Hygge, the first Distrikt

Righto. Critique / suggested changes, @joriam:

Hygge (formerly Distrikt 1, Vikinia, among others) is the first Distrikt of Witness. Initially planned as “the nervous system of Witness” as a whole, it was turned into its own Distrikt after the Fracture. It is the Distrikt that holds the State Machine, and thus the only Distrikt in Witness to have a piece of shared infrastructure within its borders.

Suggest change to reflect Covenant + Library of St. Benedict (The Covenant): the State Machine is distributed across both the Library of St. Benedict and Hygge. In fact, most of its data store is in the Covenant, which has since then built an entire knowledge economy around it - which is why military troops from Hygge guard the Library. But the front-end technical infrastructure, decision models, and the key access points, are all in Hygge. If you look at the History, you’ll see that important interDistrikt Councils still meet at Newton’s Folly (The History of Witness) in Hygge.


“No, you don’t understand. He gave me an executive order to send all of his rations to the new espionage crew. Do you get what I’m saying? Denton is so into his shit, he won’t have anything to eat. God have mercy on us.” — Larry Quoia, former Second Secretary of Witness, messaging records

  • nice touch, love it and the record of Denton’s descent into chaos. Suggest grouping this under a heading titled POLITICAL HISTORY.

On the fourth anniversary of the Fracture, Denton was murdered by his former maid Agila Nakahl, who choked him with a piano cloth. She was executed the next day. Some say the incident was political, while others claim that it was a reaction to sexual abuse. The incident is widely referred to as ‘Denton’s early retirement’.

We can’t have him dying this early. The dude continues to haunt other Distrikts, especially the Assembly, for decades (and thus incidentally pushes them towards self-sustainability) Ref: The Assembly of People

The State Machine École

I like this idea. One thing to add would be the politics around this project: more people began sending programmers / weavers (the two words seem to refer to similar functionality: I suggest we adopt one term) to the Library of St. Benedict. Partly as a way of teaching, but also partly as a way of taking back some intellectual control from the Covenant, the École was established.

The triple AI

This is the area I have most trouble with. We need to think about why an AI would exist. In this case, the State Machine exists because it is far more adept at parsing complex information flows and collecting data on the societies it governs, and capable of generating optimal recommendations to policymakers. Why, then, would it make itself redundant?

This is also the most fun part of Hygge, and IMO lends it flair. So I propose a compromise: the events of the Fracture and subsequent Distrikts led to many programmer-politicians to believe that the State Machine was inadequate or insufficient, and they began proposing significant code and model changes and put them to the public for upvotes.

Facing three different proposals and a population divided between them and the fourth option (keeping State Machine as is), the State Machine opted for a middle-ground approach: it spun up virtual machines (think VMWare, but 10000x) hosting three different versions of itself - Kiri, Sata, and Goro, with its functions split among them. It instructed the public to live with them for a while and then have a referendum on which ones made their life better.

The State Machine itself, as the underlying system, would be able to course-correct in case of extreme failure situations. And it would continue managing the infrastructure. But to everyone’s surprise, people got along with the Three Faces of the State Machine, and extensive A/B testing found that the combination of Kiri + Sata + Goro were found to outperform any single option. I suggest working in a reference to Basterfield here, so that her impact is felt in the history and not just in the Notable People section.

So now we have an interestingly layered state: ordinary citizens know and get along with the Three, but programmers and engineers and those who crawl around in the infrastructure still speak of the State Machine, brooding and watching the world through its masks.

Jonas Kimura

Hygge seems to have extraordinarily strict laws. What prevents Kimura from being thrown off a building for his crimes? Can we link him to the gangs and corps of Libria: Liberty City ? And what impact has he had on Hygge? Places to draw from if you’re depicting hackers: https://www.fireeye.com/current-threats/apt-groups.html

Ok! I’ll make the changes :slight_smile:

Ok! Noted! I need to reflect upon this and create a story in which he can still be horrible, but not in control of where Hygge is going anymore. Something like MAGA operating years after Trump falls from grace.

Ok here are my thoughts!

The State Machine was created to run Witness, not to run Hygge.

