This is truly interesting, but from a privacy perspective if mass adoption actually occurred, extremely worrying.
Not “inverse surveillance” as stated, but just surveillance in the hands of people. At mass deployment, mass surveillance accessible by anyone. Since when does that protect anyone’s privacy?
The true problem with surveillance is the asymmetry of it all. Grassland uses Game Theory to remove that asymmetry. "...no party could maintain a parasitical, data gathering asymmetry (one-sided surveillance) so long as there are other parties acting in their own self interest. A 'scorched earth policy'".
It answers the question, who's watching the watchers? Everyone. And the network itself is paying them to do it.
> ...extremely worrying.
It seems to me this forces people into Nash Equilibrium. Say I now know everything about you and vice-versa and anything we do to each other would be broadcast around the world for all to see and irrevocably encrypted in a worldwide database. And since "...all nodes are anonymous, if it's computationally improbable to be sure of the identity and intentions of all parties who may end up with the data, then all parties are compelled to expand node territory for their own self interest and eschew behaviour leading to M.A.D. outcomes"
> Since when does that protect anyone’s privacy?
It doesn't. And it doesn't care. I don't mean to sound rude but you'll have to throw away this veneration of privacy to grasp what this is really about; like esteeming the Divine Right of Kings, that's not going to work any more. People are fighting over the right to keep riding this precious dead horse while corporations and governments are riding over them in AI Ferraris because they know that the ones who'll will win are those who privately control the most data to train machines to do, well... everything that a human can do and more. They don't care about your privacy. It's long gone and turned into glue. People want to believe in it but it's dead. And the longer people think it's still out there somewhere waiting for them, the longer it'll take them to grasp the power that data affords any party that disregards that belief. Then one day they'll wake up and realize all their land is in the hands of the Conquistadors.
> you'll have to throw away this veneration of privacy to grasp what this is really about
I cannot disagree more strongly. Every day our need to fight for privacy increases.
All it takes is someone to take my actions and skew them in a political light. From there on it's tiny policy shift that could end up with me in jail for some fascist political crime that isn't a crime at all. Every day our governments take more and more: we really shouldn't be doing their work for them!
Look at what's happening in China: a whole religious cleanse is taking place: Muslim, Buddhist, and even simple meditation groups (Falun Gong) are being persecuted and imprisoned.
Russia has outlawed Jehovah's Witnesses for similar purpose, citing "extremist activities".
And this is before we even get to more "political" style actions, such as white hacking and security research. We need to be exceedingly carefully with public spyware!
If Xinjiang were under public surveillance, with the results accessible to the whole world, the Chinese government might tread more carefully.
As is, one of their biggest enablers is being able to effectively deny access to the area.
Edit: I'd note this is generally an effective tactic of oppressive regimes. It takes dedicated, brave, well-funded people to access an area that's even semi-denied. Much less regularly. And the internet / world's attention is fickle. No updates, no photos, no video? No one pushing for justice.
> If Xinjiang were under public surveillance, with the results accessible to the whole world, the Chinese government might tread more carefully.
Not really. The world already has a pretty good idea that China has incarcerated millions of people for what wouldn't be considered offences or adequate standards of proof in the West as part of wider political struggles. And livestreaming is going to end at the point of arrest. A system tracking the full history of every one of those millions of individuals' movements prior to incarceration would add no useful insight into the nuances of the conflict or rhetorical weight to arguments to take any form of action against the Chinese government, but would be exceptionally useful to helping said government in tracking down anyone that has ever interacted with $PersonOfInterest or visited $Place.
There's a reason that people campaigning for the world to take interest in human rights issues in their region do so by broadcasting explanations of their cause and how brave and inspiring their missing friends are rather than broadcasting their whereabouts.
You're looking at something and calling it 'surveillance' because you don't yet have a word to describe the entire thing. It's too new so you see it through your 'old eyes'. You noticed there was a camera involved and you knew what that was so you ran with it.
It's the proverbial blind man whose hand has happened to land on the tusk and he tells his companions, "It's a spear!". You're not seeing the whole elephant because you've never seen one.
It benefits you when other people have knowledge about the world. It benefits you if other people have the knowledge to drive safely. It benefits you that your neighbours have a basic knowledge of germ theory and don't throw their sewage in the streets like medieval peasants. It benefits you that your fellow voters have at least a basic education and can make somewhat informed decisions. It benefits you that medical knowledge isn't locked away in some vault in Alexandria but it's in the mind of the doctor whose preventing an outbreak of some disease you happen to be susceptible to.
Did you not read the things I was able to learn just from my first node? We can give so many people free access to information that would push back darkness, fear and superstition just a little further. And these people would do and build things that would make your life better.
So instead of channelling the Archbishop of Canterbury telling Bible translator John Wycliffe that too much knowledge will corrupt the commoners, open your eyes.
How many people are actually detained? How long they're detained for? What requirements have been made for the population when they're not detained (e.g. submitting biometric data)?
These are all questions to which the answer is at best murky.
Facts answering these questions would empower opponents of Chinese oppression. Whether that would be enough to offset the privacy issues is a fair debate. But it's not like you're losing anything the Chinese panopticon hasn't already taken away.
I'm interested in knowing what you think about my response to someone else below who was inquiring about my 'Deep Schizophrenia' AI model -> https://news.ycombinator.com/item?id=19537047
Any step made in the direction you suggest will lead to further avenues of monetization of individuals, and in my opinion, dehumanising them /us /yourself.
I can see a plethora of negative impacts and influences on the ability for a person to express themselves, in what would have been a "natural" set of responses, if your project is even remotely successful.
You think the whole of the human condition can be mapped and measured onto a landscape? Although I have serious doubts that you can: Great! I'd suggest you train it on what is called "Modal Music", whose very fundamentals use the same language (landscape, narrative) that you describe. It's found in many Mediterranean music styles (greek, turkish, arabic...).
I'd suggest that people are not just one thing. We don't hate all the time, we don't laugh all the time, we don't work all the time. I think treating humans as an equation is just as demeaning as the current political elite and marketeers treat us, but that's not (ironically) a big concern of mine. I think this project naive and the consequences not well thought through.
The day you tell everyone your CVV codes on your credit cards, that you use a public bathroom with the door open, that you go to work naked, that you take the curtains off the walls, that you publicly list all of your passwords, is the day that I'll take you seriously.
Personally, I view your project in a manner akin to upskirting. I'm sorry to be so blunt, but our moral compasses are facing in very different directions it seems -- despite the fact that I can see you're trying to make something beneficial, whereas I only see horror. Where you see a stop gap in the surveillance state we currently live in, I see it as a precursor to even more horrible practices, for which an increase general violence will be inevitable. Whenever there is a camera there is surveillance. By running algos over this surveillance, it will make those in power more powerful before your rose coloured view of your perfect world even gets close to being materialised.
