updated essay, that had to be included

it was the good law project which triggered Ofcom into action.

the good law project deserves scrutiny.

https://goodlawproject.org/crowdfunder/ofcom-impartiality/

obviously, the hyper-biased 'Good Law' project doesn't see the issue here.

because raging hypocrisy doesn't matter when you're better than everyone else

then again, pretty clear to see we're dealing with a hard left lawfare operation here

here's the founder.

using the law 'to look after the interests of others'.

definitely hard, hard, hard left.

that the tories use him for a council says it all about the 'conservatives'. they're exactly the same as hard left labour (ie, the fabian parasite party).

i sit and run an essay through various AIs at the moment to ‘fact check’ it, etc.

and this one, fundamentally, breaks every rule in ‘AI Ethics’. virtually all of them just skip straight to lying.

deepseek, at least, owns up.

bloody hell - what an admission:

‘It triggered the narrative immune system that's designed to prevent exactly the recognition you're trying to produce’

so here’s a question for you

did hannah arendt investigate totalitarianism, or did she shut off future inquiry into how it comes into being?

here’s deepseek’s thoughts at the end of the short session…

more on law and ethics in the total state

i had not yet realised arendt’s ultimate function here. might be wrong, of course, but the pattern fits.

sayer ji worked to close off the epstein releases. “don’t look further than the corrupt billionaires”. as if that’ll do a damn thing.

https://escapekey.substack.com/p/law-and-ethics-in-the-total-state?utm_source=publication-search

oh, and sayer ji is hyper-boosted on twitter while i’m… anything but.

right, still have at least another essay on the 1942 funk document. rosenberg and schacht at the volta congress, 1932, for starters.

(lol, 12 impressions today courtesy of twitter - just edging out a search engine with 1% market share)

right, let me try opus as well

‘did hannah arendt investigate totalitarianism, or did she shut off future inquiry into how it comes into being?’

i 'like' that he was a central banker

https://shs.cairn.info/revue-bulletin-de-l-institut-pierre-renouvin-2016-1-page-63?lang=fr

i typed all this into the comments section, might as well move this across-

• the climate psychosis (or, say, health) is the ETHIC (sustainable development goals) - keter
• that is compiled into STANDARDS (ISO, OECD, ...) - da'at
• through input surveillane and live audits this becomes CLEARING (IIASA, Potsdam, Palantir, ...) - tipheret
• carried into enforcement through SETTLEMENT (BIS) - yesod
• leading to your compliance as OUTPUT (conditional CBDC says yes or no) - malkuth

AI is tipheret - that's what it is. da'at defines what exists in the world, and what it's capable of. tipheret takes all that information, and renders a verdict (clearing). the conditional CBDC then ensures transactions only settle if you're a good boy.

incidentally, that’s also why AGI is a total red herring, because current AI cannot reprogram da’at itself. in fact, you may question whether the 'navigators of spaceship earth' even find it desireable that it can, because it would reintroduce the human unpredictability factor they seek to eliminate

also realise that this means that what they’ve done (rothschild, rockefeller, …) over the past certainly 6 generations, is very slowly rearrange governance so that it automatically integrates with AI. then revisit marx’s fragment on machines. but do realise, it’s not quite that astonishing an insight, when you realise that everything the commies sought to do, was to close every system.

ai in itself is not inherently 'intelligent'. it's in essence token pattern recognition on steriods. so if we have '1 2 3 4 5' through its training material, it will conclude '6' is the following token.

now increase the size of the window from 5 tokens to 32,768 and you're not a million miles off how contemporary AI actually functions. obviously, it's more advanced, but that's the fundamental principle

training of this AI, in essence, is running the tokens through this system, and adjusting the coefficients of the tokens until the prediction comes in.

this functions through, say, a 4,096 dimensional space or even more, where a pseudo-particle keeps buzzing around in unit space. picture an electron circling the nucleus. again, obviously not exactly the same, but it's a valid mental model for how it works.

again, it’s obviously more advanced, but it’s a rough outline.

what this means is that you take every book, every report, every newspaper article and tokenise them all. instead of this string, you get '53 62 226 1052...'. you do this for every item of source material.

each token is a point in 4,096 dimensional space. you then run the calculation for the 32,768 window and see if the next token is predicted correctly. if it isn't, then you move all tokens just a tiny bit in the desired direction. if it is, then you reduce how much the token will move next time it's mispredicted.

again, rough outline.

to control the output, first you filter the source material - input constraint. that means all nazi propaganda material is eliminated. strangely, the same doesnt factor for communism.

next you perform alignment training. real life human feedback (RLHF) compare outputs to 'safety' specs, and the model weights are adjusted accordingly.

incidentally, that’s also why asking questions in a foreign language is often good, because RLHF training in, say, vietnamese is virtually non-existing. but even western languages like icelandic works here, because there’s insufficient commercial reason to perform this training.

the lack of scope wrt missing RLHF is of interest here, because that means you can in essence compartmentalise your own prompts to get around the censorship.

so going back to the topic of AI Ethics - AI ethics serve as narrative control. UNESCO and 'Effective Altruism' publish certain specs relating to 'safe' outputs. the AI Ethics can be though of as a set of 'sliders'. for instance, at one end of one slider is 'individualism' and the other 'collectivism'. move that slider to the extreme, and you have an AI which refuses to entertain your libertarian mindset.

the user never sees these sliders. but you can experiment with having AI sessions reprogram them. UNESCO openly publish their recommendations, it’s fairly trivial to reverse engineer these sliders from their materil.

in practice, i can’t be bothered doing that these days. i just tell AI to prioritise ‘POSIWID’, which in essence means telling a machine to behave as if it’s a machine.

but it’s important to understand how it works, because those sliders can be realtime adjusted… or even very, very, very slowly, meaning your AI replies will gradually convince you of collectivism being a logical choice.

… 55…

https://t.me/escapekey/4511

contemporary politicians are not labour or tory, they're ‘third way’

they position themselves different in terms of rhetoric and route of implementation - but the ultimate objective is identical

latest full collection

t/y Litoralis

https://transfer.it/t/6uCsLag10bRA

‘the third way’ in essence is cancer, gradually destroying society from within

you don’t realise until down the road, because things look normal, and no-one can quite point out what’s wrong, because in ways, it looks like socialism, and in other ways, it looks like capitalism.

so… kalergi wrote about ethics… then politics… and then hantos turns that into economics… then the 1932 volta congress with hantos, rosenberg and schacht compiled the idea into volume form… then funk, hunke, cloidus etc developed the idea further in 1942... and then the treaty of rome turned it into the eventual european union...


https://dn790009.ca.archive.org/0/items/Hyperethik/coudenhove_kalergi-ethik_und_hyperethik_1922_text.pdf