Complexity and decline

[Voidwalkin]

I wonder: will there come a point where systems in place become complex beyond our comprehension?

Are we already there?

Take economics. I'm not an economic genius. But you can tell when people talk economics that their grasp is variable. Some evaluate the board like a chess grandmaster, some are hopelessly flawed.

Related: also concerning to me is that the sum of knowledge required to speak intelligently about some complex topics necessitates such dedication that expertise is a prereq. Expertise can lead to a trap.

Example. If the question is asked, what are the dangers of AI, you'd need to bring in an AI expert to understand what AI is/will likely be capable of. Then a different expert to understand how companies are likely to adapt the technology, given cost barriers, expected RoI. Then, if necessary, a politician to even have a sense of which(if any) legislative policies can gather enough support to pass.

Terrifyingly, all these experts must be right in their assessments and acting in good faith, or society will likely be paralyzed to inaction, act w/o effectiveness, or act counterproductively. You might be better off turning to generalists if errors become more likely as experts interact with each other.

I can totally see a significant stressor coming along a weak link in the chain breaking, setting in motion disastrous turns of events.

 

[Valka D'Ur]

Complexity: Sometimes the current world reminds me of the mess a kitten makes when they get into a yarn basket and proceed to unravel the balls of yarn while twisting them around all the arms and legs of anything that has arms and legs (furniture-wise), with the addition of flash points where if you don't untangle everything perfectly, barehanded, without cheating and using scissors to break it down into manageable sections, your entire house could explode. The kitten doesn't care that they've made a mess that you have to fix before you do anything else, and it's a hard puzzle. They want you to pay attention to them, give them what they want, and don't care that this mess is now your first priority and you don't even have a clue where to start. This is an analogy that can apply to so many things in this world.

As for AI: The more it infests everything, the more I think of "Thou shalt not make a machine in the likeness of a human mind." Whoever eventually gets the bright idea (said sarcastically) to install self-programming AI medical computers had better have read The Dune Encyclopedia and be prepared for the consequences. What that AI doctor did next is what touched off the Butlerian Jihad.

 

[Moriarte]

You might be better off turning to generalists if errors become more likely as experts interact with each other.

I'd say as complexity rises we will need better multidisciplinary experts. Generalists, who know a little about everything, don't understand the problem deep enough to be able to solve it. Teachers, journalists - their task is to make the public aware, to ignite the spark. Then experts take over and make sure the problem is solved. As complexity rises, some doctors will need to be trained in machine learning. Other interesting interdisciplinary hybrids will emerge. Full on generalism? I don't think so.

 

[Birdjaguar]

In addition as complexity arises, teams might well become more important to formulate questions and answers. The individual genius approach we have often seen in the past might be a difficult path.

 

[Samson]

Take economics. I'm not an economic genius. But you can tell when people talk economics that their grasp is variable. Some evaluate the board like a chess grandmaster, some are hopelessly flawed.

I am not convinced. If anyone really understood economics we might get a conversation that sounds more like science than alchemy.

Example. If the question is asked, what are the dangers of AI, you'd need to bring in an AI expert to understand what AI is/will likely be capable of. Then a different expert to understand how companies are likely to adapt the technology, given cost barriers, expected RoI. Then, if necessary, a politician to even have a sense of which(if any) legislative policies can gather enough support to pass.

That is because predicting the future is hard. We used to use chicken entrails, now we use highly paid consultants, but with questions like "how will this tech fit into the real world" neither is obviously better.

The world is complicated, sure. The time when one person could learn all there is to learn past some time in the 1700's I guess. It seems to me the real problem is less complexity but the fact that our current incentive systems are so bad at encouraging the behaviour we really need. We need to find different incentive mechanisms, that will involve challenging established power, and we needed them yesterday. That is the real problem, and I do not see a realistic solution. I hope someone does.

 

[Narz]

I wonder: will there come a point where systems in place become complex beyond our comprehension?

Are we already there?

I think so. Definitely too many moving parts to make accurate predictions.

All we can do is focus on prime movers, things that are obviously true and effective and hope for the best.

The problem is the 'we' tho. At the time in history when a united 'we' is actually necessary 'we' are quite divided

 

[dusters]

No, hard disagree.

The systems are not getting more complex, people's mainline education is lower. Back in 70s my mom who was barely gifted spent lots of hours to study physics/chemistry in USSR to read pop-science journals. In USSR these journals, aimed at undergraduate/graduate level STEM students were very popular. They contained general ideas, formulas, expected breakthroughs and current trends.

