Nym | The Anarchist Solution to Techno-fascism: In Practice - Casey Ford | Common S3nse - 2026

[music] Hi, my name is Casey. I'm from NIM. Uh, I'm a philosopher by training. So, this talk is a bit of a mix between uh some of my intellectual interests, sociological interests, and the work we're doing at NIM. So, I want to talk about two ideas today.

Um the [clears throat] primary one is what we are calling today technofascism and the second is uh a kind of revitalized idea of what we mean um by anarchism specifically anarchism at the level of technology and social governance and social organization. So I wanted to start with this uh quotation from Peter uh Peter Teal of Palunteer which um I want to negotiate throughout the talk. Uh Teal writes that technology is this incredible alternative to politics to change the world without having to convince, beg or plead with people who are never going to agree with you. So uh we hear a lot about fascism these days. Um, oftentimes this is this becomes associated primarily with a kind of ethnationalism which I don't want to discount but I don't want to forget that the early definition of fascism was much more simple.

It involved a merger of corporate and state interests often for the sake of militarism, imperialism. So when Mussolini when Mussolini imagined fascism in the 1920s and30s um it was a matter of bringing the corporations in civil society um out of a position of being antithetical to state interests and mobilizing for the state for the sake of a totalitarian state and a militarized project by which they through their complicity with state interests they would receive lucrative contracts etc. So um this comes from the idea of the fascis as a kind of bundle a weapon a punitive weapon from the Roman Empire. Um [clears throat] so the task today uh as I'm going to propose now um is a very specific one. It's about um technocratic technological companies like of which Palanteer is not the only one but is an exemplary case in which surveillance capitalism uh the scope of it um as a enter as a business enterprise is now merging itself with state military and a and repressive interests.

And the ultimate question I want to answer today is what are the ways in which we can actually do something concretely in practice to defend ourselves against this new form of technofascist power. So I want to take Palunteer as a model. Um it's become quite fashionable to uh be upset about Palunteer, rightly so, but we shouldn't forget that Palunteer is a company that largely emerged from the the PayPal mafia situation. Um so Palunteer first got Palunteer developing AI um AI software first kind of realized what it became today around the 2011 terrorist attacks in the United States where it hypothesized that the AI detection software that they were using to discover fraud on PayPal could potentially be deployed to locate uh suspected terrorists uh in and around the September 11th attacks. Um but the problem was Palanteer did not have access to the kind of government databases that would have allowed it to pursue this work.

Uh 10 years later um those kinds of contracts emerged and today Palanteer now holds tens of billions of dollars in government US government contracts uh largely from the US Army from the DHS um with ICE, the FBI uh and the fact that one of its first investors uh in Kelto uh was is a kind of the venture capitalist arm of the CIA. Uh so so we don't forget the um did I skip a slide? So we don't forget the kind of human consequence of this. Let's take Gaza as a model. So we're still we know that um Palunteer's AI technology has been integrated and used by Israeli state militaries in their lavender and gospel programs.

We don't quite know the extent because these are government these are government intelligence programs. Um but we can still look at the consequences of that. So following the following the terrorist attacks in Israel in 2023, uh the Israeli military launched one of the most massive bombing campaigns against Gaza, the occupied territories. But something fundamentally had changed in the art of warfare with this event. And I think it's worth thinking about regardless of what we think of the conflict.

It's worth seeing this as a virtual model of the way modern warfare is changing. So concretely what had hap what happened? Um so we know that the Israeli military has long maintained AI run databases of potential targets. We call these the Lavender program but also a program that's called the Gospel and others. Um the problem was these um these databases of potential targets of Hamas leaders uh mid-level junior uh militants uh were largely exhausted within a matter of days of the Israeli military's bombing campaign in Gaza.

Uh so there was a rapid there was a rapid mobilization um by in by Israeli intelligence services and uh military officers to expand the lists. So with the with the help of these AI systems, the lists were quickly expanded to include a lot of mid and low tier um Hamas members ranging all the way down to like tunnel guardians to cash uh to cash transfer participants um to messengers, kids on the street um in 620 sectors. the bombing campaign was spread. Uh the rules of engagement which at the maximum before permitted about 10 uh civilian casualties at a maximum were suddenly raised to 20 as the baseline. Uh so what this meant is that uh these targets were sorry one second.

