Comment by gaiagraphia | Hacker News Reader

Comment by gaiagraphia | Hacker News Reader

"That's another thing we've learned from your Nation," said Mein Herr, "map-making. But we've carried it much further than you. What do you consider the largest map that would be really useful?"

"About six inches to the mile."

""Only six inches!"exclaimed Mein Herr. "We very soon got to six yards to the mile. Then we tried a hundred yards to the mile. And then came the grandest idea of all! We actually made a map of the country, on the scale of a mile to the mile!"

"Have you used it much?" I enquired.

"It has never been spread out, yet," said Mein Herr: "the farmers objected: they said it would cover the whole country, and shut out the sunlight! So we now use the country itself, as its own map, and I assure you it does nearly as well."

(From Sylvie and Bruno Concluded by Lewis Carroll, first published in 1893.)

I wonder whether the above is applicable to modeling the world and its behaviour in other ways; fully understanding everything of the complexity of the world would require the power of the world itself.

And, inspired by that: "On Exactitude in Science", Jorge Luis Borges

…In that Empire, the Art of Cartography attained such Perfection that the map of a single Province occupied the entirety of a City, and the map of the Empire, the entirety of a Province. In time, those Unconscionable Maps no longer satisfied, and the Cartographers Guilds struck a Map of the Empire whose size was that of the Empire, and which coincided point for point with it. The following Generations, who were not so fond of the Study of Cartography as their Forebears had been, saw that that vast Map was Useless, and not without some Pitilessness was it, that they delivered it up to the Inclemencies of Sun and Winters. In the Deserts of the West, still today, there are Tattered Ruins of that Map, inhabited by Animals and Beggars; in all the Land there is no other Relic of the Disciplines of Geography.

— Suarez Miranda, Viajes de varones prudentes, Libro IV,Cap. XLV, Lerida, 1658

All of you would love the "obscure cities" series of comics (https://en.wikipedia.org/wiki/Les_Cit%C3%A9s_obscures), as well as some works by J. Abeille (https://en.wikipedia.org/wiki/Jacques_Abeille#Les_Jardins_st...)

Regarding your last sentence, "fully understanding everything of the complexity of the world would require the power of the world itself." maybe that's not true? In particular, maybe reality is highly redundant and can be modelled using highly compressed representations.

You can take the Hari Seldon approach and abstract away the macro level 'crowds/nations' and simulate them when they're not in the observer's sphere of influence/vision using a (much) higher level of abstraction.

not really though. In the real world you can see the results of action over time. So, if you stop simulating what the user is observing, when they actually do observer it later, you're required to run the entire simulation to get that observed part of the world back to the state it should be in after time passed. That includes the influence of the butterfly on the other side of the world. It effectively adds up to computing the entire thing.

To take a semi-long term example. If you go carve your name in a tree, when you come back 50 years later it better show your name distorted over those 50 years. If you cut open the tree. It better show a record of every weather event over those 50 years. Every piece of the world you inspect will need the simulation run and not just for it but for all the things that affected it and all the things that affected the things that affected it etc etc recursively until you've simulated every in the universe.

You're assuming a simulation where you either care about accuracy for all elements, and/or where the observer can tell if you "cheat".

Minecraft is an example of the latter: if you leave a chunk, mobs despawn and decay and growth stops.

But a simulation need not work that way. E.g. instead of continuing to run a simulation, it can simulate macro-indocators, and continue to simulate micro-details with a strong bias towards reverting to a prediction from the macro-indicators, and each element that gets there can be discarded.

To go back to the Minecraft example, you set a "decay rate" for mobs that reflects a believable rate of dispersal, or simulate changes to vegetation towards what your generating function will produce until they match.

Then when the player re-enters at t+a_lot, you just run your generating function where time is one parameter, and apply whatever reduced delta is still relevant.

Yes, you'd still need to simulate a lot, but you'd be able to prune most things.

So the question is meaningless without defining your requirements.

Judging from first few chapters Seldon did manage to create probabilistic model, and not a deterministic simulation (he was not sure how the interview with government representatives will go in the end). So I guess even syfy writers do not believe that it is possible to simulate reality in deterministic way.

Since reality is not deterministic that possibility simply does not exist.

Reality is not meaningfully redundant at the data level but the computation is nonetheless compressible because the dynamic feature space is vastly smaller than the data.

