If you look at war economies as the norm and consumerism as an attempt to sustain the economy in peacetime, the problem with consumerism becomes obvious: it has no objective measure of progress. Most of our economic institutions and stats were created for fighting wars but they're incomplete. War provides many objective measures of progress: the front, consumption of materiel, weapons testing, etc, so this is not a problem. It becomes a huge problem in peacetime.

Liberal democracy is essentially a coalition of conservative movements that form a united bulwark against genuine modernity. Its ‘progressive‘ credentials come from incorporating foreign movements (‘multiculturalism’) and subcultures (LGBT, etc). But they’re all anti-modern. Genuine modernity would simply accept that political arrangements are arbitrary and that language and culture are artifice and of purely instrumental worth. Liberal democracy is just a collection of groups that oppose this for whatever reason. An anti-modern coalition.

Social theorists massively overestimate the role of conflict and competition in people's lives. Our lives are overwhelmingly dictated by routine and we very rarely engage in conflict or even competition of any kind. Even when we do engage in competition it's highly constrained. For example, when competing for jobs or other positions, we generally don't even know who we're competing against or have any ability to affect them.

Because liberalism has such an implausible model of individuals, institutions, society, etc, it, for a long time, seemed to make good on its promise of 'neutrality'. Culture, in liberal societies, was largely left to the whims of elite philanthropy. They were fairly conservative. Thus, Christianity and Christian mores survived the secularization of the state because liberal societies are actually dominated by elite philanthropy and elite philanthropy remained nominally Christian and geared to 'moral reform' along nominally Christian lines.

AI doomer's don't seem to be able to model their opponents well. If an AI is not a mind, it will still be subject to reliability and misuse issues, but not as a mind. So it won't go rogue or deceive you or plot against you, it will just fail.

https://twitter.com/dwarkesh_sp/status/1668717423963557888

"what do you think happens as artificial neural networks get smarter and smarter?" They don't get smarter and smarter because they're not minds.

In the ‘Sparks of AGI’ paper they get GPT-4 to draw a unicorn in TikZ, but if you try to do it now it protests that it can’t draw pictures. However, I got it to draw progressively better wine glasses by converting descriptions into shapes before writing the code. Basically, you tell it to start with a description of the parts of the object, then make a list of the shapes it needs to represent the object, then consider the relative position of each shape, then write the TikZ code.

We have two broad approaches to the social:

1. Treat it as a natural object and attempt to study it using what are supposedly the methods of the natural sciences

2. Claim that the natural sciences aren't exhaustive and that some other mode of understanding applies to society The first approach leads to pseudoscience, which is practice leads to bureaucracy, the surveillance state, etc. The second approach merely leads to criticism of the first approach, since nobody can define this other mode of understanding or see any feasible practical application.

Our predicament came about because getting the physical side of things right gives you a massive advantage (industrialization) whereas getting the social side of things right does not, at least until social and demographic decline set in (centuries after you’ve gotten it wrong). Indeed, getting the social side of things right NOW would only reverse decline and not offer the kind of massive advantage over other nations that industrialization supplied (except in the very long term), and it would take multiple generations to see the results of the change.

There's a perfectly natural generational progression among the rich from being interested in whatever venture(s) they made their money in to becoming, through diversification, more interested in 'the economy', and so politics and grand 'philanthropic' social engineering projects. The model people on both the left and the right often have of capitalists as greedy libertarians who are interested in their business interests is wrong. Incumbent wealth is interested in socially engineering the overall conditions needed to expand their wealth.

I can see how the pandemic would've made people more 'anti-authoritarian' and so pushed them towards 'classical liberalism' and stoked anti-China/anti-'autocracy' sentiment, but the irony is it thereby also made people more conformist (since that's the state ideology). Liberalism launders its own incompetence into renewed support for liberalism:

1. 'Lockdowns are statist/authoritarian'
2. 'We must return to our liberal values' (support the state ideology)
3. 'We must fight authoritarianism [abroad] with renewed vigor' (support empire)

Most problems aren't going unsolved because of lack of intelligence (as ordinarily conceived) but a mix of institutional inertia and conceptual confusion. They don't require faster computation or more complex pattern matching to solve, but a change of perspective. Ironically, language models might be good at this very task: resolving conceptual confusion. It's just simple pattern matching in language (comparing technical use to ordinary use). Human beings aren't so much bad at it as we just refuse to do it for a variety of reasons.

The model that most people in the West have of the world is one of individuals interacting - no power, not even any real organization - and then everything that happens that doesn't fit that model is an isolated aberration, quickly dismissed. You see this in the economy. The market is individuals buying and selling. Some of those individuals happen to be multibillion dollar corporations and sometimes they do things, individually, that go against the model, but we never revise the model. Even 'radicals' accept it.

Geniuses: ”Conspiracies are not real, it’d be impossible to keep that stuff secret.”

Guy who worked on literal NWO-style conspiracy to insert population control messaging into popular TV shows openly recounting his experience in a freely available handbook for practitioners: This is the source: routledge.com/Entertainment-…

Reading an intro to language policy. The author claims the 'globalization' of English is 'organic'. But later he talks about how international recognition for 'language rights' makes it harder to fight the encroachment of English. He doesn't consider whether THAT is deliberate. This is an obvious strategy for the Anglophone world: promote individual and minority rights in language and culture as a way to Balkanize other cultures and ensure English and Anglophone culture remain the 'default'.

The liberal project requires not only the obfuscation of power, but revisionist history and the defacement of theory. The usually takes the form of denying that planning works generally (theory) or worked in particular cases (history). Liberal regimes need to deny that planning works, both theoretically and as historical 'fact', because most planning in liberal regimes takes place outside of formal government and is illegitimate under liberalism's own conception of political legitimacy.

I think what the Twitter Files really show is that political discourse is downstream of the regime-media interface. The character of political discourse of the last couple of decades has reflected the problems of trying to get social media companies to censor the 'right' people. Traditional media and social media are fundamentally different in that the former is inherently exclusory. Manipulating the traditional media is about creating 'experts'. Social media is inherently inclusive, so you have to explicitly identify individuals/groups to be excluded.

The ad revenue model killed positive perceptions of tech progress. Everything stems from the ad revenue model. Google was at the core of it. Its strategy to avoid being seen as an ad company was 'big bets', research, philanthropy, and dump trucks of money for open source projects. It turned tech progress into marketing for marketing.

The accumulation of knowledge is not a given, it's programmatic. Different programs will produce the accumulation of different kinds of knowledge (of which there are infinitely many). This is key to understanding differences in development and the rise and fall of the West. The liberal view that the accumulation of knowledge is just what happens when you give people 'freedom' and they're no longer suppressed by 'dogmatic authority' is completely misguided (and comically inaccurate as an account of history).

Some people seem to think that alignment is just like flipping the color of the map in a game, so that Russia could suddenly become 'pro-West' and a major problem for China, but it's incredibly implausible that the US could achieve any significant control over Russia. US hegemony is strongest Europe, Japan, and South Korea, where hegemony was established through occupation. This isn't likely to happen in Russia's case. The US would adopt the usual method of pressuring the leadership to adopt policies that will let in US media, NGOs, etc.

It's interesting how EAs sort of lucked into elite status, finding themselves at the nexus of finance, politics, and charity where oligarchic power is maintained, but it doesn't appear to have been intentional and they don't appear to be cut out for it. The formula of 'earning to give', in particular, is sort of like accidentally summoning the devil.

Completely losing trust in the news media is liberating. I don't mean just losing trust in the 'mainstream media' and then following some other self-declared 'non-mainstream' source either. It's all propaganda. Trust nobody.