I am utterly horrified that you seem proud by these words of yours:
> I now know most of my neighourhood's average height, individual walking gait identification pattern, estimated salary based on car model, family structure, daily schedules, how many (visibly) pregnant women, on what days in July the guy across the street mowed his lawn, the pattern he mowed it in the 8th time and that it was the same day my other neighbour had 5 guests over for a get together. I can rewind and replay it from multiple angles in 3D. If I really want, I can convert it to spreadsheet format, etc.
Every single thing you mentioned here can be used against us. I hope that you spend some time reflecting upon this. Ultimately, I hope that you gain some wisdom and think differently.
The political science aspects of this are super interesting. I would prefer perfect privacy for everyone and consistently argue for increased privacy rights, but if that is impossible I would definitely prefer everyone has equal access to surveillance on each other.
Having considered how I would attack this it seems trivial for the super rich and mega corps to defend against this and maintain the asymmetry however. They would merely need to buy up enough land that outside cameras could not see all of it or build walls to block the view. Then when traveling outside their land they would just need to travel by vehicle that occluded any view of them, and simultaneously have a few other duplicate vehicles move with people in them who don't have to look identical but close enough that the cameras and system cant differentiate.
Its overly expensive to a normal person, but whats a few hundred thousand or million a year to a Zuckerberg or Bezos to maintain their privacy? I know Zuckerberg already does this to some extent with his house[1]
To add to this, I think making it asymmetric on how easy it is to surveil someone vs protecting themselves from it might be the right way to balance out privacy, but I don't see it being possible without putting legal limitation on the rich.
Given our current system its unlikely you could ever make it asymmetric enough to stop the rich unless it was as bad as needing 1 million dollars to defend against every 1 dollar used to attack. At that point the rich have enough power to influence the government and just get them to make this illegal and everyone who uses it face jail time. Given that the nodes can easily be traced back based on the physical location data they are giving out, it wouldnt be hard for the government to send police to any area with a node and destroy it and jail whosever property its on.
Someone eventually: "..and somebody's built this anonymous AI thing that tracks people and physical assets and lets you know where every citizen, rival public servant or object has been or whatever they've done before and they've posted it online for free. It's disgusting!"
Politician/News Agency/Insurance CEO/Hedge Fund Quant: "Oh that's disgusting. An application that tracks people and physical assets and lets you know where every citizen, rival public servant or object has been or whatever they've done before and they've posted it online for free. Where? Where is this thing?"
Someone eventually: "I don't know, one of those stupid '.network' domains"
Politician/News Agency/Insurance CEO/Hedge Fund Quant: "Ugh! Those stupid '.network' domains! (shakes head convincingly)... But I mean there's so many of them though! Which one? Which one is it on?!"
> The true problem with surveillance is the asymmetry of it all.
This argument reminds me of those pro-gun arguments: we can't prevent guns from falling into the hands of wrongdoers, so let's allow everyone to arm themselves – soon we'll reach an equilibrium where the fear of retaliation acts as deterrence.
I don't agree with this kind of thinking at all, the way to deal with destructive technologies is not further proliferation.
I don't think the comparison is entirely fair.
The gun argument directly turns towards violence. This is an entirely peaceful approach to an inevitable societal end-state.
Eventually, everyone at all times will be under constant surveillance, even in Western nations. Just look at the UK. Walking around London there is an undeniable asymmetry of constant, pervasive surveillance.
Unless you are suggesting the violent approach (people walking around smashing CCTV), I don't see why you would oppose this.
And if voters are opposed to having more information at their disposal they'll regret that decision. Politicians will see the opportunity fate has dropped into their lap and realize they have one of the following two choices:
Option 1: Oblige voters and pass a prohibition that regular people are not allowed to surface the data through Grassland. This would maintain the status quo for that particular polity but ONLY if NO OTHER polity in the world, to whom CAPITAL can flow chooses to act in their own self interest and takes Option 2 i.e. 'defect' in the Prisoner's Dilemma[1]
Option 2: Realize that having constituents incentivised to surface this data themselves means denser node coverage and less friction; Level 5 self driving cars (with better than human situational awareness); more efficient stock markets because of information decreasing risk and the opening of new markets and industries never before thought possible; less ignorance, violence, fear, unaccountably and resources needed between law enforcement and their communities; and a guaranteed (as long as the node's running) source of income for their voters. And not to mention the fact that they won't be forfeiting AI dominance and capital to those political bodies that have chosen Option 2 themselves.
All misunderstandings and fears of Grassland stem from an ignorance or misunderstanding of Game Theory. And until now this was little cause for concern. But now it's very dangerous to the public interest. And will become even more so when Deep Schizophrenia gets released.
And also the present asymmetry is in the average person's favour. Google's interest in tracking my whereabouts begins and ends with which ads I'm most likely to click on.
My family, employers and people I've irritated recently have considerably more interest in making and acting on other inferences (correct or otherwise) about my life from movement data.
Your thoughts on Grassland echoes the ubiquitous, widely accessible and on balance beneficial surveillance predictions David Brin made in "The Transparent Society". https://en.wikipedia.org/wiki/The_Transparent_Society
But isn’t balance here mostly about a balance of power? And isn’t a balance of power in terms of surveillance more than just a balance of information?
In a transparent society the most powerful will still find their ways to be less transparent than everyone else, while leveraging this difference in transparency to the maximum extent possible.
Information alone is not power. It is also about the ability to act on it, to use or abuse it.
What if Zuckerberg lives in a different neighborhood than you? In fact, most of the people your device would surveil if you did live near Zuckerberg would be his household staff.
Everyone has rivals. Everyone has people who are interested in selling to them. Everyone has people who are interested in finding out what they do, where they go etc etc. And no one more so than the rich and powerful.
Regardless, I think you're misunderstanding the algorithm. Nodes broadcast their data to other nodes because it stands as unassailable proof an AI performed a difficult computation on human behaviour and earned those coins. So you don't have to have a node in a part of the world you want information about. As long as there's someone there who either likes cryptographic money or wants to know more about the activities of that person beyond where the node tracks them, you're going to get that information one way or another.
People in expensive neighborhoods tend to be less interested in getting $10 than people in less expensive neighborhoods because they already have money, so this thing will predominately target the not-zuckerbergs of the world. If you wanted even coverage, you'd have to compensate for the fact that rich people tend to live near other rich people who are less motivated by the same money than poor people who live near other poor people.