After collapse of USSR these journals disappeared. Right now we have a journal "Illustrated science" instead of "Modern science". It has a lot of pictures instead of formulas. It has dumbed down considerably in past 40-50 years. These journals back then were state funded, but now they are private funded.

My guess is that to read PhD level papers you need at least BA in the field yourself. To read BA papers you need to familiarize yourself with introductory courses using correct books available from a general library.

The problem is that people don't read these scientific books for fun anymore. When I started studying in 2009 the best thing was that I was given a library card to huge amount of scientific books.
I picked up introductory books on Philosophy, linguistics, history. When I switched universities I picked up more - on physics, chemistry, psychology.

If it takes 2-3 months to familiarize with professional terminology in a field and therefore be able to read BA level papers, more people should do it. Because otherwise people just study their own field and are almost clueless. Many linguists don't know how gravity works, historians don't know why metallurgy is complex and vital to wars, because it requires chemistry basics. And so on.

Interdisciplinary courses should be the norm, at least 20% of courses nowadays at MA level should be interdisciplinary, but in smaller universities I don't see it yet.

My take is USSR/Latvia specific, so I am not saying it is a general trend in the West.

 

[Samson]

The systems are not getting more complex

Hard disagree. I agree with a lot of what you say in your post, but I think the systems have become inherently more complex in recent decades.

Whatever level you look at the world is becoming more complex. The internet gave us access to all the information in the world, if only we had time to consume it and determine what was true. Computers allowed us to make decisions that would have taken humans time to be made for "free". Mobile computing allowed us to always have this decision making support. Globalisation allowed us to source every part of our product from a different corner of the world for less than the CEO spent on lunch. Just in time supply chains reduced capital requirements such that you had to do it to compete on the FTSE.

 

[dusters]

Hard disagree. I agree with a lot of what you say in your post, but I think the systems have become inherently more complex in recent decades.

Whatever level you look at the world is becoming more complex. The internet gave us access to all the information in the world, if only we had time to consume it and determine what was true. Computers allowed us to make decisions that would have taken humans time to be made for "free". Mobile computing allowed us to always have this decision making support. Globalisation allowed us to source every part of our product from a different corner of the world for less than the CEO spent on lunch. Just in time supply chains reduced capital requirements such that you had to do it to compete on the FTSE.

What I was trying to say - as systems were getting more complex, we had to increase our general education level along with them. That didn't happen. Therefore complexity is starting to become a problem.

And funding, funding. In Latvia we have little to no money for pure science. Pure science doesn't give immediate results, it takes years. So we are making breakthroughs mainly in fields which are marketable. Physics, chemistry, biology. The soft sciences aren't funded as much. So we rely on new knowledge from Germany/France/UK mainly in fields like semantics, ethics, pedagogy.


***
Personal Edit - I want to clarify that in early 90s I got 80386 PC. In one of offices my mom worked they had 80286. These computers had 2 MB of RAM and 8 MB of RAM.

By 1997 computers with 32 MB RAM and more become the norm. Soon Pentium 3 and Windows 98 happened. Futurologists spoke of geometric progression in computing power.
At age of 11 I was programming basic stuff in BASIC and attending a school for future programmers. I realized what miniaturization was by playing Civ 3. There is a tech with that name.

So things got smaller and smaller quickly and more complex meanwhile. To keep up, one had to study more. Electrical engineering was the main one, I thought at age of 16, but robotics should be important too.

So in early 2000s as a kid who could repair his own PC at age of 13 and I repaired PCs of adults as well, I realized that IT and robotics is the future.

A friend of mine gifted me Bill Gates book "Road Ahead" on my 13th birthday. It was year 2000. Bill wrote about/predicted broadband internet and lots of things correctly. I was impressed and decided to teach kids how to adapt to the new age of information when I graduated high school.

 

[Zardnaar]

I think the technocrats elite type things are more complex. Day to day life has gone the other way.

 

[Voidwalkin]

In addition as complexity arises, teams might well become more important to formulate questions and answers. The individual genius approach we have often seen in the past might be a difficult path.

There isn't any structure to assemble the team on an effective timetable. Creates vulnerability to inertia and opportunity for self-interested parties to sabotage.

Paper is published. Those with most expertise begin to interpret applications first, begin predicting societal impacts(probably quite poorly, this being outside their expertise)

Takes time to establish consensus. Months, if not years. During which time any adversely effected party can sponsor a counter-study.

