to enforce the lockdowns - and those social means exist because of socially justified purposes in the societies.
Unless it was literally impossible to block internet access at all (some kind of unjammable starlink with undetectable receivers/transceivers?).
> Unless it was literally impossible to block internet access at all (some kind of unjammable starlink with undetectable receivers/transceivers?).
Yah, these types of things would be hard to detect (just point an LNB at the southbound sky... no dish... cheap electronics...)
What it's not is turnkey or filled with a good feed of information that people would want at baseline. There's no adoption model besides being a dissident, and people tend to start planning for that too late.
Starlink dishes are trivial to detect with even 1990s military technology when active. There's no way any satellite uplink can be hidden from a motivated adversary.
Any receiver using the heterodyne principle (this includes most current digital radio modems) is trivially detected by its local oscillator, often to some considerable range (potentially hundreds of metres).
In the UK, the BBC used to exploit this, with radio detection vans patrolling the streets for unauthorized receivers, to fine people for watching TV without a license. So did the USSR with shortwave receivers, where the penalty was potentially more than a fine. A satellite receiver is the same, in principle.
> Any receiver using the heterodyne principle (this includes most current digital radio modems) is trivially detected by its local oscillator, often to some considerable range (potentially hundreds of metres).
I'm sorry. This is not intrinsic; this is a design choice. Yes, most LOs just care about getting under legal emission masks, but there's nothing to prevent one from bringing it much lower.
Further, as I explained: the LO frequency of the LNB for the first conversion is often the exact same frequency as you'd use to downconvert digital TV broadcast.
> In the UK, the BBC used to exploit this, with radio detection vans patrolling the streets for unauthorized receivers, to fine people for watching TV without a license.
They never found anyone in a way that could be used for prosecution. "A freedom of information act has revealed that TV Detector vans have never to date provided evidence to be used in court against the TV Licence Non-payer, and have therefore never succeeded in enforcing the TV licence."
It's not intrinsic, no. It's possible to design a receiver with its local oscillator emissions to be down to an arbitrarily low emission standard. (But who actually does that?) And not all reception techniques require a local oscillator at all. Still. My point was merely that with an average radio, which was designed to just not be harmfully interfering, such as typical satellite TV receiver? It's quite possibly lit up like a beacon if you know what to look for.
> with its local oscillator emissions to be down to an arbitrarily low emission standard. (But who actually does that?)
Better testing to compliance standards has pushed LOs down way lower than they were in the 40s-60s. We have relatively cheap high quality spectrum analyzers and big integrated measurements. So-- kinda everyone pushes things down pretty low now.
> My point was merely that with an average radio, which was designed to just not be harmfully interfering, such as typical satellite TV receiver?
Yah-- this LNB looks just like a satellite TV receiver, with the exact same LO frequency. So your rogue downlink looks exactly like something that there's hundreds of innocent versions around.
Then a SDR inside your house chooses a different signal from the transponder and gets the data.
The point is, there's a whole lot of (mostly legal) satellite receivers around, so looking for satellite receivers doesn't seem like a great way to spot dissidents immediately.
It's trivial to force all the registered satellite broadcasters servicing a country off-air, if your already willing to arrest the former prime minister in public view, so anyone left with a hot dish would be easily identifiable.
So, you turn off all the state TV coming by satellite... so that you hope that all the satellite receivers get unplugged from the wall (else, they're likely to be listening to reacquire and get program guide etc)... so that you can drive around in vans and look for the -remaining- satellite receivers as a conduit of information coming in.
I mean, I guess it's possible. But probably not your top priority during a time of potential unrest. And it probably won't work that well as a way to find people with the links.
Any non zero amount of incoming radio signal to a hot dish would be suspicious, and easily identifiable, if all the registered broadcasters are offline.
I suppose the heterodyne frequently (the 465 kHz of AM, for instance) can be effectively hidden. Put your whole receiver into a metal enclosure, have a high-pass filter before the antenna connector, and a low-pass filter before the outputs and power connectors.
Such a device would likely look peculiar enough to be easily recognizable, though, so you'll need to hide it, possibly inside some innocuous-looking other device or enclosure.
The whole setup becomes somehow more involved, but still cheap enough.
> I suppose the heterodyne frequently (the 465 kHz of AM, for instance) can be effectively hidden. Put your whole receiver into a metal enclosure, have a high-pass filter before the antenna connector, and a low-pass filter before the outputs and power connectors.
Yes-- and every piece you mention is "quality radio design 101".
> Such a device would likely look peculiar enough to be easily recognizable
I'm not sure about that.
The biggest most suspicious thing is, if they're resourced to investigate everyone with a piece of coax going outside-- where does yours plug into?