Uh what this effectively meant is that um in order to accelerate the kill list um a wide ranging cross database intelligence operation allowed um military targets or military personnel to target individuals based on cell phone data suspected location. Um oftent times with delays of up to 7 hours. um they set aside uh typical warning signals like rooftop taps that would allow people in buildings to escape uh before the bombing commenced. Um so we saw I think we see this kind of really intense incubator for what metadata surveillance which we are all undergoing um across the world is fed into automatic weapon systems um with automated automatedly and AI generated kill lists um with very little human interference or adjudication [clears throat] in the United states. Uh Palunteer is largely responsible for um building up AI generated databases of a similar scale which are being used by ICE uh from originally from a prototype they called Falcon uh which is now a kind of homebrewed immigration OS device that allows uh ICE officers to have a kind of live uh hunting device that pulls on uh driving driver's license records, tax records, medical records.

um phone records uh to target potential immigrants um in their live locations. So the important part of this is exactly what Palunteer excels at um AI AI AI database analysis statistical analysis but drawing from databases that are largely um non-public medical records tax records um and all of this is being sort of outsourced to a third party company. So this is the radical change. This is the change at which um technology companies are not being an alternative to politics. They're being integrated within a political state um in order to both fur further the profit interests of those technologic companies and also the state military and police apparatuses uh that enable it.

[clears throat] So why do I propose anarchism as a way in which to address this distinctly new form of power? So by anarchism we simp we we certainly do not mean some sort of chaotic situation. By anarchism I go back again to the original definition as a form of self-organization especially in local rather than grand centralized state formations. uh for the self-defense against tyranny of against hierarchical forms of power and the fabrication of a way of life or a coexistence um without centralized forms of control and in this case without centralized forms of surveillance. Crypto is obviously one space and we can see the instantiation of something like a modern form of anarchism.

But I'm going to be focusing on the underlying communication systems uh through which a lot of the metadata of our personal lives is leaking and being scraped and extracted uh for interests that are not our own. So rather than Teal's idea that technology is an alternative to politics or what they call a kind of ability to escape from politics in order to exercise governments at a different level that's not an impass. I like to propose that right now with technofascism technology is an accelerant to politics particularly an accelerant to increasingly totalitarian politics. But at the end of the day, the onlogical reality is that the that people, human living beings are both the source of political legitimacy and the target under which political control is exercised. And anarchism is a return to people as that source of political resistance and alternative forms of political existence.

So we shouldn't forget that under technofascism there's a kind of double capture. On the one hand we're being captured continuously by technological enterprises um whose uh primary goal is to extract data from us that can then become modified. And at the same time, we're being captured by uh totalitarian state forces of varying degrees uh which attempt to control the movement of our bodies, the movement of our thoughts. Um and both of these forms of capture under technofascism are now complicit. It's a form of it's a kind of it's a kind of alliance.

So what we need and what we propose is we need a sort of anti-palenteer. um not just the grass not just the grassroots activism that was there from the beginning against Palanteer that forced them to change the locations of their headquarters multiple times due to public uh uprest. And not just the grassroots activism uh that we see in places like that in places across the United States in London particularly that are where people are organizing to refuse um governments giving contracts which is to say U to give uh actual human tax dollars to these companies in the form of contracts um so that they can have access to public and personal records. Those forms of activism need to continue. But uh in the spirit of anarchism, we also need forms of technology, multiple forms, not just one.

Um that can that can actually counteract um the way in which we're being targeted from the ground up. So that's what I'm interested in. I'll skip that. The idea is to sort of try to starve, poison or pollute um the collection pool from which our data is extracted. So that um or so that given that all surveillance kind of targets us individually and collectively as a signal to be found within a noise of internet traffic or noise a noise of digital traffic.

um to see the task as introducing enough noise into the system that even AI learning models have trouble reconstructing meaningful patterns or the point at which becomes too expensive. So the proposition is we sort of need um a thousand types of noise generators or jammers. So, NIM is one. Um, it's very different than a normal VPN, which is a kind of C which is a kind of centralized proxy. And the problem with centralized proxies is that uh they make it uh relatively easy to reconstruct a pattern even even if a even if a pattern is interrupted at one point.

You simply need to observe traffic coming in the shape of traffic coming in and coming out to reconstruct the metadata trail that links you back to what you do online. Tour is a bit better but tour still has one has a very significant problem is that your traffic takes a single route through a decentralized network. So NIMD does something different and we call it a noise generating mixnet. How many how many people have heard of the mixnet nims [snorts] mixet? Got one.

Okay, great. So uh mixet is actually quite old. It's a bit it's web one. was uh the idea was hypothesized by David Chum in 1982. I think Nim there's been multiple kinds of mixed nets that were developed from reailers in the 1990s.