If was doing a good or a bad job at running Hygge is irrelevant, as the political agitation of the time was completely us-versus-them and any hint of compromise was shunned. So the populus turned against the State Machine and the State Machine noticed — the State Machine understood the human processes at hand.

So what was the State Machine’s response? The ultimate compromise, a part of itself that is Hygge-focused. A part that is Hygge through and through just like the pro-Dentons wanted. The advent of the triple division was just part of the calculation: the State Machine understood that the current Hygge populus needed ‘teams’ to compete. They needed somebody to cheer and somebody to complain. By dividing the AI, it gave the population the language it wanted.

I wanted this separation to be a suggestion from the State Machine, because it would give the pro-Denton’s a sensation of “we were right, see? even the machine agrees”, which is exactly the power of the State Machine to understand human behavior.

What do you think?

Oh my whole spiel was that Hygge knew excatly who Jonas Kimura was, after all he was a top student from its most prestigious institution. But Hygge has no idea who Jonas Kimura is right now and where he’s at. He graduated at 24, so now he’s in his late 50s. For all they know, he could be overseeing the prison.

1 Like

Cool. Do write that argument in, since this dialogue reflected in political history makes it richer. @alberto, we may need econ support.

I wrote in a social-democratic system with strong redistribution… but then I realized there is a “resource allocating AI”! @Joriam, did you want to make a digitally planned economy, rather than a market one? À la Red Plenty? Then I would suggest to rename the AI something like Gosplan. :slight_smile:

Since we’re running short on time, and need this core out the door, I rearranged some things for flow and made some small additions and changes. I especially gave SATA more veto power over Kiri to tally with “Sata has been blocking this increase due to lack of funds.”

Few things I have removed, because they felt flimsy or did not cohere to later explanations:

“Electable projects must have teams attached to them, but these are unknown to the public at the time of voting — a necessary precaution so that the AI can measure the team’s chance of implementation success.”

  • why? You need to know if your team is CIVICSMOD or a bunch of hippies off the street. At no point does hiding the data about the team allow anyone to analyze the team’s chance of implementation success.

Those vote weights can change depending on particular achievements or flops, but they tend to remain stable.

  • This would mean an incredibly deterministic society where people do not improve their lot or learn anything new over time. Especially given the later phrase “The solution, historically, has been to focus on motivating the working- and middle classes by providing opportunities for social mobility.” These two ideas are inconsistent with each other.

Apart from the main “democracy of ideas” system, there is a sector of public governance that is perpetually employed, called the maintenance crew. This is widely regarded as a low position in Hygge society, even in its crucial importance for the overall functioning of the Distrikt.

  • why? What do they do? Why are they a low position if crucial?

At the same time, politicians leveraged the popularity of the generous welfare system to keep very wealthy individuals in minority, and prevent them from unduly influencing policy.

  • we have politicians now? Text elsewhere indicates we have working groups of people proposing and executing projects; this seems more in the realm of Sata as a default function of minimizing disparity.

"Wealthy Hyggians mostly responded by relocating elsewhere on Witness, typically on Libria. " - why? Their wealth would not exist outside Hygge if their money is controlled by Sata.Have we also not established that those who remained in Hygge were the most rabid Denton fans?

“Such punishments are conducted by Sata itself, and they’re noteworthy because they can’t be changed or cancelled after their initial sentencing, even if the subject is proven innocent later on.”

  • why would anyone implement something this draconian?

Have not taken out, but have a problem with: What is AI weaving? What is mechanomics?

1 Like

I think I managed a reconciliation. Any further tuning you can think of?

Hey @yudhanjaya, sorry I didn’t have a free second yesterday, so I just didn’t log in between your provocations and your intervention! Wish I could have saved you some time!

I had a look at everything and it feels right to me! I just went through the file and added a few comments here and there, and also changed the structure of paragraphs — I think a structure with more “#” titles rather than “###” titles is less confusing. I also standardized the quotes: they should only come after “#” titles.

To me there only one single detail here that’s stopping this from being ready: all the “XX” I added are dates… and I don’t know how to date them! Could you help me with that, @yudhanjaya?

Sure thing, man. Good job on making it more legible. Will sort out the dates as soon as I’m done on Avantgrid.

