I am part of a citizen’s collective that promotes direct democracy. They did not theorize the narcissism-authoritarian link explicitly, but gave a few tricks that worked here to make some bad actors flee:

• Avoid one-on-one conversations with potential bad actors, have a public channel and put things there even when they insist on communicating directly (they hated that)
• Talk explicitly about how positions of powers are going to be distributed but also how they are NOT going to be. The earlier you have them tag you as a dead end for their political career, the better.

And I think that we failed on that account here: recognize that they are going to go from friendly to hostile in the blink of an eye and be ready for it.

One thing I will do differently in the future is that I will not waste too much time with people who can’t clarify their positions and disagreements.

Ce ne sont pas des armes de destruction massive. Oui ce sera utilisé de façon ciblée par des tueurs de la même façon qu’aujourd’hui ils mettent une bombe dans une voiture. C’est déjà faisable pour quiconque peut acheter une grenade à main, un drone civil et est un peu bricoleur. Ça ne changera pas énormément les choses. L’impact est bien moindre aue la prolifération d’autres technologies

I am really happy that this question led to so much elaboration. It does come from a person I know IRL who talks a lot about the psychology of power structures, having had to deal with too many psychopaths himself. If you are interested in the profile of authoritarian followers, which is different from leaders, there is an abundant literature on RWA profiles (right wing authoritarianism, but a bit ill named as stalinists followed similar patterns)

Power/authority needs to be based on trust, and it needs to be lost at the same instant as the trust that supports it is. The overhead of getting everyone together to hold a vote of no-confidence is way too high.

We should reverse the logic of the ‘signing onto law’ where a final formality gives a president, a chancellor or a queen an actual but rare veto power.

There should be something like a representative assembly that has to give a ‘go’ vote for coercive power to be exerted. Nowadays it can be very lightweight: remote voting can be secure easily if it is not anonymous (representatives, one can argue, should vote publicly).

It should be almost automatic when trust is there, but if it is absent, mere doubts should be enough to block an action.

We would live ina very different world if the representatives of a neighborhood had to give the ‘go’ for a police operation

can Anarchists build and maintain public infrastructure?

Internet

Wikipedia

Many open source projects

One could argue that international research efforts are generally done in a non-coercivie way

Anarchist ways can maintain public infrastructure, but they need to be built differently with that modus operandi in mind.

Over the past years, reading more about the dark triad/quadriad, I am becoming more and more convinced that authoritarianism is the political expression of narcissism and that it is 100% of the explanation, that there is nothing more to it. Want to fight authoritarianism? Stop narcissist. It is not a matter of ideology, of left or right, of reformist vs revolutionary, it is just a matter of psychological profile. Stop the narcissist, that’s all.

How do we build systems that are resilient to sabotage without falling into authoritarian logic?

I had a eye-opening moment with this videp, whose title (“Can 100 people self-organize without a leader”) is actually misleading, as it (IMHO) failed to demonstrate what it wanted to test, but demonstrated something much more interesting. The task given to 100 people was too simple to require multiple people (a “hack” they forbade has shown that one person was enough to do the full task) yet, a hierarchy “naturally” emerged. Even though the sample population is biased towards people who would not be very hierarchical.

My main takeaway was that an organization that does not want a hierarchy does not only need to make it possible to self-organize, but needs to actively “weed out” hierarchies. That’s hard, I don’t know of any examples of it.

La solution est simple: rendez explicite la responsabilité juridique et criminelle quand ces systèmes font des erreurs. Et rejoignez la CPI, putain. Au lieu de l’handicaper de toutes les façons possibles et mesquines.

Une règle ou contrainte sans moyen de l’implémenter ne sert à rien.

It surely sounds like it. Which is annoying if they are crying out wolf, because China can and probably will (or did) put backdoors in its equipment.

Batteries pack with radio is weird. But really I don’t understand why we don’t have technical details.

Yes, well, from what sources do you gather that?

Et comme l’article initial ne trouve pas important de mentionner l’auteur de ces illustrations, le voici: Mikhail Belominsky https://fishinkblog.com/2012/06/26/mikhail-belomlinsky-childrens-russian-illustrator/

(I hate it when a technical take makes me side with authoritarian propaganda, but well…)

There is zero technical information in that article, yet plenty of people jumping to politically-loaded conclusions. Reminds me of the time when there was a (totally legitimate imho) scare about Huawei backdoors but zero technical details about what was actually found.

So from what I understand, some inverters “phone home”. A despicable habit of too many hardware in the industry, but the phrasing suggests without even confirming that it may be more nefarious than “mere” telemetry that plagues any connected device out there.

