3 ms·
OK, I'll do so. To put it brutally it's because on Hacker News and probably your social circle, there are too many people whose social status and income depend
by native_samples 5y ago
OK, I'll do so.
To put it brutally it's because on Hacker News and probably your social circle, there are too many people whose social status and income depends on being clever, or at least the perception of being clever. This leads us to seriously over-index on things that look like science or data analysis, without even being able to see alternatives that in my experience easily occur to the less (over?) educated.
Consider this article. It's actually a good case in point. The article can be summarized like this:
1. In the 1970s there were these modellers who predicted disaster. Lots of very influential and educated people took them uber-seriously and thought the Limits To Growth was proof positive that everything needed to change.
2. The model turned out to be useless and the disaster it predicted never happened, because the numbers it used for natural resource availability were garbage.
3. This major error did not reduce the modellers confidence in what they were doing (this seems to be a common theme with modellers). Instead they just doubled the number of natural resources and kept on going. That sounds rather arbitrary and not like a real fix to me, but alright. The new models indicated that nothing makes much of a difference for decades (until 2020), conveniently making the predictions unfalsifiable on any timeline that the original authors may have cared about.
4. Now a new researcher has published a new paper on these models, which also says nothing useful. We read that the models "aren't terribly enlightening" because we don't have any data past 2020 yet so which scenario we're supposedly in cannot be determined.
5. Despite all the above, the author concludes that "the most important takeaway from these models is that we need to take definitive action".
Wrong! A rational person can conclude nothing from the evidence in the article. Given that we can only judge the models by prior performance, we should anticipate that the models are probably wrong again. Moreover, the model scenarios don't have any probabilities attached to them and are described in sufficiently vague terms that nobody would ever be able to conclude which scenario we actually followed, again, making it an exercise in stargazing, not science.
Therefore, we cannot actually say we've "known for at least 50 years that our current pattern of consumption is not sustainable". We've known no such thing. The conclusions of the article and many of the commenters here seem like motivated reasoning as a consequence - or at least, it's certainly not Bayesian reasoning!
- kokey 5y agoWe use increasingly sophisticated models to make the same mistake they made 130 years ago when they projected the increase of horse manure.
- cjohnson318 5y agoThings always seem fine until they aren't, and then they're real bad. Invariably, the nature of a difficult situation is that they're difficult to get out of, so we tend to be very interested in avoiding difficult situations. This is why people are so interested in avoiding climate change, rather than waiting for more real-time data about how much climate change actually sucks so they can update their Bayesian priors and "avoid climate change when it happens again."
- pksebben 5y agoThis is specifically what I'm referring to. Temperatures are already getting lethally hot and we're not really doing all that much about it. It's not resource scarcity that seems like the killer, here.
- native_samples 5y agoIt's worth noting that the Limits to Growth didn't talk about climate change, just generalized 'pollution'. One now obscure reason for this is that back in the 1970s people didn't believe in global warming and the term "climate change" didn't exist as we know it today. Climatologists were worried about global cooling instead. At that point their records showed a generalized decline in temperature from the 1940s onwards and they believed it might continue indefinitely, leading to a new ice age. This was taken quite seriously, as newspaper records from the time show. These days it can be hard to understand why they thought that, because if you get a modern graph of the long term temperature record it doesn't show this cooling effect. The reason is that climatologists have repeatedly issued new versions of the historical time series, each time reducing the extent to which this cooling period can be seen. The differences between versions are the extent to which models are allowed to modify the raw data. The climatologists do this because they are very certain in their theories, but a long period of cooling in the 20th century is not compatible with them, and thus they look for reasons why the data must be wrong. When they come up with such a reason it gets merged into the model and the next version "corrects" for it. Modern graphs show basically no cooling, leading to some people concluding that the belief in global cooling was some sort of fringe belief, or even an urban legend, because obviously, how could they have thought that when no such cooling is visible. Given the long track record of educated people having near unshakeable confidence in models even in the face of failed predictions (of which World3 is a salient example), this progressive replacement of measured historical data with synthetic data is well worth pondering deeply.