Less Wrong is a community blog devoted to refining the art of human rationality. Please visit our About page for more information.

The map of future models

6 Post author: turchin 03 July 2016 01:17PM

TL;DR: Many models of the future exist. Several are relevant. Hyperbolic model is strongest, but too strange.

Our need: correct model of the future
Different people: different models = no communication.

Assumptions:
Model of the future = main driving force of historical process + graphic of changes
Model of the future determines global risks

The map: lists all main future models.
Structure: from fast growth – to slow growth models.
Pfd: http://immortality-roadmap.com/futuremodelseng.pdf

 

 

 

 

Comments (22)

Comment author: ChristianKl 05 July 2016 03:19:16PM 1 point [-]

When the economy grows every year by 3% that's exponential growth and not linear growth.

Comment author: Piecewise 04 July 2016 04:08:04PM 2 points [-]

I think you're incorrect about the wave model when you say it doesn't take technological progress into account. I think what it assumes is that technology will either not progress fast enough to halt the downfall, or that the downfall itself will halt or hinder technological progress. After all, tech progress requires an environment with not only the luxury of time but also an abundance of materials. The exhaustion of certain material resources, the loss of trade with producers of certain resources, loss of communication with others in the field slowing the exchange of information, all these things could strangle the meaningful advancement of technology in the crib.

And all this also assumes that we could come up with and implement meaningful solutions in the first place. There is no guarantee of that. If, for instance, the downfall is triggered by a pandemic as is listed as one of the threats, there is no guarantee that we'll be able to manufacture a cure in time. And the loss of a significant portion of humanity would cripple our various industries. And this would cripple research and development by limiting the resources available to them.

Many people seem to forget that the advancement of technology depends on a complex web of supporting industries. Even something like an iphone requires materials and production steps that take place across the entire world, let alone the high tech components for something like an AI running supercomputer. I think the wave model is less a prophecy of doom that fails to take advancement of tech into account and more a prediction of the results of a truly catastrophic event. The idea that, as we struggle towards the singularity, something kneecaps us and we fall. And that this either prevents us from ever reaching that singularity or at least it greatly sets us back in our attempt to reach it.

Of course thats just my interpretation.

Comment author: turchin 04 July 2016 08:16:09PM 0 points [-]

The model just say that the tech. progress is not the main and overwhelming power. In this case all you said is valid.

Your point about "postapocaliptic" world model is interesting and I will consider its including. But I think it should be different model.

Comment author: turchin 04 July 2016 11:13:00AM *  0 points [-]

Update: How to estimate validity of future's models?

  • Refutability – check early predictions of the model. Popper's criteria. The model should allow falsification.
  • Completeness – the model should include all known strong historic trends.
  • Evidence base – “predict past” + built on large empirical base.
  • Support – Who: thinkers and futurists. How: independently
  • Is it planning (normative) model? No wishful thinking. Planning must be based on some model of the future.
  • Complexity. Too complex = false. Too exact = false.
  • Too strong predictions for near future contradict Copernican mediocracy. Because if we randomly chosen from the time when the model works, we should be somewhere in the middle of it. For example, if I predict that nuclear war will happen tomorrow, I bet against the fact that if didn’t happen for 70 years, and it’s a priory probability to happen tomorrow of very small.
  • Principle: strongest process wins. We should choose the model that predict most powerful changes. For example nuclear war would have stronger effect on demographic than space exploration.
  • Quickest process wins. Technological development predicts large changes in next 20 years and resource's depletion in, say, next 50 years. In this case tech, development is better prediction model than resource analysis.
Comment author: MrMind 04 July 2016 08:24:00AM 0 points [-]

While most of these models are intuitive to grasp, I noticed I was confused by the spiral. So I followed the link, but the article doesn't seem to explain it at all.
According to the text, we are living in an era of ever accelerating "yangization", whatever that might mean. But the picture instead shows history as a process with two ever-dwindling parameters (cockiness / horse-iness), which are never explained in the text.
Now I'm more confused than when I started.

Comment author: turchin 04 July 2016 10:56:18AM *  0 points [-]

One of the proponents of spiral development was Karl Marx. He claimed that history will come to the same constructions but on the higher level of development. E.g. he claimed that primitive communism of tribal people will be reached again on the highest level of society development. He probably got the idea of spiral development from Hegel.