“Rogue device” suggests that it is additional hardware. They imply that the add connectivity channels that were not present in the device. Are we talking offline devices that were stealthily loaded with a 5G simcard or a Lora device waiting for a bricking code? It is implied but not stated, which makes me extremely suspicious.

If Chinese authorities can remotely brick solar inverters, it is a matter of national security to disclose the models and the modus operandi asap. It is irresponsible to not help us mitigate the potential of attack. Also, if there are “rogue devices” designed to sabotage your grid, that’s international sabotage, that’s state terrorism. It is important to state it if it is the case, instead of implying it.

“This is a serious issue that the industry needs to address, and it’s even more reason for Congress to maintain tax credits that are onshoring the production of inverters and the entire solar supply chain in the United States."

I suspect that this is the core reason actually. Don’t get me wrong, manufacturing crucial equipment locally is definitely a good idea, but I suspect strongly that these accusation are just a way of dodging the embrassement that Chinese companies’ market share is annoyingly high in a market that westerners were too slow to recognize as critical.

C’est à la fois hilarant et inquiétant. C’est hilarant parce que ce genre de manipulation est faisable d’une façon beaucoup plus subtile et que là c’est un boulot d’amateur complet.

C’est inquiétant parce que il a une équipe derrière lui qui est capable de faire ça de façon bien plus compétente.

Mais c’est une chose étrange que ça a en effet l’air d’être difficile de rendre ces LLM plus racistes qu’ils ne le sont naturellement, qu’on a du mal à rajouter ce genre de biais.

Je ne me lasse pas de raconter l’histoire de Grok. En fait, ça vient du fait que Musk a l’origine critiquait (avec raison) le fait que les LLM tels que ChatGPT ou Claude sont obséquieux, politiquement corrects et refusent de prendre parti sur des questions politiques. Ce sont des modèles entraînés par des boîtes américaines et la norme là-bas c’est d’être le plus mou possible quand on parle de politique, de froisser personne et d’avoir des clients qui sont à la fois conservateurs et libéraux.

Du coup, Musk a proposé d’entraîner son modèle sans la partie diplomatique et politiquement correcte, en pensant que si tu enlèves la bien-pensance (forcément gauchiste), on aurait un modèle qui est aligné sur ses valeurs à lui. De façon hilarante, ça crée un modèle qui est le plus à gauche quand tu le testes sur des choses comme le political compass, et qui n’a en effet pas la langue dans sa poche lorsqu’il s’agit de critiquer des milliardaires tels que Musk. Il se trouve que le dataset d’entraînement principal de ces choses-là, c’est-à-dire un peu tout Internet, tout Reddit et compagnie, est en fait très à gauche (du point de vue de Musk)

La Chouette dort

L’un sans l’autre c’est comme Luke sans Lucky ou Jumper sans Jolly

Bon ben je vais candidater on dirait!

slrpnk.net (l’instance Lemmy que j’utilise) utilise et promeut XMPP.

Mais oui, l’expérience que j’ai eu de XMPP recoupe la tienne. Moi c’est Element/Signal que j’arrive à promouvoir pour qu’on évite d’utiliser Whatsapp et FB. Je pense que je peux pas descendre en dessous de ça niveau utilisabilité sans les perdre.

Si tu ne connais pas, le cycle de la Culture, de Iain Banks décrit un univers comme ça. Les vaisseaux sont autonomes, embarquent ou non des humains pour une variété de raisons.

Oui, cette “utopie” décrit déjà la réalité d’une classe privilégiée.

Je pourrais enfin passer plus de temps à m’occuper de l’éducation de mon gamin, au lieu d’avoir à remplir le frigo.

Ensuite, après avoir trouvé la Chouette d’Or, je passerai mon temps à créer des jeux vidéos artistiques, et probablement me reperdrais dans le monde de la réalité virtuelle, s’il se débarrasse de son coté corporate.

Je participerais aux inévitables débats politiques sur la place respective des humains, de leurs habitats, de la nature. Je pense qu’une fois le massacre de la biodiversité arrêté, on va parler de “rewildering” et se rendre compte que plein de débats moraux, philosophiques et politiques se posent!

Si mes compétences peuvent servir à la recherche médicale, j’aiderais là aussi…

Rah! Le drama récent m’a fait me replonger dans la chouette d’or après deux décennies d’abstinence ! Je me suis couché deux fois à 3h du mat en explorant la piste Noirmoutier que j’avais in peu hâtivement écarté autrefois !