Of course, if you're really dedicated-- get a satellite TV receiver, empty out the nice metal box, put the receiver inside ;).
Sounds useless for the internet (where every protocol is fundamentally two way), but great for ‘Radio America’ type broadcasts. Of course, so is shortwave.
definitely not solving the same ‘problem’ at all though.
The more random it is, the harder it is to have certainty, but at high enough transmission rates and powers the raised noise floor is impossible to miss and is definitely something that can be triangulated.
If a normal radio transmission is a point light of a specific color, truly random spread spectrum is a grey glow.
Direct sequence spread spectrum being many different color flashing lights here, I guess.
> The tricky thing is that receiving radio transmissions is also usually detectable
A LNB is pretty quiet to begin with... and the LO is on the same frequency whether you're watching satellite TV or tapping into Othernet. You could, of course, make it much quieter still. Just because a lot of receivers leak a lot of oscillator and mixer frequencies doesn't mean it's intrinsic.
> and a transceiver has to, by it's nature, transmit.
Which is why I was talking about a receive-only approach. (Well, also for reasons of scalability).
Well, I thought I made that pretty clear in what I posted.
"Maybe the internet as a whole is intractable. But perhaps we can get a meaningful digest of snapshots of the important stuff that can work over slow links, intermittent links, and one-way links?"
Since the whole point of the internet is, you know, people can request what they want and communicate with others, and they already HAVE people shoving ‘important stuff’ their way, what exactly do you think that adds to the picture?
> what exactly do you think that adds to the picture?
- Othernet is already a system out of local control that can be received with relatively trivial equipment.
- 2000bps feels perhaps a bit too slow, but with an order of magnitude more bandwidth, you could have a -lot- of textual communication. It could carry a big cross-section of what everyone is saying and allow people to choose what they see.
- An ideal system would support multiple transports; Othernet, intermittent wifi connectivity, sneakernet, etc. It would do the best with what it can to get data in or out, and allocate the scarce resources to the traffic which is most likely to be useful to others (this is tricky, with sybil attacks, etc... but I don't think it's necessarily intractable).
That is-- I love the interactive internet, but wouldn't it be cool if it gracefully degraded to still provide "postcards" and "broadcast radio" and a bit more as capacity went down, intermittency went up, and inability to return traffic was sometimes a factor?
> That is-- I love the interactive internet, but wouldn't it be cool if it gracefully degraded to still provide "postcards" and "broadcast radio" and a bit more as capacity went down, intermittency went up, and inability to return traffic was sometimes a factor?
This already exists, it's called ham radio. There are even large communities of folks sending 'postcards' to each other on a regular basis.
I'm probably failing to articulate what I'm describing. Not really. I've been part of the SSTV scene. I know about packet and things like PACTOR etc too.
I remember operating a UUCP site, and I remember how well network news gracefully dealt with quite a wide variety of sites with different ad-hoc setups. I miss that. I wish more of today's internet was a bit more like that: all the interactivity in the world when the network is good, but gracefully degrading to offline copies, snapshots, etc as the network is intermittently disconnected or one-way or whatever.
Secure ScuttleButt ("SSB") does a lot of this, but there's no real one-way provision and it's a bit wack to use and discover your way through.
The reason I think you’re getting so much confusion is it doesn’t seem to make any sense to folks what you’re proposing.
It seems like an edge case either already served by existing tech except for some very, very small set of situations where most won’t consider it practical or useful anyway. They’ll just knuckle under the societal rules which forbid it, because the cost of doing it in a way they won’t likely be caught and imprisoned isn’t worth it to them.
Most people just care about having food in their belly every day, feeling safe and loved, and being able to believe that what they’re doing is okay day to day and will work out to be all good in the future.
Doing what you’re proposing doesn’t really help them with that.
> It seems like an edge case either already served by existing tech except for some very, very small set of situations where most won’t consider it practical or useful anyway
Yes, it's all about the edge cases that don't affect the commercial value of the network much. It's about dealing with natural disaster, intermittent and very low quality network connections by modern standards, people in remote areas, or censorship. It may even be about space missions and distant probes in the future.
SSB is a neat system if you're on a sailboat with connectivity every few days. Othernet is a neat system for pushing what satellite communications can do to the limit. Of course, systems being technically neat is not what gets them adopted.
Networks, network capacity, and network reach have grown so much that we've forgotten about UUCP, etc. We have better algorithmic tools and more computation to build networks that tolerate disconnection, intermittency, etc. But perhaps we don't have the commercial motivation to use them, even if they would make the networks and societies that use them more robust.
I think to answer your underlying question of ‘why’ a bit.
You say ‘robust’ for societies as if it’s a given positive, but there is an implied statement hidden there.
Robust against what?