I was going for a faux-market economy. It looks markety, it has markety characteristics, but the AI has enough control over the economy (mostly via universal income) that it can prevent and control many of the nastier aspects of the market economy. Again, all in Hygge is about reducing pain, rather than increasing glory.

What do you think? If you wanna add something here, just go for it!

Good call! Love it!

YES! That’s the whole complaint of the entirety of Hygge.

It’s a system that is focused on preventing shit from happening, including good shit. It’s incredibly deterministic and monotonous, as every overbureaucratic society is (I live in Germany man), and though people are happy that the previous shenanigans are not happening (again, I live in Germany man), they are just not satifyied.

Sata just does it. It requires no approval for economic acts. It’s just AI — it doesn’t need to justify, just apply what its systems deem correct.

Perhaps Goro gave Sata some internal lesson on fear. And its usefulness.

A lot of your 'why’s and even eliminations had the same theme, Yudha, which was me creating groups that could benefit or be dissatisfied with systems. I thought this could lead to interesting stories, like somebody from the Maintenance Crew that dreams of becoming an elected politician or somebody who was unjustly accused of theft be summarily kicked out of Hygge and be forced to adapt to another system. Anyway, it’s cool! Perhaps it’s for the best that the worldbuilders create their own groups and dramas :slight_smile:

@yudhanjaya @Joriam this is still economically inconsistent. We can’t have both a social democratic economy and “near total allocation of resources by an AI”.

I made a social democratic economy because the universal basic income thing stood out. UIB is a redistribution measure, but where is the production sphere? So, I tried to imagine a system that (1) would produce a redistributable surplus and (2) would be self-enforcing. Earth’s real social democracy is only partially self-enforcing, because the rich try to undermine the redistribution system.

Redistribution can be achieved by AI, as Joriam suggests and as in Red Plenty. But then, same thing: what prevents people who stand to lose from the AI’s intervention from tampering with it? This is what happens in Red Plenty, after all. Factory managers in practice default back to Gosplan practices, paying only lip service to shadow pricing. On this, I have no idea.

@Joriam So I have a bit of a problem with plot being as some sort of silver bullet.

It’s just AI — it doesn’t need to justify, just apply what its systems deem correct.

No human community is going to fully give over the reigns to decisions it cannot understand or justify, unless there is a combination of propaganda and tight economic and political control. To channel Gramsci, you need a cultural hegemony in place. Think the U.S.S.R. and, going back further, the Catholic Church of the Crusade Era. And this is logically inconsistent with the idea of a state where people have so much political determination that they are able to vote, Kickstarter-like, on projects instead of politicians.

If a successful totalitarian government has been set in place, you will need all of those (context: I come from Sri Lanka: wartime governments are terrible and terribly effective things, but giving political power to people is the least of what they do). Even then, you have dissent: from terrorist spinning civil wars. You also need to control borders, or at least prevent awareness of other systems (think North Korea) and Hygge is open via the Migrant Train.

Assuming all this is done: why three AIs, then, or a vote system? A ruling class, to channel Gramsci again, would only have to give enough goods and services to the ruled that there is no outright rebellion. And Hygge is far from that.

So there is no logical consistency here between the initial setup you have for Hygge and the end shape of the society. While it may be the society we would like it to be, that does not mean it is the society that would evolve given the social forces that would come into play. Systems interact with one another: they do not exist individually, even in stories (and stories where they don’t interact are easily detectable by their whiff of unreality).

Secondly, as someone who trains machine learning systems: you are never going to have all-perfect AI. It is not a silver bullet for good or evil. Even the State Machine is not without flaws, and historically we are bad at dealing with outliers. AI trains for the middle of the bell curve on a distribution of data: they are thus incredibly efficient at dealing with most cases, but never all. German dissatisfaction is a good example: Germany has systems and a society (no bombs, no race riots, actual roads, a decent passport) that most of us in the Global South would (and have) kill for, but they are unsatisfied.

I would suggest we do @alberto’s Red Plenty economy; something centrally run, with people’s votes on projects turned into redistribution plans by AI and then executed by humans. That is what I see as letting us preserve most of Hygge - including Goro to provide entertainment and a deliberate aspect of randomness to life.

1 Like

I deeply disagree with this.

I think all human communities will give over the reigns after the simple shift when the AIs start doing the job way, waaaaay better than humans.