By time the politicians get it the water is already muddy. Those who'd prefer to dispute conclusions in bad faith/ideological reasons have their pretext and alarmingly, the average voter is probably totally detached from the basic knowledge needed to soundly evaluate, and easily mislead.

If we are already past the point, then the current structure of consensus-building is ghastly inefficient for the job.

 

[Voidwalkin]

That is because predicting the future is hard. We used to use chicken entrails, now we use highly paid consultants, but with questions like "how will this tech fit into the real world" neither is obviously better.

Huge problem I suspect stems from experts being no better or worse than average at predicting human behavior, making most of their conclusions shaky. As we should probably expect(see below)

I'd say as complexity rises we will need better multidisciplinary experts

You won't get that by nature of how skill develops. 10000 hours are necessary for even the most talented to reach just the top 80% percentile in any field. Isn't going to be enough time left over for them to be expert in much anything else.

 

[Gori the Grey]

I think we're already there. And feeling the effects of it.

For me, it takes the form of the information overload. So, for your example, we want to find an expert in AI. But who? If we go poking around, the internet will give us various candidates. How do we evaluate their relative expertise? There are some ways, of course. Rate of publication in refereed journals. But then there will be some other expert in the field, nearly as well published, who thinks, and has said, that the one you're honing in on is a crackpot. How do we evaluate that expert? Infinite regress.

And then we need a business expert. But which one? How do we establish their credentials? Rinse repeat.

Anyway, we settle it however we do. Now. How do these people talk to each other? How does the AI expert put his knowledge in any kind of meaningful connection to the business expert? They'll each freely confess that they aren't experts in one another's field. Where do they take their first starting point for the conversation? Does the business expert, who knows the kind of product that businesses tend to invest in, start asking the AI expert whether AI fits that definition? Does the AI expert explain what AI is capable of and the business expert start speculating on which businesses might be interested in such capabilities? There's no given way to intersect these two fields of knowledge. And then who is to judge whether the two of them come up with any valid "take" on the matter?

So why do I say we're already feeling the effects? Since people despair of getting anywhere in this morass of data, they're more disposed to just putting their trust in one person who can make himself sound like he knows everything. They long for the simplifying effect of just having to listen to one person. People who have no hesitation about just making up facts out of thin air, but then voicing them as though they are gospel truths, will be able to get whole hordes of people to follow them. And experts are powerless to refute this person's claims because "well, that's just your opinion, man" and "yeah, but there's this other website that says this" and "the experts are always bickering among themselves; how do we know that this over here isn't the truth; it's got as much validity as any other claim." and "look at how many hesitations and qualifications that expert shows when she speaks; this guy over here says his truth forcefully."

In the course of my lifetime, we went from information being a scarce commodity to being in massively superabundant supply. The implications of this shift are numerous, far-reaching and profound and mankind is scrambling to cope.

 

[Voidwalkin]

So why do I say we're already feeling the effects? Since people despair of getting anywhere in this morass of data, they're more disposed to just putting their trust in one person who can make himself sound like he knows everything.

Whole heartedly agree on larger point of mass failure to evaluate at every turn.

But more commonly I think the average person puts their trust in nothing at all, or their blame on something, rather than trust in an individual or movement. Cynicism becomes common as people lose faith in the ability of structures to respond positvely to change. I think the dramatic rise in conspiracy theories is totally reflective of that. Cynicism is so common now I'd even say it's wounding us.

The presumption I see more and more people making is that nothing ever changes and they're both right and wrong there. Our ability to express agency is dramatically limited on anything at scale above the individual level because of impossibility in evaluation.

But they're entirely wrong that nothing ever changes. It does, more rapidly than ever before, even permanently altering the most basic things like how we speak to each other. Which is to say, there's less speaking than ever before, and far more writing. Before anybody can judge whether or not these are actually GOOD, it's already far too late; the super tanker is headed down a strait too narrow to turn around.

 

[Birdjaguar]

We live in a muddled, fast changing world of both determined and otherwise occuppied people. We have no plan or even a guide to the future beyond our hopes. So, we wrangle our way through and course correct as we go along; at the same time we add increasingly to what we know. Pessimisrts will cry out in despair that the end is near (in some fashion or other); optimists will see a bright future of solving problems big and small; most of us will just keep on keeping on and try to put our house in order one day or week at a time. The "super tanker in a strait too narrow to turn around" is just pessimistic realism at work.

Unity of purpose has never been a strong suit of humanity.

