Many people these days talk about an impending “fourth industrial revolution” led by AI, the internet of things, 3D printing, quantum computing, and more. The first three revolutions are supposed to be:
- 1st industrial revolution (~1800-1870): the world industrializes for the first time via steam, textiles, etc.
- 2nd industrial revolution (1870-1914): continued huge growth via steel, oil, other things, and especially electricity.
- 3rd industrial revolution (1980-today): personal computers, internet, etc.
I think this is a misleading framing for the last few centuries, though, because one of these things is not remotely like the others. As far as I can tell, the major curves of human well-being and empowerment bent exactly once in recorded history, during the “1st” industrial revolution:
(And yes, there’s still a sharp jump around 1800-1870 if you chart this on a log scale.)
The “2nd” and “3rd” industrial revolutions, if they are coherent notions at all, merely continued the new civilizational trajectory created by the “1st” industrial revolution.
I think this is important for thinking about how big certain future developments might be. For example, authors of papers at some top machine learning conference seem to think there’s a decent chance that “unaided machines [will be able to] accomplish every task better and more cheaply than human workers” sometime in the next few decades. There’s plenty of reason to doubt this aggregate forecast, 1 but if that happens, I think the impact would likely be on the scale of the (original) industrial revolution, rather than that of e.g. the (so small it’s hard to measure?) impact of the “3rd” industrial revolution. But for some other technologies (e.g. “internet of things”), it’s hard to tell a story for how it could possibly be as big a deal as the original industrial revolution.
Footnotes:- E.g. answers differ depending on how you ask the question, we should worry about response bias, and it’s not clear whether AI scientists or anyone else can make reliable long-term forecasts of this sort.[↩]
We might also consider that those lines were trending upwards already, before the drastic upturn in the late 18th Century.
The acceleration in energy demand, led by ‘manufactuaries’ of looms and pottery kilns, and fulfilled by water- and then by steam-power, gave the society of the day the power to do more of what they were already doing.
What then, are the effects of another acceleration in the current trends to concentrated wealth and greater inequality, decreasing life expectancy, pervasive surveillance, and the hollowing-out of democracy by an ‘information revolution’ that is no longer about the information we might use to our benefit, and all about the information we are fed?
I think the log plot is basically the only plot worth talking about, and moreover the chart that displays the entire history (4000 BC – 2000 AD) can’t actually settle the question addressed in the post. The reason is that we’re trying to see how much the slope of the line on the log plot changes within the span 1800 to 2000 — when the putatively distinct revolutions occurred — and all the plots linked above squish that span so much that it’s impossible to tell. In particular, the existence of a sharp jump around 1800 does not rule out the existence of equally sharp jumps at (say) 1910 and 1985, which would reasonably be interpreted as equally important revolutions, simply because they would all be jammed on top of each other.
To see this, consider this log plot of the entire history, with Luke’s real data in blue, and some fake data I generated in orange: https://imgur.com/a/FONxA . My fake data is exponential growth with doubling times of 10k years before 1700, 500 years between 1700 and 1910, 30 years between 1910 and 1985, and 5 years between 1985 and 2000. The fake data has been constructed to clearly exhibit discrete revolutions of roughly equal importance. Yet the real and fake data look basically indistinguishable if plotted on the entire range starting with 4000 BC. (Adding some noise, black death, etc., would make it even more difficult to tell them apart.)
However, it you look at the log plot only for the range 1600 to 2000, you see what’s going on much more clearly: https://imgur.com/a/CQ4cU . In this case we can see the distinct transitions in growth rates in the fake data, and we see that the real data is maybe well described by a single kink at 1800 (although it’s not so clear).
Finally, I want to argue that the real data is actually compatible with three very discrete and equally important revolutions. All you need to assume is that the effects of a revolution take time to spread geographically. When averaged over the Earth, this will tend to blur the kinks in the graph as different regions get on the triple-kinked growth curve at different times.
In order to argue against the importance of “modern” industrial revolutions like the putative IT revolution post-1980, I think you need to look at better and more fine-grained data, such as the (not super impressive) growth rates of countries like the US that adopted IT first.
I can see a case for charting GDP/cap and kcal/cap/day on a log scale, but life expectancy, % not living in extreme poverty, and % living in a democracy seem like odd fits for a log scale. (War-making capacity is harder to say, given how the index was constructed.) Thoughts?
Certainly I agree more fine-grained data would help; I just didn’t find any for these global variables. Looking at regional data would also help, but that would require more effort than I was willing to expend, and would require a lot more judgment calls.
Also, are you aware of economic historians who think the IT revolution was comparable in magnitude to the British industrial revolution? If so, I could check what arguments they make.
Re: the importance of the IT revolution, note that I did link one article about that, via the “hard to measure” phrase.
Sorry, I should have made it clear that I was just talking about GDP/capita. I agree percentage quantities are not well suited for a log scale unless they are very small (or very large and you are plotting log of the difference from 1).
Note that, especially when you focus on metrics with both upper and lower bounds (e.g., % not in poverty), the choice of metrics gives you a lot of freedom to find a revolution almost anywhere. For instance: fraction of homes with electrical power, fraction of population with access to most of the world’s knowledge (the internet), divorce rate. Extreme poverty is pretty fundamentally important and objective, but democracy maybe not as much.
To be clear, I’ve never seen good data that there were multiple revolutions in a meaningful/important sense (although I’m a complete novice at this). Occam’s razor and the data I’ve seen suggest there was just the one, and in particular I see no reasons to reject the expert consensus (?) that the impact of IT has been very modest. I was just trying to caution that the data you’ve discussed is not super conclusive and seems compatible with multi-revolution models.
Sounds like I don’t disagree much, then? But maybe it’s not a sufficiently clear caveat to say “As far as I can tell,” and link that phrase to a post titled “Three wild speculations from amateur quantitative macrohistory,” which says “That said, the long-run estimates I rely on are pretty uncertain, and my particular choice of measures to capture ‘well-being’ is obviously questionable (and to some, no doubt objectionable), so I will still refer to the basic picture presented above as ‘speculative,'” and which links to a longer report full of even more caveats.
Nope! I don’t think we disagree on conclusions, and any apparent disagreements on which evidence best supports those conclusions easily lies within the ambiguity that’s unavoidable when compressing the history of humanity into a brief blog post 🙂