Not sure I understand the reason this kind of model keeps getting spawned every few years. Idealized models with no relevance to empirical observations are nowhere near <i>predicting</i> civilization collapse. They can't even explain the observed past trajectories.
Neat package.<p>For reference, the field of looking at future civilization direction w/r/t climate change is called "Integrated Assessment Models," and DICE (from Nordhaus) is one of the most used and also open source <a href="https://williamnordhaus.com/dicerice-models" rel="nofollow">https://williamnordhaus.com/dicerice-models</a> . Here's the UNFCCC page on them. <a href="https://unfccc.int/topics/mitigation/workstreams/response-measures/integrated-assessment-models-iams-and-energy-environment-economy-e3-models#eq-1" rel="nofollow">https://unfccc.int/topics/mitigation/workstreams/response-me...</a> .<p>All models are wrong, some models are useful.
Watch this video to see what this is about:<p><a href="https://www.youtube.com/watch?v=kVOTPAxrrP4" rel="nofollow">https://www.youtube.com/watch?v=kVOTPAxrrP4</a><p><a href="https://news.ycombinator.com/item?id=29525410" rel="nofollow">https://news.ycombinator.com/item?id=29525410</a>
Reminds me of an old Apple II MECC program from the early 80s called <i>Limits</i>. The user could "tune" various portions of the model and run projections as to when the world would reach its "limits" and collapse. Seems people have dug these sorts of automatic speculations for some time.
It may be instructive to compare individual health, where we have great experience with decline and collapse.<p>Consider a physical activity that imposes a strain resulting in damage to tissue (muscle, tendon, cartilage, what-have-you) that heals fast enough not to accumulate. As you age, your rate of healing declines gradually. At exactly the point where rate of healing matches the rate of damage, you start into a sharp decline all out of proportion to any apparent, experienced event, as damage begins to accumulate at an unceasing rate.<p>All contact-lens users hit such a point where their eyes seem "suddenly" to begin a rapid downslide if they do not abandon contacts.<p>Many systems operating at global level have this character.
To paraphrase Brahms when he first heard the Dvořák cello concerto: if I had known that it was possible to write a package like this, I would have tried it as well!<p>I love this. It would be interesting to pair it with a decent climate model, though.
So an elementary model which cannot possibly hope to model all the real world variables is gonna predict the future? No no, this is too much but it's a cute toy, to play with models.
Can we get the editorialising out of the title (currently: "Predict civilization collapse with Python and World3 model").<p>Putting aside the issues with the data etc, this model isn't even attempting to predict civilization collapse. The closest it comes is an attempt to model future population which isn't obviously correlated with level of civilization.<p>The repo itself doesn't discuss "civilization" and it's unclear why the submitter thinks that is what it is doing.
this model is very sensitive to certain inputs if smidged a little bit yield vastly different projections. Models are more convincing when they are robust to that sort of thing. For instance go here: <a href="http://bit-player.org/extras/limits/ltg.html" rel="nofollow">http://bit-player.org/extras/limits/ltg.html</a> and try increasing the "output consumed (fraction)" input a little bit.
Bullshit, bullshit everywhere.<p>An abstract model can predict nothing, especially when the system is not fully observable, discrete, fully deterministic.<p>Abstract modeling is the modern age astrology.