Do rich people have insurance companies, ex-spouses, business partners, rival colleagues, rival corporations, government agencies who'd like to know more about them yet want plausible deniability that an anonymous tracking system affords them? This is what I mean by saying that Game Theory balances everyone out. You have to consider that there is always another party and they're all motivated by their own interests. And I don't mean to be curt with you but I'm pretty sure rich people leave their houses every once in awhile. Don't you?
Good, it's democratizing surveillance. Anyway, society need to evolve from privacy focused to almost everything transparent. Information want to be free, it's not practical to fight it.
People are getting so hung up on surveillance. You have to think like a data scientist. And all the things seemingly useless data can tell you.
And Grassland nodes are just software that can read a feed streamed from any digital camera in the world P2P. So in order to stop it, you'd have to somehow ban the ownership or use of all cell phones by everyone, everywhere.
It seems technically interesting but societally problematic. For example it appears to greatly simplify stalking and harassment. How can a domestic violence victim avoid their attacker if their movements are so easily monitored? Perhaps use this to identify people walking away from a gay bar or a political meeting to set up a bashing? Note when somebody leaves home in order to rob their place? And bear in mind that for some individuals the cost of being identified back won't outweigh the value in being able to locate their victim.
The situation you're describing means you can know exactly what your stalker/harasser's routine is too. You can also prove undeniably that they are a stalker or harasser. In fact, Grassland will be watching them even when you can't. It can make statistical predictions to you about when they sleep, where they work, where they're going to be at 11 tomorrow etc etc.
Yes, but you don't know your stalker exists until you are attacked and perhaps they just don't care if their routine is known. Consider that your stalker/attacker may enjoy the power of forcing you to monitor them back. So now we have a new avenue for trolling.
The system as described empowers the obsessive aggressor far more than improving the lives of people who don't care to focus on the minutiae of the lives of others.
To join the hypothetical, I'd imagine you can deploy AI/ML to try to avoid your stalker(s). Google Assistant already tells you when to leave for work and which route to take, imagine having it vary the time and the route to avoid stalkers it has noticed from the data it gathered.
But then the stalkers would also have the same system to try to figure out what your digital bodyguard recommended to you.
You guys don't really see the big picture here. You're hung up on minutiae. And it doesn't matter. Because there are other people who'll read this. People who understand how seemingly useless data can tell you so much. Data the human body and other physical objects "stream" every time they move. Data that pops into existence and is gone in an instant, so no one cares about it. They "get" it ;)
Not about stalker, but related question. What if the offender don't really care if you can see them? They just do a suicidal attack to the victim, like what a terrorist always do. Or what if the offender are prepared for the jail time?
You can't stop people who ignore consequences, and this tech enable them to perform what they want without an organized crime group or terrorist group.
State surveillance is an issue, but this tech creates different issues.
Hi David, I posted an issue on GitHub with a few questions, but I have one better asked here: what’s with your license? I think I understand it generally in that it looks like you want the right to harass people who integrate grassland with social media; but what is the Deep Paranoia text generation thing you refer to? And can you talk about how your group theory link relates to the project?
> Any abuse of this software and/or algorithm(s) evinced by parties engaged in violation of this Agreement, if discovered will be taken as aknowledgement, consent and agreement by those parties to allow any member of the Grassland community to attempt to seek out and identify such parties in order to target and generate any digital news, social media content or other types of information that the household and/or family of the members of such parties may consume using the 'Deep Schizophrenia' narrative
It reads like, "if you break this license, we'll dox you and your family"
Except that their modification of MPL2 has a giant loophole. If you include MPL2 code as part of a larger work, you can license the larger work as GPL. At that point, their restrictions are scrubbed from the license of the GPL codebase.
It doesn't matter though, you can't claim right to harass in a copyright license agreement. The whole thing is bunk, but it goes to show how messed up the world view of its creator is...
At least he is making it open, governments won't. While this is worrying, it is inevitable (satellite surveillance, road cameras etc - everything is in place, it's just about time), so I'm happy we're getting it this way and start working on learning to live with it. I'd hate to have to learn how to live under an oppressive government again (I'm from a former eastern bloc country), so this is definitely welcomed by me - this way I can keep track of the secret police.
It's funny that the words they use, "dystopian", "doomsdayish" etc. etc. ... are all from fiction. These are literally categories of genre-fiction. They're trying to make straight faced predictions about the future using what is literally in the book store under "Fiction". Something that is by definition not real, a fabrication, a lie. How can they possibly expect me to take them seriously or grasp what it is that I'm doing? They've taken stories they know are fake and mapped them to the real world. How is that any different from religion? I'm giving them mathematical arguments and they're giving me Dr. Seuss.
That's messed up. Well, Hitchen's Razor says that, "What can be asserted without evidence can be dismissed without evidence". Here's my razor, "What can be asserted with fiction can be dismissed with fiction". This is a bulwark against those who insist on living in a world of fantasy and superstition. But for the rest of us, those who want to live in reality, I'm giving you the real world.
Yes, frankly I see the current strong tendency to use metaphors and labeling as one of the greatest diseases of this age. IMHO it simplifies and degrades any topic it touches to the point of nonsense, which of course means that any subsequent discussion is nonsense. Metaphors and labels also carry emotions that are very complex AND totally individual, which makes them even worse for discussions.
The TL;DR of the linked comment/explanation above is:
* OP built a neural network (Deep Paranoia) similar to GPT-2 + DeepDream that can generate text for any narrative, and identified that anyone could use such a system to reinforce any (false or not) narrative they want
* OP then built a symmetric surveillance system (Grassland) focused on "truths" (derived from proof of work from IRL video feeds) to help protect humanity against a future in which auto-generated false narratives are widespread, intending to provide common people with something closer to omniscience such that they can see more (spatially and temporally) to be able to verify fact against potentially false narratives
It's doomsday-ish, but seems like a natural evolution of tech in this space. And understandable why building the former (Deep Paranoia) would prompt someone to build the latter (Grassland).
Obviously you can't have a "we have the right to harass you" clause in a copyright agreement, though. Threatening to do so is quite extreme.
The words people throw around here, "dystopian", "doomsday", etc. etc. are all from fiction. These are literally categories of genre-fiction. They're seriously trying to make straight faced predictions about the future using what is literally in the book store under "Fiction". Something that is by definition not real, a fabrication, a lie. How can people possibly expect me to take them seriously or grasp what it is that I'm doing? You can't take stories you know are fake and map them to the real world. How is that any different from religion? I'm giving people mathematical arguments and they're giving me Dr. Seuss.
And here you've put words in my mouth to suit your narrative. The story you want to believe. A story that's fake by defintion.