Um the cyberpunk movement was a big fan of mixnetss. So NIM uh developed the first generalpurpose mixnet that is available through a single VPN app. Uh so what does it mean? What does a mixet do exactly? I'm going to have to replay that.

This is too fast. Basically, think of a mixet as a proxy system of five hops. You have an entry and an exit. But in the middle, you have three layers of nodes that we call mixed nodes. So in a basic mixet, when your traffic passes through the a mixed node, well, so the first condition is when your traffic leaves your device, all of your packets are identical.

They're all the same size. They all use the same encryption format. Um, so you can't deanonymize packets based on their shape. And then when they arrive at a mix node, each mix node shuffles your packets together with the packets of other people. So when they leave a node, any given node, any observer of that mix node will not be able to tell your packet from my packet apart.

And this happens three consecutive times. And each time it's also on a MC node, there's a randomized delay, milliseconds. Um, so that an observer also can't trace a single packet based on timing signatures. And to make matters worse or to add more to make matters worse for surveillance or to add more noise to the network, we also introduce covered traffic um empty packets that uh look identically from the outside to your packets so that you cannot be a don so you can't be deanonymized based on the amount of traffic you're sending at any given time. If you're not sending traffic, if you're idle and sleeping, the app is still sending traffic from you.

So it seems like a continuous traffic pattern. So all of this we all of this cover traffic timing delays packet mixing we call the introduction of noise into the network and of course the more people who use the network the more anonymous every single individual is like increasing the size of a crowd. This is my visual representation of what I imagine the internet looks like from the outside where sort of inputs and out inputs and outputs can't actually be distinguished from one another. Still working on that visual design though. And maybe I'll stop for questions.

Thanks. [applause] Is there a moderator? No, they left. Okay. Any questions before the next talk?

Yeah. Thanks. Uh so right now we've launched it as a commercial as a commercial VPN app which has two modes. So the mixnet is one of the modes and then the other mode is just a normal two hot not a normal but a a kind of decentralized VPN mode because mix nets can be quite slow. So you don't need a mixet for all the kind of everything you're doing online.

And so it's really ideal for it's really ideal for crypto transactions, private messaging, email, uh certain high threat model situations and for everything else. You stay protected, but like way better than any centralized VPN can in a way that reveals nothing uh to NIM or to anyone looking at the network. Uh we we we have a community that's developing router support. Uh so that's still in the early stages. Um we're looking for more people to help us with that because we're focused mainly on the app right now.

Um, so as long as the app is running on the device, it'll cover everything.

Yeah. So it's on a it's a as a VPN on all uh on all platforms.

Yeah. Hey Yeah. So all nodes, so NIM doesn't run nodes that route traffic. Uh all nodes are run by by uh sort of commun independent community of like privacy enthusiasts around the world. Much similar to the tour model, but except these nodes are incentivized by the NIM token.

So we have a tokconomics model. Uh it's not just simply a volunteer-based system. Um there are always there are always legal risks that come with a with running an exit node. Um but the important thing to keep in mind is that in a decentralized system um nodes can never nodes can never contain any meaningful information that could be useful to a law to to law enforcement or to anyone demanding records because they can't connect what an individual any individual to what they're doing online. Um we do have legal support and we have ways of uh so uh they sign no log so they sign no logs agreements.

Uh that's also a kind of legal protection in some jurisdictions. Um does that pull you answer your question though? I mean the the Well, it would stay it would still be in it would still be it would there would always be a layer of encryption uh at every node point. It would only be de- enrypted. the the final payload would only be de- enrypted at the the the at the um point of delivery.

In terms of content, the content of the payload. Yeah. The Yeah. Like any like any proxy system, the uh IP address of the exit will be the one attached to the delivery at this point. Yeah.

one of them. Yeah.

Well, yeah. So, I would think that so one distinction I would make is when it comes to privacy technology, um, and this is why this is why I think my our idea of noise is a is a distinguishing category. So think does the privacy technology that I'm using protect simply me and my activities or is it also contributing something to the anonymity and privacy protections of others and I think it's in the second one that noise introducing noise into a network has this interesting effect where you using you using the network is actually contributing something to the network itself um that concretely can help uh the anonymity of others and so I would propose like are there other systems that can be designed that can function like this also not just not just us. Yeah, good. Cool.

Thanks. Uh if you want free passes to try, just come find me. I got some, I think. Uh, thank you.