Spiral development is combination of idea of progress and idea of circularity of history.

About yangization... I think I have to update the link ))

Comment author: MrMind 04 July 2016 03:32:31PM 0 points [-]

If I remember my philosophy lessons in high-school, Hegel was the proponent of the idea that history faces moments where it develops a thesis, which eventually makes the opposite thesis grows, until finally the two are reunited in a synthesis. That is, a progression, circular in nature, but still expanding.
It's curious that the spiral in the picture grows inwards, not outwards. Weird.

Comment author: turchin 04 July 2016 08:02:13PM *  1 point [-]

I think your memory of Hegel is correct.

By the way he also thought that Prussia of the beginning of 19 century is ultimate end of history. He also thought that the history is a story of "establishment of the Spirit", and so the God is in the future - so he was near prediction of future superhuman AI, but idea of Prussia as ultimate state prevented him from following his own logic to its end. This of cause is just my interpretation.

The picture is result of internet search of an image by keywords "spiral development". But the fact that it is inward could make sense if we think of it in the context of exponential acceleration. If it will be 3D, it will be a conus.

Comment author: MrMind 04 July 2016 08:26:53AM 0 points [-]

Ah, but in the spiral there is a global risk, a sort of "yangization catastrophe" that we will enter if we are not able to tame history with a process of "yinnization":

We might expect an explosion: we reach the center point and most of mankind is wiped out by some cataclysm, perhaps self-inflicted.

Comment author: Elo 03 July 2016 11:10:51PM -2 points [-]

I like that this is a collection of graphs. Visual representation is excellent.

Comment author: turchin 03 July 2016 11:13:39PM 1 point [-]

Yes, I am constantly searching for better ways of presenting ideas. "Wall of text" is proven failure mode for me.

Comment author: James_Miller 03 July 2016 09:04:59PM *  0 points [-]

Excellent. My view is that the future is either human extinction or unpredictable.

Comment author: MrMind 04 July 2016 07:23:43AM *  0 points [-]

Is that P(x| NOT human extinction) = ?? or just P(NOT human extinction) = ??
Because the second case seems contradictory to me.

Comment author: James_Miller 04 July 2016 04:49:40PM 0 points [-]

Unpredictability conditional on NOT human extinction.

Comment author: turchin 03 July 2016 09:08:22PM 1 point [-]

I used to say "extinction or immortality", where under "immortality" I mean galactic supercivilisation dwelled by immortal posthumans

Comment author: James_Miller 03 July 2016 09:41:43PM 1 point [-]

In attempting to get people to sign up for cryonics I have learned that most people don't want to live for a very long time.

Comment author: turchin 03 July 2016 09:48:03PM *  1 point [-]

Yes ... (( Ben Goertzel suggested another term: "potentially indefinite life extension". It is almost the same as immortality, but people may feel freedom on ontological level with it.

Nobody wants to be trapped forever in unpleasant conditions, and to become immortal is the first step to such situation )) So we should address this concern.

Comment author: qmotus 05 July 2016 08:50:36AM 0 points [-]

I wouldn't call cryonics life extension; sounds more like resurrection to me. And, well, "potentially indefinite life extension" after that, sure.

Comment author: turchin 05 July 2016 10:28:51AM 0 points [-]

Another good wording is "respawning", which came from computer games, and means ability of a character to appear again, something like resurrection.

Comment author: Yosarian2 03 July 2016 04:52:59PM *  1 point [-]

A 3% a year economic growth model is actually an exponential growth model as well, it's not linear at all. The exponential growth rate is a lot lower then Moore's law, 1.03^n instead of 2^n, but it's still exponential growth.

Comment author: turchin 03 July 2016 06:07:43PM 1 point [-]

That is true, but those who use this model predict typically not for long - 5-10 years, where exponential nature of such growth is not much evident. But in demographic and long term economic prediction it becomes clearly exponential. Probably I should reformulate and separate slow, but exponential growth, from actually linear predictions.

Comment author: Yosarian2 03 July 2016 07:59:00PM 1 point [-]

Yeah; don't underestimate slow but exponential growth, I would say that 3% a year growth is what really started at the start of the industrial revolution and why we've gotten as far as we have.

Personally I expect overall civilization wide exponential growth at some intermediate level for the next few decades faster than 3% but slower then Moore's law. Lot of uncertainty in there of course.