Don't you see how messed up that is? Well, Hitchen's Razor says that, "What can be asserted without evidence can be dismissed without evidence". Here's my razor, "What can be asserted with fiction can be dismissed with fiction". My software is a bulwark against those who insist on living in a world of fantasy and superstition. But for the rest of us, those who want to live in reality, I'm giving you the real world.
And it's not 'deep paranoia'. It's 'Deep Schizophrenia'
The words you guys use "dystopian", "doomsday", etc. etc. are all from fiction. These are literally categories of genre-fiction. You're seriously trying to make straight faced predictions about the future using what is literally in the book store under "Fiction". Something that is by definition not real, a fabrication, a lie. How can you possibly expect me to take you seriously or grasp what it is that I'm doing? You've taken stories you know are fake and mapped them to the real world. How is that any different from religion? I'm giving you guys mathematical arguments and you're giving me Dr. Seuss.
And you've put words in my mouth to suit your narrative. The story you want to believe. A story that's fake by defintion.
Don't you see how messed up that is? Well, Hitchen's Razor says that, "What can be asserted without evidence can be dismissed without evidence". Here's my razor, "What can be asserted with fiction can be dismissed with fiction". This is a bulwark against those who insist on living in a world of fantasy and superstition. But for the rest of us, those who want to live in reality, I'm giving you the real world.
As I stated to the parent comment, I didn't mean to offend. My choice of words could use some improvement. Here's further explanation. https://news.ycombinator.com/item?id=19537047
Your licensing terms are less offensive, and more invalid from incoherency and encouraging of illegal actions as well as showing a deep misunderstanding of what valid licensing and contract terms even are.
(I responded to your other question in a different reply)
> And can you talk about how your group theory link relates to the project?
You mean GLn(F)? That's a really, really cheesy kind of math joke that doesn't really work well. It's the symbol for 'The general linear group of degree n over any field F'.
But if you stay up too long writing code, well then you might realize it can also be used as a pun/pseudo-acronym on the word 'Grassland'. Take the first three letters, GLN and what do you get? ...GrassLaNd. I know, right. Keep your shirt on. This party ain't over yet.
And the 'F', well that's a 'field', right. like grass in a field. Ooohhh... Yeah... get it?... Oh Yeah, I bet you're impressed now.
And Bonus! The transformations of that group are 'symmetries'...
'Symmetry' as in the opposite of 'asymmetry'... You know... like how Grassland tries to promote symmetry... uh-huh... yep. That's right. I bet you're rolling on the floor with laughter now.
...So yeah, it's not at all a good joke. And yes, I am really fun at parties.
Deep Schizophrenia is a deep neural network model I'm going to either open source or make an API for people to use that can be used to generate new narratives of any size. And it doesn't have the semantic and narrative "fall-off" you get after a few sentences with models like OpenAI's GPT-2 and other "attention" based or LSTM models. It's called "Deep Schizophrenia" because like Google Deep Dream, GAN's and style transfer models do to create new images of people, lanscapes etc., it sort of "warps" narratives to generate new ones. It's as if the model is having an hallucination (hence the nomenclature contrasted to 'Deep Dream') but instead of changing images it's changing the semantic and narrative embedding space.
You're talking about some kind of online trolling and harassment. That would be stupid and accomplish the exact opposite of the intended affect. But I'm talking about customized, generated digital content using the very same content channels that everyone else on here uses. Social media, blogs etc etc.
Let me explain how Deep Schizophrenia works.
I was able to construct a continuous, fractal, space filling curve that satisfies Peano's/Cantor's definition (https://en.wikipedia.org/wiki/Space-filling_curve) but I was able to give it just enough additional "structure" so as to let it be treated as being differentiable 'everywhere' (If you're not clear on the definition of 'differentiable', for now, just know that it's very important for training machine learning models). This lets me normalize each document in my corpora from narratives the size of a single, abstract sentence like "The quick brown fox jumped over the lazy dog." to the entire novel 'The Fox and the Hound' so the entire narrative, no matter the length can, as a whole be embedded into a common narrative unit 'space' by simply adjusting the iterations of the curve according to word count. In Deep Schizophrenia, individual word tokens aren't decoded from nor treated like discrete values (as in CBOW, Skip-Gram, BERT etc.) but as the localized, resultant values of wavelet functions 'passing through' the dimensions of the space. So this lets me use techniques similar to GAN or style transfer but which are heavily modified to take advantage of my curve's structure in order to generate new narratives by 'warping' them along these dimensions while still maintaining both narrative, gramamtical and semantic 'cohesion' So no "fall-off". To borrow a metaphor, think of these 'semantic' wavelets like draw strings on some n-dimensional piece of fabric, when you pull on the string, the entire garment, from 'hem' to 'hem', gets pulled, bunched, stretched or twisted cohesively, as one garment, which is what you'd expect; how consistently it conforms to one's expectation of narrative, grammatical and semantic cohesion is largely dependent on how much memory one can afford to throw at it during training.
And the training set is prodigiously annotated and tagged with themes, prominent characters/persons, archetype categories etc. etc. So these inputs can be modified to get different predictions (stories). GPT-2 was trained on about 40 GB's of largely unannotated data. And it's considered state of the art. But I have over 247 GB's of annotated narratives all of which could potentially be trained with (I estimate the GPU costs to train any significant portion of it would be around $250K and take months. But it would be worth it).
I chose a fractal structure because narrative expression has what appears to me to be a sort of infinite set-ness (https://en.wikipedia.org/wiki/Infinite_set). This allows Deep Schizophrenia to recursively fill up the story from the 'inside' with each pass, moving down from higher levels of abstraction as it increases the cardinality (detail) of the narrative set. So in theory, with some clever memory management, it could allow you to fill up a narrative indefinitely (a-la George R. R. Martin). Theoretically, it should also be possible to take a complete, encoded narrative, normalize it onto a fraction of the curve, say [0,0.5] and, with some software architecture I haven't quite figured out yet, generate a "sequel" on the interval (0.5,1].
Narratives, common stories, myths etc. inform and tell you more about people's beliefs and ideologies in ways mere factual data could never hope to tell you. Did it matter that Boston Tea Partier's were actually dumping the tea in order the protest the new lower prices of the East India company's tea (thanks to the British lowering the tariffs) which were negatively affecting the sales of their own speciously sourced tea? Not really because what's the story that made America what it is today? The one that reaffirmed the heroes of the American Revolution. Does Turgenev's 'Sketches From A Hunters Album' and the affect it had on people's understanding of the morality and brutality of Russian serfdom any lessoned by the fact it's a fictional narrative. No, I would argue the fact its a fictional narrative loosely based in reality is what makes it that much the stronger.