It’s like when European monarchs started becoming ‘enlightened despots’ around the time of the Dual Revolution, because the combination of industrialization and Democracy was so undeniably powerful that even the people who would theoretically be beheaded by this new system started adopting it.

It makes no sense for a monarch to adopt industrialization+democracy, it undermines all of their powers, both the land power and the birth right. It’s not dangerous, it’s suicidal. Still, it sold like hot cakes. Why? Because industrialization+democracy was steamrolling over feudalism+monarchy.

As soon as AI-centric governments start steamrolling over human-centered governments, everybody will be forced to accept the new way to do things.

At the same time, part of the cleverness of those new systems (and that part doesn’t exist yet) is to understand humans very deeply. Any AI system will have to give the impression of control to humans — they don’t need to control the world, they just need to believe they do. Thus the kickstarter-like characteristics of the system. Humans control the sexiest actions, while the AI control the reduced, efficient and necessary maintenance crew (that ended up deleted).

The three AIs, instead of just one, is another form to manipulate human expectations and irrational behaviors. People will take sides, people will complain about one and praise the other, people will make Tshirts. This is a key feature to control political turmoil in times of peace.

In the final Hygge I created, the term ‘ruling class’ finally applies in reality. There’s no ‘ruler class’, the ‘ruling class’ is temporary, fluid and obtained via means that are not capital.

I think all the logical inconsistencies that you pointed out come from the same root: you don’t think humans would subjugate to an AI. But I would, and it doesn’t even need to be perfect, all it needs to do is a better job than any human group ever did.

What do you think?

I didn’t read Red Plenty, but buy what I researched here it looks like a cool reference — but this book seems to point at a rather flashy, rather opulent view of a possible economics system, and I wanted Hygge to have the marasmus of the soviet block after the USSR’s fall from grace. Or maybe even the current economic stagnation of Japan. Japan is megacapitalistic, but there’s just no extra money flowing for 15 years - a kid can only aspire to be as rich as their parents, dreaming more than that is too much. I wanted a productive society, including some of the most crucial products in the world (like Japanese manufacturing and technology), in which people wanted to be rich — but they just couldn’t reach the richness.

But above all, I wanted it to be realistic, so I’m not really arguing here, just presenting where I came from with all those ideas.

Do you think the Red Plenty’s economy could be interpreted in a less flashy way, @alberto?

Oh, no. Homo economicus- the rational Man - does not exist. If it did, we would have no war, no religion, nothing strictly logical, and no stories: the eagles would airlift the ring into Mordo.

At the same time, part of the cleverness of those new systems (and that part doesn’t exist yet) is to understand humans very deeply. Any AI system will have to give the impression of control to humans — they don’t need to control the world, they just need to believe they do. Thus the kickstarter-like characteristics of the system. Humans control the sexiest actions, while the AI control the reduced, efficient and necessary maintenance crew (that ended up deleted).

Now this I agree with. This is the basis of the State Machine, and indeed Ian M. Bank’s Culture. (By the way, CIVICSMOD is the crew that does that maintenance function): they are far, far better than the human equivalent.

But to assume that we will end up here without turmoil is merely to set up a Potemkin village. At major junctions you must ask: what about people who disagree? What will they do? Given that you have written people directing their ire at the State Machine, do you believe it feasible that the triple-AI future will arrive without a fight or even skepticism? The European monarchs certainly did not go without a fight; they were eventually forced to co-opt. See Cromwell, the French Revolution, or fundamental keystones in our history. Even then there were Luddites; and today we have tech-Luddites dotting the development sector as decisionmakers.

But it looks like we are in agreement as to the source of ideas. And here is an excellent place to trial ideas from Red Plenty and Ian Bank’s The Culture: a centrally planned economy planned and run by AI.

At the risk of plugging myself: might I recommend reading: https://slate.com/technology/2020/09/state-machine-yudhanjaya-wijeratne.html and taking a crack at this? If needed, @joriam, more than happy to have a call and discuss after.

Disagree with your disagreement. Better for whom? “Human communities” are not monolithic agents. Most allocation decisions result in winners and losers. The losers tend to fight back. The winners tend to try and stay on top. I imagine that most Wall Street bros would dislike what you, Joriam, would call “a better allocation”.