Robust against censorship is rarely desired by the majority (and certainly not openly), because the censorship is with the consent (and often at the request of!) the majority.
Just look at the CSAM/Kiddie Porn discussion in the US.
At least as long as the majorities interests and the gov’ts interests align enough, which is usually for far longer than anyone wants to admit.
So as long as it doesn’t step on the majorities ability to get what they want when they want it more than they are willing to tolerate, robustness against it will always be a feature with a limited market and a lot of opposition. An anti-feature in some ways, or at least it will be spun that way.
We may consider robustness against this good - and I personally do - but in my experience that is far from a common attribute, especially if you boil it down to specific examples.
Should Pakistan be able to robustly send how to bomb US targets + propaganda to everyone? Should someone be able to do the same with child porn? Or anti-Islamic rhetoric? Or anti-women’s rights? Or anti-free speech?
I think I explained "robust" in the post above pretty well. It's about making a network that works and delivers information even when things go wrong or when there's not much infrastructure available.
Primarily for the use case where there's just not much reliable infrastructure. I like the idea of getting a local snapshot of news, important source repository updates, and other things important to me during little snippets of connectivity.
But this also makes things useful for the case where the infrastructure is impaired. Perhaps by natural disaster. Perhaps by state interference. Perhaps by individual commercial actors.
Yes, "how does one determine what the network distributes broadly" is a complicated question. It should be democratic and resilient to being coopted. I understand this is very, very hard. But I don't think it's intractable, and even without it the other building blocks of a network like this are attainable (general resilience and reliability).
We now have the ability to train entire language models that contain information, can regurgitate it in the form of articles, and can even interact through conversation. Right now they’re big, but with continued improvements we could likely have useful and accurate models that fit on USB sticks and can be distributed through Bluetooth.
You aren't. The entirety of English Wikipedia (sans history) compressed with zstd is currently ~93 Gb from https://www.kiwix.org/. If you exclude the images, it's ~13 Gb. It's a much better bang for the buck than any LLM storage-wise, and that's before we even consider the hardware requirements to actually run an LLM vs an offline wiki reader.
Storage is effectively free. I can fit the entire Vicuna LLM on a $5 (US retail) SD card right now, and in a decade that’ll be less than $.50. LLMs will also presumably get more efficient at storing fact data outside of the network parameters.
People can download Wikipedia pages but weirdly they don’t. They prefer the interactive web to big downloads of static encyclopedia data, even though the latter is better and easier to distribute in a censorship-resistant manner. That’s why Tor is much more popular than high-latency censorship resistant networks that share static blobs, even though low-latency networks are relatively insecure and much easier to block. The question then is not how to save $3 of storage media: it’s whether you can simulate the interactive experience of the Internet in a satisfying way without a reliable low-latency network connection. Maybe the answer is no, maybe it’s yes. Imagine a really high-quality future LLM that you can actually interact with and ask questions about specific topics, maybe even using voice communication, and that can play games with you and generate other types of “web like” experiences. Is that really less compelling than a download of Wikipedia?
People in regions where Internet access is a problem to begin with do not download Wikipedia because they don't have the ability to do so. That's why Kiwix has side projects such as a Raspberry Pi based content server that can be set up in e.g. a school and used to serve content to all students using whatever cheap phones they might have, or a reader that runs on WinXP. And, yes, that stuff is actually deployed in various places in Africa.
Now imagine the hardware that you'll need to run that high-quality future LLM. How many Kiwix hotspots could you set up for the same money?
Mind you, I'm not saying that there aren't compelling use cases for LLMs. It's just that it's a thing that's nice to have after you have Wikipedia etc. Indeed, ideally you'd want to have Wikipedia indexed so that the LLM can query it as needed to construct its responses - otherwise you're going to have a lot of fun trying to figure out which parts are hallucinations and which aren't.
The situation I’m considering is not people who are sitting in regions where Internet connectivity is fundamentally poor, because there is no infrastructure. The situation I’m considering is one where Internet connectivity exists but is being deliberately throttled or filtered: either through periodic Internet shutdowns (as is happening in Pakistan right now as described in TFA) or through ubiquitous content and site filtering (as happens routinely in China and increasingly in other nations.)
There are places where Internet access and compute resources are fundamentally limited by lack of infrastructure. Talking about LLMs in those places doesn’t make sense. But on the trajectory we’re currently on, those places will likely become more rare while political Internet filtering will become more common.
The real question is how you do this in a way that doesn't encourage one-sided propaganda, sybil attacks, etc.
I'm reminded of systems like Secure Scuttlebutt and the Othernet satellite transport. https://othernet.is/
I know the internet is fundamentally two-way. It seems there's always a few competent users who know how to get information out of a conflict zone or disaster area... getting information in beyond a few choke points seems to be the hard part.