 

[amadeus]

I think how good of a grasp we have on things depends on whether we are talking on the individual level or the collective level. Individually, some things are beyond our understanding: pop open the back of an old TV set and you’ve got an electron gun, a tube (by the way, don’t touch that without checking to see if it’s not discharged first) and a few wires. Easy enough for the layman to fix. Nowadays, no clue. If it’s broke, throw it out?

But on the collective, societal level, I don’t think our pace at developing new systems has grown exponentially relative to our ability to understand them. I’m not even sure we can develop things that are totally beyond our ability to understand them, or else how did they get built in the first place? With specific regards to AI, I think a lot of it is smoke and mirrors. I remember the dot-com bubble, and a lot of what was said 25 years ago sounds to me very similar to what’s being said today.

 

[Voidwalkin]

But on the collective, societal level, I don’t think our pace at developing new systems has grown exponentially relative to our ability to understand them. I’m not even sure we can develop things that are totally beyond our ability to understand them, or else how did they get built in the first place?

It's not that we build things we don't understand. It's that we don't understand how those things will interact with the systems in place.

Take the internet. Its designers understood it. They probably didn't expect algorithms to be written to expose people to deliberate agitation for engagement. But it did and there's no denying this has been a contributor to polarization on so many issues.

With specific regards to AI, I think a lot of it is smoke and mirrors. I remember the dot-com bubble, and a lot of what was said 25 years ago sounds to me very similar to what’s being said today.

Maybe. I think it will change something. Maybe not abrupt dramatic changes but we will see something.

Guys at work sceptical of automation tell me all the time they'll never be replaced by a robot, but wages haven't kept up with productivity growth since the 70s. Doesn't have to be an android in the line for change to hurt.

 

[Quintillus]

There are multiple ways to look at it.

On the one hand, as our collective understanding has grown, there are fields that the average person can understand today that they couldn't have 100 or 200 years ago, because the knowledge didn't exist. Germ theory, for example. Biology and medicine in general. Is medicine more complex than in 1900? Doubtlessly. Is that a bad thing? No, and that's even considering that it's much more impossible for a doctor to know everything that is known about medicine now than it was in 1900. This is why there are GPs and specialists, and you might go to a GP first and be referred to a specialist, who might refer you to a sub-specialist. Maybe even then they can't figure out what's going on, but maybe then can, and there's a good chance that you wind up with a much better outcome than in 1900 when our understanding of medicine was much simpler.

So in that case, the increased complexity has been a good thing, even as it's harder for any one person to understand the field.

Another case is computer processor design - this has been machine-assisted for decades, I think the last fully human-designed CPU was in the early 80s? Late 70s? Humans provide design ideas, but computer programs do the final layout; it's simply too many transistors for humans to come up with the optimal design. Literally billions of them, probably tens of billions now, so we delegate that to the computers. It may be that no one at Intel understands exactly how the transistors are laid out on their chips, but the designers do know the high-level and mid-level aspects of the design, and that is good enough. Complexity has been managed to allow progress to continue.

Where I think it's tougher to say it's a good thing is the decreased ability for the average person to understand how the tools in their everyday life work, even if they want to know that. We've always had things that defied explanation, and were explained by magic or pantheons. How does fire work? It's a gift from the gods. How is the image displayed on your television? It's a gift from the engineers at Panasonic. How come the river never runs out of water? Neptune ensures it doesn't. How does your phone load CivFanatics, and how does CivFanatics work? I could give a halfway decent overview of that, but for a lot of people it's just something that works.

So I guess the conclusion this lead me to is, yeah, there are a lot of everyday gadgets that we have a surface-level understanding of. But there are people who know the details along the way. And we've always lived with things that worked without our understanding how, which we accept as part of life. We might worry that some day the Internet might break and be un-repairable, but the ancients worried that the gods might turn against them and cause a draught that could lead to starvation. Which is the more valid fear? True, the Internet does depend on human activities, but I think the ancients had the more plausible fear, even if the reason for the draught wasn't necessarily the gods' anger at them.

 

[GoodEnoughForMe]

The secret is that it’s always been ‘too complex’ but at least now we have easy access to information for those who want to learn

 

[Farm Boy]

I think the technocrats elite type things are more complex. Day to day life has gone the other way.

That's onto something. The razor's edge is probably sharper. But a razor's edge is thin and ridiculously high maintenance. It's maintained with hammering, stone, and sweat. Take those away, no razor's edge. There is something indelible lost when a music video can be watched billions of times, but the listeners never gather at a (table) to sing.