Well, imagine that the shadow pricing reformers in Red Plenty would have had the full support of Kruscev, including the KGB on their side. Then they could have overridden the hesitations of the factory managers.

That’s not flashy at all.

  • The system is imposed by fiat. It gets imposed because someone who gathered power before the AI worked acceptably well decides it would give them an advantage. In the Soviet Union, this advantage was the race with the west: a small, but steady delta in productivity would crush the West by the sheer weight of exponential growth.

  • Eventually, the system gains enough consensus with enough people that the politics check out: he AI can stay, becomes a piece of the governance, and burning it to the ground exits the Overton window.

This is “it does things waaaay better”, but now with the politics added, in turn supported by the economics. If the AI decides your house has to be razed down to make space for a Migrant Train track, you will protest (good for the Distrikt, bad for you personally), but if you protest too hard a white van will show up and make you stop. The goons are paid by the taxes of a majority who is better off by the current system.

I’m even feeling bummed we’re having to make final decisions because of a release date, I really love this discussion for the sake of it!

I’m not dying on this hill, but here’s my final (and also abandoned) argument for pro-AI systems in the future: the human brain might not be able to calculate all the factors that would lead the losing side of a society revolting against a system or regime — but an AI could. We don’t have anything like this right now, but one day an AI could have such an elegant risk calculation system that it would know exactly how much and who it could displease and still remain operating and being efficient.

I do like this final twist @alberto gave with a… rarely used, but quite noticeable Peace Police that is nothing but peaceful.

We do not need to make final decisions. This is a wiki, it is meant never to be final. :slight_smile:

Well, much as I love debate, can we get a move on on the writing?

The best way to trial this out is to encapsulate this debate itself as political factions. Suppose the triple AI are trying to create this optimal mix of minor markets + distribution of major resources by fiat. Say there is a prominent political faction that argues this is inconsistent in the long run and that AI-optimized fiat should be the only way forward. Say there is another faction that is pro-markets with decidedly Keynesian welfare state leanings.

Now Hygge is not one politically homogenous mass, but a place that has more than one opinion in play, and it matches @Joriam’s statement above:

People will take sides, people will complain about one and praise the other, people will make Tshirts. This is a key feature to control political turmoil in times of peace.

We have some bounds to ideas. For example, the differing opinions here show that assumptions like:

I think all human communities will give over the reigns

Are not going to fly, even at very small scales, without some ugly historical context which draws from how revolutions and power have worked in the past to set up these conditions. Neither utopias nor dystopias appear out of the ether.

The European monarchs certainly did not go without a fight; they were eventually forced to co-opt. See Cromwell, the French Revolution, or fundamental keystones in our history. Even then there were Luddites; and today we have tech-Luddites dotting the development sector as decisionmakers.

We also have some historical context and other systems set up that do not need to be canonically replicated or deprecated - please read the History. For example:

CIVICSMOD is the crew that does that maintenance function

one day an AI could have such an elegant risk calculation system that it would know exactly how much and who it could displease and still remain operating and being efficient.

Yes, welcome to the State Machine. I feel like we’re going around in circles now. This merely increases my confusion as to why we need more arbitrary Deus Ex Machinas when we canonically have one setup up and primed to fire.

Write and see how the economy of Hygge develops. See if there are rational / and/ or historical arguments for design decisions. Put yourself in the shoes of decisionmakers and see what tendencies they might have and to where they might be most willing to push decisions. We need this setup to be economically consistent with itself, whatever form it takes: otherwise we might as well switch genres altogether and build economies based around magic rings.

Hey, this process doesn’t need to be halted!

After your additions we already have a BUI system with a market on top of it. We have products and disciplines that can trade with other parts of Witness. As much as I’d like to see this Red Plenty economics (again, I haven’t read it) more fleshed out, I don’t think this is a requirement to push this forward to the wiki. It’s “safe enough to post” as it is, we can improve it later.

That sounds real good. I’ll add one hint at this in Sata’s text and later on we can give the specifics.

I also added a small section about the free market below Economy.

Proposal: let’s delete Hygge — and create another ‘first distrikt’ totally from scratch.

Context here.

1 Like