The macro characteristics of entity dynamics in the world is that of a sparse metastable graph. There is considerable compression possible at the computational level, since most entities in the world are fungible in every discernible aspect, but the data model is not similarly reducible because what can be compressed away is context dependent. Since the computation is less tractable part, this makes the modeling problem eminently viable albeit still extraordinarily difficult. To model the effects of the relatively small number of nodes and entities that can shift the equilibrium you need to compute over the true graph, not some aggregate and summarized roll-up of the graph. There is a real tension between anonymization/privacy (and reducing computational costs) and the ability to build models of human systems that remotely reflect real-world dynamics. Models based on anonymized aggregations are notoriously rubbish when ground-truthed.

It is also important to treat these models as sparse metastable graphs. I am always slightly horrified by the number of serious academic disciplines that often default to modeling human systems as if they were a computational fluid dynamics (CFD) problem. Global human systems don’t have CFD-like structure but CFD is attractive because it is easy to reason about computationally and widely understood. A lot of scientific research on models of human dynamics (e.g. epidemiology) that were ultimately debunked had a root error in the use of CFD-like assumptions about how humans interact and behave.

It is technically possible to build a far higher fidelity model of physical world dynamics than I think most people would expect or intuit. Unfortunately, that is not the hardest part of making it useful.

> the farmers objected: they said it would cover the whole country, and shut out the sunlight!

This is actually a solved problem now with Augmented Reality, whereby you can walk in some field and get an overlay with the additional layers you care about at that moment.

This is the problem. How good is a model compared to reality.

I heard one argument against Climate Change being that it is 'just an estimate', that reality is more fine grained. How can you trust a model since the model is just an estimate of reality that leaves out a lot of details? They were calling into question all models, in order to debunk Climate Change. Not sure they realized how much of the technology is based on models.

This. Think in terms of the compute power and raw energy needed to accurately model climate. The best at the moment use cells that are about 3 miles on each side. Hopelessly crude.

https://www.climatehubs.usda.gov/hubs/northwest/topic/basics...

Is it hopelessly crude? I mean the model is definitely wrong by some high level behavior might be useful enough

How do you know it is wrong? To be definitely wrong, wouldn't you need another more accurate model to compare to it, to disprove it?

Run it on past data to see how it predicts against known recent climate and watch for curve fitting.

I guess you are saying that those failed? I doubt that. Since those types of runs are done all the time to improve accuracy.

Or more to point, it isn't like there was one-run, we take that as gospel, then run on past data, and failed, so models are wrong forever. They are always being refined.

And some of the longest range calculations, that were done decades ago, are matching. So what model are you specifically saying failed?

""The hallmark of good science, however, is the ability to make testable predictions, and climate models have been making predictions since the 1970s. How reliable have they been?

Now a new evaluation of global climate models used to project Earth’s future global average surface temperatures over the past half-century answers that question: most of the models have been quite accurate.""

https://climate.nasa.gov/news/2943/study-confirms-climate-mo...

How many does it take to simulate an atom? What about multiple atoms? I believe with current computers the answer does not scale linearly. But like they did here you can always use one atom to simulate that same atom.

Would the "map" itself be shown at full scale in the "map"? This sounds like it's turtles all the way down - ad infinitum.

It reminds me of virtualization. An emulated CPU in a hypervisor will never perform as well as the physical CPU that is hosting it.

virtualization doesn't emulate a CPU; instructions typically run natively. IO is trapped and handled exceptionally.

Hardware virtualization can still emulate peripherals. If your virtual machine thinks it has an ethernet card that doesn't actually exist, it may be nice and fast for CPU operations, but networking will take a hit.

They're talking about an "emulated CPU"...

That's not true. Some machines are faster for some workloads when virtualizing. This has been known for decades. I recall first reading about it around 20 years ago, but I can't find the reference I'm thinking of. Nevertheless you'll find plenty of hits for faster in virtualization if you search.

A cycle exact and semantically precise one, sure. One that is allowed to prune unobserved results, on the other hand could potentially just lazily evaluate or simulate approximate outcomes and skip whole chunks of the simulation. To make claims about the computational power needed to simulate something, we really need to first know at what level of precision, and whether there are e.g. any observable effects of taking shortcuts or not (e.g. if "outside time" are not observable "inside", then skipping anything that can be lazily evaluated only if actually referenced from something observable even if it makes computation slow might be fine)

Reply on news.ycombinator.com