This is a great topic and one that I think anarchists should think about a lot.
I just finished reading Domination and the Arts of Resistance: Hidden Transcripts, which argues, among other things, that various forms of resistance are developed and encouraged in spaces that are out of the view of institutions of power. It was written in 1992 and his historical examples come from his field work in Burma, and the US experience of slavery, so quite different from our present techno-dystopia, but still interesting. Are there internet sites that are conducive to the development of hidden transcripts, and to what extent are they surveilled by our enemies?
One example of an anarchist strategy that contributes to illegibility is the refusal of demands and leaders. One important historical strategy of the state is to channel revolt into something it can negotiate with and co-opt / destroy. There are an abundance of examples of states and their security forces basically appointing a leader within a leaderless protest movement / revolt, or elevating a wannabe leader. Anarchists have done well when they have successfully avoided these maneuvers.

I love all of Scott's books that I've read, and I'm interested to hear what people think of this topic.

Maybe people need to listen to this: https://www.youtube.com/watch?v=gl9DgAM2zSk
Maybe people need to listen to this again: https://www.youtube.com/watch?v=gl9DgAM2zSk

I don't believe this leviathan stands as strong as one might think and the speed it moves is likely untenable and much sooner than later it's gonna falter. I see a lot of denial out there right now and real impossibility. It seems that big cracks will likely appear sooner than later. I could be wrong but, I hope not.

Han Byung-Chul writes a lot about digital transparency and how it has lead to a Psychopolitics worse than biopolitics. He hates social media and the cult of the neoliberal self. I recommend his very short book Psychopolitics for a reading on what digital legibility means with regards to a self imposed Panopticon and the loss of the Other as we all look for Sameness to be cultivated but with “diversity” of identities. Very convenient for all the marketing making use of metadata.

i like the topic, i HATE the terminology. legibility? doesn't work for me in this context.

i am incomprehensible
i am uncontrollable
i am illegible?

Has anyone noticed the way the Zapatistas have disappeared from the global media? They're still there, and maybe stronger today. But, they stay relatively safe by staying invisible, avoiding coverage. There's a recent article in the Guardian where they allowed a few questions to be asked. But mostly, they keep it all as secret as possible. It's a strategy straight out of Hakim Bey. The same thing is found in Graeber's work on Madagascar. Who knows how many of these sites exist? I have occasionally heard of free parties and festivals going on underground in Britain. They aren't being repressed because the state doesn't know about them. Just occasionally, Vice or the Guardian find one and run a story. But mostly, they're invisible.

Today, I'm seeing four main strategies emerging. First, people are abandoning (or never opting into) social media and Web 2.0. The strongest form of this is the off-grid movement. This is a lot bigger than people realise – hundreds of thousands in the US by some estimates. Weaker forms are even more common – people living marginally within less-regulated areas and avoiding having traceable identities. There are millions of undocumented migrants living this way in America. It's also how most career criminals operate. See McKenzie, “Getting By”, the chapter on the “missing men”. To reduce police attention, men involved in crime avoid having a fixed address, a doctor, benefits, or state registration. They live with relatives or partners, often between several addresses, and rely on the relatives or partners to supply any medicines they need. Second, people are staying connected but using increasingly sophisticated encryption and other technologies to reduce surveillance. This is a huge trend on various scales since Snowden. It's taken a blow in anarchist circles since Anonymous got largely broken up. But technologies keep improving, and mostly, the state is losing. Hundreds of thousands use drug marketplaces – including long-term sellers dating back to the original Silk Road – and are never caught. There has just been a case in Britain of a man jailed for abusing children on an industrial scale, and posting about it on the deep web. Police have been looking for him for five years, and at one point there were 100 police on his case, across several countries. They only caught him when they managed to identify one of his victims offline. Cybercrime has moved to Russia, where local users operating internationally are pretty much safe for now. The third strategy is the hikikomori strategy – people who stay inside, stay in one place, rarely or never go out. They live on the internet and rely on the anonymity techniques discussed in point 2. The fact that they're not out in public reduces all the threat surfaces except the virtual one; and a more extreme version would involve disconnecting from the web as well. Finally, there's people who embrace visibility and seek to manipulate it. This is particularly common in idpol and the alt-right. The fact that they're so visible in plain sight, but blend in with a large, mutually protective crowd, renders them somewhat safe – and hard to separate between the truly dangerous and the largely innocuous. The spree-killer who posts threats on social media before they act is missed because there's so many people posting non-actionable threats on social media, and too many complaints and warnings for the state to filter. The trick here is to blur the meanings of discourse to such a point that one's permanent visibility does not render one legible, but rather, renders legible a set of signifiers prone to constant emptying of meaning.