And we can now tell people much better stories, with better customized content (targeting) faster and much more efficiently than any human could hope to accomplish.
(sigh) A lot of hard work and bootstrapping. My friend, don't waste this moment. There is a much more pertinent question you should be asking me.
There's a lot to unpack here. But you have to understand that I built Deep Schizophrenia (though I didn't call it that at the time) SEVERAL YEARS BEFORE I built Grassland. Partly because I realized what D.S. could do to people. Let me explain....
We'll take a few of the arguments some people have commented here as an example. I imagine most of them are athiests. But it's irrational to think 100 years of Nietzche is going to make a dent in 3 million years of evolution. We're hardwired for silly beliefs (no offense). Every culture and social group has their own pantheon of gods. They all just have different names for them. You can talk about 'privacy' and you can talk about 'Privacy'. Encryption and closing your blinds will give you privacy. That's rational. But there's no Privacy god who's going to make a data scientist suddenly unknow your pilfered Equifax credit history. The Privacy won't make the former employees of Cambridge Analytica suddenly unknow how to make your aunt vote for candidate X. And they'll never outright say they believe that. But they do by their actions.
Like 4chan with Nazism, they at one point merely cajoled one another with this mocking, ironic disattachment to the idea of a Flying Spaghetti Monster. Because they thought they were too smart to believe in it. But then some where along the way, they actually did.
These are all different stories. Human beings are extremely susceptible to the power of a story. They're like those funghi that take over an ant's brain till the ant is controlled by the funghi. If you want a story to placate your fantasy, if you really want that then I've built Deep Schizophrenia (Well, I actually built it to generate romance novels. Romance is a big industry) and it's ancillary software to figure that out for you and provide the story/rhetoric that reaffirms your fantasy back to you. A virtual, virtual reality.
But for the rest of us, those who want to be able to have data about the real world with a statistical guarantee of validity that we can calculate and create a clear separation between that and things that are mere stories, narratives and rhetoric told by humans and therefore subject to bias and interpretation (I enjoy the Lord of the Rings but I don't literally believe in Mordor), for those people there's Grassland.
And that's part of the reason why I built Grassland and why I built it in such a way as to make it extremly costly to get false data into the system. Because I knew eventually either I would release the Deep Schizophrenia software or because simultaneous discovery is so common (https://en.wikipedia.org/wiki/List_of_multiple_discoveries) possibly someone of untoward quality would discover it, use it in secret and not tell people about it like I did. And there'd be no safeguard against it if I didn't build Grassland.
I'm not saying the things I build are perfect or they're going to fulfill your fantasy of a perfect world (again, Deep Schizophrenia can give you that fantasy if you're hell-bent on stupidity). But what I try do is give mathematical arguments to support my conjectures.
Hence why I wanted a license to prevent people merging the software with things that would counteract Grassland's purpose. Yeah, maybe I wrote it wrong. It's a little difficult solving some of the world's oldest mathematics, AI, computer vision, cryptocurrency and surveillance problems in one's spare time. Adding a law degree to the mix must have slipped my mind. I'll fix the license. I'm only one guy...
So as I understand, the higher the wordcount, the deeper the "spaces get" by generating filler content matching the space and fit for the dimension of the space.
> ... the higher the wordcount, the deeper the "spaces get"...
During training, what I would say is that the "space" gets denser. Imagine you live on a cliff overlooking a lake/sea (some body of water with known boundaries). You notice on some days the winds produces long waves that are spaced far apart, while other days the waves are very short and choppy. If you wanted to encode this, it would take more memory to encode the latter than the former despite the lake being the same size.
If you have more questions about the Deep Schizophrenia model, I'd be happy to discuss. You'll see my email at the bottom of the site.
I agree there are a lot of privacy concerns here, but isn't this what governments and mega-corporations are already doing? This way at least the data will be accessible to more than just privileged groups.
Furthermore, it seems reasonable to hypothesize that this could add a whole new depth to the market of privacy and security. If this "Grassland" became the norm, then privacy concerns would be less niche, the effectiveness of our privacy efforts would be more obvious as we would have immediate feedback, and it would, therefore, be easier to ensure we're maintaining privacy from governments/corps (or anyone else for that matter).
So rather than reflexively lashing out at this as a "scary threat to privacy" maybe this is actually a counter-intuitive evolution to enhanced privacy and connectivity.
Governments and mega Corp have some level of responsibilities. If someone is judged illegal, irresponsible, they can be prosecuted somehow. You have some limit to their power and some restrictions to what data they are allowed to gather and use, even if it practice there is obviously some issues at this level.
What the OP suggests is to use the power of the mob to apply surveillance to potentially anyone, completely removing the concept of privacy. That sounds incredibly dangerous and nefarious to me.
At least there are some limits to the use of the data when it is restricted to a few elites. When any loony or malcontent can use it then it just makes things much worse for those at the bottom of the heap.
You're looking at something and calling it 'surveillance' because you don't yet have a word to describe the entire thing. It's too new so you see it through your 'old eyes'. You noticed there was a camera involved and you knew what that was so you ran with it.
It's the proverbial blind man whose hand has happened to land on the tusk and he tells his companions, "It's a spear!". You're not seeing the whole elephant because you've never seen one.
It benefits you when other people have knowledge about the world. It benefits you if other people have the knowledge to drive safely. It benefits you that your neighbours have a basic knowledge of germ theory and don't throw their sewage in the streets like medieval peasants. It benefits you that your fellow voters have at least a basic education and can make somewhat informed decisions. It benefits you that medical knowledge isn't locked away in some vault in Alexandria but it's in the mind of the doctor whose preventing an outbreak of some disease you happen to be susceptible to.
Did you not read the things I was able to learn just from my first node? We can give so many people free access to information that would push back darkness, fear and superstition just a little further. And these people would do and build things that would make your life better.
So instead of channelling the Archbishop of Canterbury telling Bible translator John Wycliffe that too much knowledge will corrupt the commoners, open your eyes.
Exactly. The real world is comprised of data, and the more we restrict it to private interest, private interest gets to reap more benefits.
Information is only leverage against you if it's not public.
It's funny that the words they use, "dystopian", "doomsdayish" etc. etc. ... are all from fiction. these are literally categories of genre-fiction. They're trying to make straight faced predictions about the future using what is literally in the book store under "Fiction". Something that is by definition not real, a fabrication, a lie. How can they possibly expect me to take them seriously or grasp what it is that I'm doing? They've taken stories they know are fake and mapped them to the real world. How is that any different from religion? I'm giving them mathematical arguments and they're giving me Dr. Seuss.
That's messed up. Well, Hitchen's Razor says that, "What can be asserted without evidence can be dismissed without evidence". Here's my razor, "What can be asserted with fiction can be dismissed with fiction". This is a bulwark against those who insist on living in a world of fantasy and superstition. But for the rest of us, those who want to live in reality, I'm giving you the real world.
Yes, this is what I mean on the site in postulate '3. Data Symmetry' by a "scorched earth policy". Someone can't use it unilaterally as exclusive leverage if everyone else knows it too.
From wikipedia: "A scorched-earth policy is a military strategy that aims to destroy anything that might be useful to the enemy while it is advancing through or withdrawing from a location".
There's no unilateral benefit to hording valuable data about the natural world because Grassland's AI independently causes information inflation.
David Brin also touched on the idea in his 1989 novel Earth, where police bodycams were required, but citizens ubiquitously chose to wear bodycams as well to keep everyone honest. I always thought the idea felt rather inevitable.
The words you guys use, "dystopian", "doomsday", etc. etc. are all from fiction. These are literally categories of genre-fiction. You're seriously trying to make straight faced predictions about the future using what is literally in the book store under "Fiction". Something that is by definition not real, a fabrication, a lie. How can you possibly expect me to take you seriously or grasp what it is that I'm doing? You've taken stories you know are fake and mapped them to the real world. How is that any different from religion? I'm giving you guys mathematical arguments and you're giving me Dr. Seuss.
And you've put words in my mouth to suit your narrative. The story you want to believe. A story that's fake by defintion.
Don't you see how messed up that is? Well, Hitchen's Razor says that, "What can be asserted without evidence can be dismissed without evidence". Here's my razor, "What can be asserted with fiction can be dismissed with fiction". This is a bulwark against those who insist on living in a world of fantasy and superstition. But for the rest of us, those who want to live in reality, I'm giving you the real world.
In addition to that, and I'm going to approach this delicately because some people are uncomfortable with facing this fact. But since you're on here I'm sure you know this. If you have a cell phone, used a credit reporting agency (Equifax?) or used almost any website in the past 10 years, you're already being tracked and your data has already been stolen in, on average, one data breach a month to advertise or politically manipulate you. The horse has already left the barn. Your life is right now being sold and resold every millisecond.
But I'm doing you a favour. Because I'm giving your life away for free.
And that data is freely disseminated to everyone not just because it stands as unassailable proof an AI performed a difficult computation on human behaviour and earned those coins but because it's important that everyone knows the truth about events in the natural world for science, democracy, health, justice and all rational decision making. An embargo on stupidity and ignorance in that node's corner of the world at least. Advantage must then come from helpful innovation not parasites rent seeking on people's data.
To keep up with your analogy: I'm here mending the fences of my farm to keep the horses from running, and you are using a bolt cutter on my gates. It's without my permission, and goes directly against my wishes.
Some of us are still actively working on making the world a secure place. It can be done... but not with a project like this. It certainly doesn't help that this style of thinking plays right into the Mr Zuckerberg's hands.
> But I'm doing you a favour. Because I'm giving your life away for free.
You're obviously a very intelligent and talented individual to have created this, frankly amazing, system, and to articulate your reasons for doing so.
But I fundamentally disagree with the notion of total surveillance as a remedy to asymmetric surveillance. You obviously think this would benefit the world, but the endgame looks like a dystopian nightmare to me.
I really wish you would have used your talents for "good rather than evil", looking at ways to prevent surveillance rather than increase it.
> I really wish you would have used your talents for "good rather than evil", looking at ways to prevent surveillance rather than increase it.
It just takes a little more thought on the problem to understand that removing all privacy is an instant and irreversible one way ticket to a dystopian outcome.
OP suggests privacy is dead? Lol. I bet there are curtains on their house. I bet they goes to work wearing clothes. I bet they have sex behind closed doors... All of these are basic indicators that we have parts of ourselves and our lives that we like to keep private.
Ask any woman whose private parts have been published on the internet without their consent and they'll mention the shame and the violation they feel.
You're looking at something and calling it 'surveillance' because you don't yet have a word to describe the entire thing. It's too new so you see it through your 'old eyes'. You noticed there was a camera involved and you knew what that was so you ran with it.
It's the proverbial blind man whose hand has happened to land on the tusk and he tells his companions, "It's a spear!". You're not seeing the whole elephant because you've never seen one.
It benefits you when other people have knowledge about the world. It benefits you if other people have the knowledge to drive safely. It benefits you that your neighbours have a basic knowledge of germ theory and don't throw their sewage in the streets like medieval peasants. It benefits you that your fellow voters have at least a basic education and can make somewhat informed decisions. It benefits you that medical knowledge isn't locked away in some vault in Alexandria but it's in the mind of the doctor whose preventing an outbreak of some disease you happen to be susceptible to.
Did you not read the things I was able to learn just from my first node? We can give so many people free access to information that would push back darkness, fear and superstition just a little further. And these people would do and build things that would make your life better.
So instead of channelling the Archbishop of Canterbury telling Bible translator John Wycliffe that too much knowledge will corrupt the commoners, open your eyes.
I think what you've done here is interesting, fascinating even - but also terrifying, and better belonging in the realm of sci-fi, rather than the real world.
You talk about undeniable benefits knowledge, giving specific examples - yes, there are many examples where knowledge is useful and can improve lives, but I really don't see how knowing what time my neighbour takes a dump will benefit myself or humanity (a crude example, but you get the idea).
> Did you not read the things I was able to learn just from my first node
I guess you mean the "Use Cases" section on the linked page. Yes, I read those, and they played a large part in my comment you are replying to. For example:
> A citizen or public servant who wants to know what any citizen or (rival) public servant is doing at this moment or has ever done before
Sorry, I value my privacy and that of others. And this was the nail in the coffin:
> I now know most of my neighourhood's average height, individual walking gait identification pattern, estimated salary based on car model, family structure, daily schedules, how many (visibly) pregnant women, on what days in July the guy across the street mowed his lawn, the pattern he mowed it in the 8th time and that it was the same day my other neighbour had 5 guests over for a get together. I can rewind and replay it from multiple angles in 3D. If I really want, I can convert it to spreadsheet format, etc.
That data may be trash to you but to a sociologist, to an economist, to an epidemiologist, to an anthropologist, to a historian, to a city planner, to a property developer, to an investment firm, to an insurance company, to a kinesiologist and on and on... it's TREASURE.
Then don't install a Twitch streaming video camera in your bathroom. And don't order a custom t-shirt with your banking password printed across the chest. This isn't rocket science.