Legibility to the state is limited by its reliance on external appearances, quantification, and existing, already-observed patterns. It cannot see what's going on in the inner life (e.g. motivation), and its generalisations are clumsy. This is, historically, what's created the gap between “the map” and “the territory”. Technologies are now more precise in creating a growing pool of data and images, but this evidence is meaningless without being processed and interpreted – a duplicate simulation of most of the territory does not perform the same role as a map. Since there's too much data for humans to handle, algorithms are brought into play. But algorithms aren't very good at spotting complex patterns, at least at this stage. There's an excellent article I'm having trouble finding again – called something like “Drones have Blindspots Too”. I think this is great in countering the paranoia which the state (probably deliberately) induces through surveillance. In many ways, the biggest danger is not that the algorithms will actually catch anarchists doing illegal stuff, but that they are broad, clumsy, and leave lots of blind spots. They miss real dangers, but they also generate dozens, hundreds, thousands of false positives for each real finding. And with uncontrolled police powers, broadly defined laws, and gullibility (or malice) of judges and juries, profiles become self-fulfilling prophecies. The person “fitting the profile” will be framed, or accused on flimsy evidence, and wrongly convicted. And this will feel back into the profiling mechanism, making it more “robust”. For people who are known anarchists, or part of visible subcultures, this is an extremely dangerous situation. Being someone profiled as “a risk” for planting bombs or sabotage or rioting may be as bad as actually doing it and getting caught. And someone who does something in a way that's never been done before will never be caught by an algorithm. At most they'll stand out as eccentric – and often, not even that. Legibility only extends as far as existing patterns. It's constantly vulnerable to unexpected tactics and escalations.

Still, it's important not to be too pessimistic. New technologies both increase and reduce legibility. Surveillance drones make it a lot harder for off-grid communities to survive simply through distance. CCTV and face recognition allow real-time profiling and pre-emptive repression. A lot of the cracks and fissures which anarchists used in the 1960s-1990s have been reduced or closed. Legal squatting, dole autonomy, hoppable trains, unfenced sites to park a van on. Repression is often enough to finish off a dwindling movement. But where the social force driving the movement is still active, it adapts to the new terrain, eventually flowing around the blockages. We are seeing this now with counter-terrorism, and immigration, and spree killings. The system thought it had fixed these in the mid-2000s, with securitised borders and public spaces, deterrent measures, and profiling. Around 2010 they were pretty pleased with themselves. Immigration to Europe was way down. There hadn't been any more 9/11s and most people planning big attacks in major cities were caught before they could act (assuming they were real plots and not frame-ups). But the “problems” are now worse than the mid-2000s, or the 1990s – because the causal factors have become worse, and actors have adapted. Terrorists use guns, knives, cars, and act alone – so the old ways of catching them through network mapping and tracking precursors, or infiltrating cells, no longer work. Since the Snowden revelations, more people use encryption. The FBI is afraid that the internet is “going dark”. Anti-laundering measures have simply fuelled the growth of cryptocurrency. Crackdowns on drug imports have led to increased local production, often in remote or abandoned locations. In the Syrian crisis, the push factors simply became so strong that the borders no longer worked. If the system survives, I'm expecting to see a lot more counter-technologies with anarchist uses in the near future. 3D printing, autonomous robots (think robotic tree-spiking), EMPs, DNA editing, fingerprint editing, realistic prosthetic masks indistinguishable from real faces. Already the deepfakes app can be used to alter images from crime scenes. In the short term, this will be used to frame people – but once people become aware of the problem, it will make CCTV evidence far less reliable.

Tactical innovation is also important. And here, the contradictions of idpol are most apparent – celebrating “transgression” and “disruption” while in fact running very predictable scripts. These scripts are often the ones which disrupted capitalism in the 1960s-70s. But today, we need to find ways to disrupt scripts which are prevalent today. For example, lockdowns are more vulnerable than they seem. What would be the fallout if a half-dozen peace protesters turn up during a lockdown, protesting militarisation of public space – making sure the media and social media are watching the whole time? Either the police would tolerate the protest, or they would attack viciously – and either way, it's a PR disaster. And all it would take is a handful of fearless people. Or, ahead of a big political trial relying on CCTV evidence, an anarchist quietly releases a believable deepfake implicating a top pig or politician in some kind of crime. After about a week, we out ourselves – and justify it on the grounds that we're showing that CCTV evidence can't be trusted. We need more of this kind of thing!

& Alice Becker-Ho's Essence of Jargon

Both address this subject with lucidity..