If you're out and about in view of others, you don't have privacy. By definition of what we call "privacy" that is the exact opposite of privacy.
Future generations will look upon this like we look upon the guy who disassembled a bunch of smoke detectors in order to turn his back yard into a radioactive superfund site.
Looking for a link to the PDF of it, but I’m reminded of David’s Gelernter’s “Mirror World” [0] concept from the 1990s, was lucky enough to encounter it in one of my classes.
Hadn’t thought of the proposal as more than sci-fi but here we are :)
I didn't have a name for the type of software that Grassland is and that's why the first sentence on the site is so long. I'd never heard of this "Mirror World" concept before but the similarities are surprising now that I've read a bit about it. I've added it into the description as "A world mirroring system..."
Kudos for the concept. Intriguing form of counter-surveillance that could actually work in a lot of existing surveillance states. Had a few questions :
1. Who in the system is responsible for maintaining and updating the object detection model? It seems like a centralized point of failure for an otherwise decentralized system. Might want to check out existing techniques for Federated Learning for ways to counter this. (https://www.openmined.org/)
2. How possible is it to mount a Sybil attack on this system? Why did you select PoW considering it's weaknesses? One possible problem could be that early adopters will necessarily have to be co-located for the system to have any value(Hence consolidating value in a certain geographical area, making it very easy to regulate/shut-down).
1) Yeah, I'm not sure how Federated Learning would solve the problem. Federated learning still requires someone to decide on a single model and it's still trained under the coordination of a central server[1]. And how do you ensure the data people are giving the model is good? You need some sort of scheme to determine that. So how do you decide on the scheme? Decentralization? Isn't this decentralised-turtles all the way down?
2) I'll explain in detail how it prevents Sybil attacks later. But It involves signing all the tracklets and digest tuples (see website) as well as a sort of Russian Nesting Doll of hashes. PoW is the only viable one we have. I like to use systems that are battle tested. I don't know what you mean by PoW weaknesses. Could you clarify? You don't need to co-locate. Read what I learned about my neighbourhood with just one node under "Use Cases" on the site. As to people shutting it down, I have clearly delineated on the site's front page what my theorem predicts will happen if they do. They're handing data, unprecedented knowledge of human behaviour and AI dominance to those who don't. So yeah, shut it down. Fill your boots, mate...(https://i.imgur.com/rDJWv5E.gif?noredirect).
The one saving grace of such a system is that it should make it trivially easy to identify who is running these nodes, so that they can be forcibly ejected from your community.
Grassland software is just the proof of a theorem whose 3 postulates/axioms I clearly outline on the front page of the site. My theorem predicts that a population that shuts it down is just concentrating cryptographic money and an AI with perpetually increasing and unparalleled knowledge of human behaviour into the hands of those that don't. In short, the rest of the world would react with -> https://i.imgur.com/rDJWv5E.gif
It says "successfully submitting fake data in the network is so computationally intractable that any self-interested node would find it more profitable to be honest", but that doesn't sound quite right.
I see the diagram and how it should enforce that the input frame is valid wrt the hidden layer, but it seems I could generate a fake overlay animation (for n POVs) and superimpose it over (or object detect and scrub it from) the input frame sequence.
You'd have to do this repeatedly. Many times a second. And you could have to always have a different one each time. And you'd still have to hash the hidden layer so you're still spending money. And you'd still have to make sure your objects didn't behave on the map in a funny way. They couldn't walk through walls or fly off into space. Or walk into the ground. Because other nodes will see that. It'd too difficult to maintain over time.
But I'm speaking to the set of cases where there is some external economic incentive (outside of the crypto token value) to either incriminate or exonerate some actor's behavior.
I think the scrubbing/exoneration scenario is straightforward, and the physical anomalies you mention (for the incriminate overlay attack) would be handled by CG or 1:1 offline model recording.
Maybe the network needs to report on cases where sensors disagree and visualize/surface these n different versions of history.
Any thoughts on leveraging something like Golem for the compute and Streamr or Erasure for the payment/sponsorship?
> I think the scrubbing/exoneration scenario is straightforward, and the physical anomalies you mention (for the incriminate overlay attack) would be handled by CG or 1:1 offline model recording.
You could but it would take a lot of effort even if you just wanted to scrub the data of one. You'd also have to know what you're about to scrub before you even do it. You'd also have to make sure there weren't other nodes around too. Keep in mind, every few months the power requirements of the models increase. The next model will have semantic segmentation (https://medium.com/nanonets/how-to-do-image-segmentation-usi...) so by Eon 1 things are going to be much, much harder to fake than Eon 0, which we're at now.
I'll admit, it's never going to be perfect. But as you mentioned, we can find ways to make it better. But one thing I believe we can guarantee mathematically is that if someone's going to make an attempt to submit fake data into the Merkle Tree, we're going to make them have to pay out the nose just to attempt it. So they have to have a lot of resources at their disposal. But in the end, they're also involved in raising the value of Grassland's currency for everyone else. Because they're funnelling more compute costs into the system, trading their money for Grassland's coin. So everyone else in the network is, in a certain sense, rewarded from the penalty extracted from the attacker's dishonesty.
> Maybe the network needs to report on cases where sensors disagree and visualize/surface these n different versions of history.
Yes, I'll think about this... Do you have further thoughts on this?
> Any thoughts on leveraging something like Golem for the compute and Streamr or Erasure for the payment/sponsorship?
I don't really know much about them. Most of the 'blockchain' stuff out there seems quite overly complicated to me and I'm not sure what they even do to be honest (Not to disparage these systems in particular. I think I've only ever heard of Golem before. And the other two I've never heard of). But I'd rather work with systems that I know and aren't going to be changed from under me because of the whims of some other dev team. And Grassland has quite a unique problem space so it was necessary to build the system from the ground up so I know exactly what each piece is doing and what it could potentially do because I don't put in any more code than I need to.
I'm not sure I follow the reasoning that making attacks expensive is strong enough protection. State actors have plenty of resources to throw at this, much more than private citizens - doesn't that mean this effectively becomes one way surveillance again? Since if the stalker has enough resources, you can't effectively stalk them back
One interesting tweak you could consider, don't know if it's possible: it would be great if everyone could see logs of access as well. If everyone is giving up privacy symmetrically, it might be an interesting dynamic if everyone can also see who accessed their information (or, if it's anonymous, simply know that someone accessed it).
Neural networks plus proof of work is interesting, but there is one, foundational problem: cryptocurrencies have no particular value. Bitcoin has a bubble of people who held a particular ideology and didn't understand currency, and folks who needed money laundering took advantage of that. Without a similar bubble, it's hard to convince someone to give you a currency that you can, say, pay taxes with in exchange for some numbers that you can prove were hard to calculate.
Beyond that, it's pretty straightforward to shut down. If a municipality decides it doesn't like this, you pass a law against it. You can't fully anonymize sparse nodes because the appearance and disappearance of objects under various situations lets you pinpoint where it is and either send the police up to break it or give the company running the poorly secured security camera an injunction to fix it.
> Bitcoin has a bubble of people who held a particular ideology and didn't understand currency, and folks who needed money laundering took advantage of that.
I'm pretty sure that the Bitcoin ideology specifically seeks to enable money laundering. It's not like people who needed money laundering took advantage of a design flaw; the point is to let people launder money.
actually all money transacted is transparent because the blockchain is public domain. What is used for money laundering is dollars. It's a common fallacy that there is a lot of money laundering with Bitcoin disproved by EU studies (google them).
It may allow organizations and governments to hold each other accountable, because generally there's an inappropriate coordination or centralization of the data-gathering machinery and the political/financial machinery.
As for privacy, I think it will serve to add more players to the data game, balancing the powers.
More power to those who have the resources to get the digging done for them. Even with a perfect AI interface, you could only ask "Alexa" so many questions in 24 hours. Lowly individuals would gain absolutely nothing, they have better things to do than sniffing each other's data trails. That investment of time would in the overwhelming majority only be worthwhile for malicious use cases.
Even in the simple stalker/stalker evasion scenario: the stalker would enjoy digging into the data, but for the stalkee, keeping track of the stalker for avoidance would be a terrible chore.
Where is gait identification or license plate tracking illegal? If you're visible to the naked eye from a public space in the US it wouldn't be a bad idea to expect it.
I'm glad you like it. Yes it can :) Keep in mind it's still beta. There's still more work that needs to be done. But if you have any questions just open up an issue on github or email me. My email is on the bottom of the website.
Quite amused to see that Presto card! Always great to see something done here in Ontario. On the other hand, the privacy/security implications of this are very frightening.
:) You know I heard saying once. I think it goes something like this, "Every mathematician receives their best ideas in the three 'B's. 'Bed', 'Bath' and 'Bus'"
Unfortunately, no. Not at this time. Your node calibrates itself and interprets what it sees in each 2D video frame into the real 3D world by 'knowing' where its 'virtual self' is placed in the 3D map of the real world that you'll find in the GUI. When you set up your node, you have to do set this calibration once, manually using the GUI and its accuracy in space is down to half a centimetre on the earth's surface and its accuracy in time is down to a millisecond.
If it was on the top of the car, the node would have to be constantly updated automatically. But GPS is no where near that accurate. I guess you could use a combination involving LIDAR to constantly tell the node where it is so the car's position is updated in Grassland.
But it would make more computational/economic sense for the government and economically self-interested citizens (interested in mining Grassland's coin) to line the streets with ip cameras which grassland nodes can read from to analyze and broadcast car and pedestrian locations/speed etc. etc. millisecond by millisecond. Which self driving cars could pick up and 'know' everything going on around them even things they can't 'see' themselves.
If the accuracy is down to half a centimeter , this is something that would need to be re-calibrated several times a year. The tectonic plates move over 1 cm per year, and any type of earthquake can shift land by significantly more than a half centimeter
What's stopping me from finding all the unsecured IP cameras in my city and using them to "mine" data? Further, if I could do that what's stopping others from reusing the nodes that I set up?
> What's stopping me from finding all the unsecured IP cameras in my city and using them to "mine" data?
Absolutely nothing.
> Further, if I could do that what's stopping others from reusing the nodes that I set up?
Right now, nothing. Nodes could overlap other nodes. But they're still using compute cycles so they're still spending money. The only issue is, who gets theirs into the tree first. And if the hash is the same, you better be first or you're just wasting time.
Keep in mind, I've still got a lot of work to do. And until today, I've been working on this by myself. And pretty much since last summer. You can see the list of tasks that I need to complete on the site under "Beta Version".
Thanks for answering my question. And I want to say I appreciate what you're doing. Information is power and it's being leveraged against individuals practically unchecked. If I had to guess some state actor would have come up with (or already has) with a technology like this eventually. Everything in that system would be kept private and be leveraged against the individual without consent.
Tools would be built with it that contain their own biases and uses. The appeal of using this system against others would be too great, either for those in power or those who seek to sell the ability. For that reason I like the ubiquitous access _everyone_ would have.
Distributing this to everyone reminds me of the end of "The Stars My Destination"
To continue with my original question though; I'm guessing whatever system has faster compute cycles will win the race of who hashes first? Will you be able to tell who's hash gets there first
> I'm guessing whatever system has faster compute cycles will win the race of who hashes first? Will you be able to tell who's hash gets there first
Yes, but it's not just about hashing power though. And there's no nonce like Bitcoin. It's mostly about deep learning inference power. You can see the process described on the website under "Unrewriteable History And Reasonable Storage Requirements" (though I'll probably change that title soon and call it "[something] Merkle Tree"). I've copied the last bullet point here...
* The process of producing the next Merkle Tree is a competition any node can compete in. A person who cultivates real trees is an Arborist so we'll use that term. Our "arborists" will also need to perform inference on the frames that were randomly selected for evaluation based on the last Merkle root, and will need to publish each frame's digest tuple and the "reserved digest" (see diagram). If the original node holding the reserved digest signs the arborist's digest, it generates that node's reward coins and a percentage goes to the arborist. But that will only propogate if that arborist's tree is also accepted by a majority of the network. Therefore, as long as (1) there is always active competition to compute the Merkle Tree by statistically verifying the work of all other nodes and (2) over 50% of every active node prevents inflation of the money supply by rejecting trees that don't pass their proof-of-work, then all rational (self-interested) nodes will have to be honest.
Interesting project! You could run the neural networks directly on the pi. Expect 3 fps for object detection, try ncnn or deepdetect + ncnn for ease of use. That's without GPU. With a Jetson Nano expect up to 50 fps. You could detect more classes as well, as needed.
Sorry, I'm going to change the video format later. Maybe a GIF or something. But I don't want to touch it now while so many people are looking at it just in case I mess things up.
I've uploaded the two of them to Gyfcat for now. Let me know if that works for you
Interesting idea to create a currency of proof of work for a task of intrinsic value. I had thought about creating a currency around making successful predictions, some computational activity that is not random but has some demonstrated value.
Not “inverse surveillance” as stated, but just surveillance in the hands of people. At mass deployment, mass surveillance accessible by anyone. Since when does that protect anyone’s privacy?