Abstraction learning in general is an area where I’m not yet fully satisfied with my own understanding, but I’ll see if I can set up anything interesting around this.
johnswentworth
Karma: 9,916
Given how much the comments on this one diverge, sounds like there’s a lot of confusion around it (some of which is confusion around how words work more generally). Guess I’d better talk about it.
I will be focused more on the abstraction aspects than the game-theoretic aspects, though.
Services are not “one sector”, they’re generally considered a category of sectors, and they comprise a majority of all first-world economic activity. If something is true of the service sectors, then it’s true of the majority of the economy.
Your eleven business sectors are not all dominated by huge conglomerates. Automotive production is, but automotive sales aren’t. Construction isn’t. Electronics isn’t. Finance is mixed bag, depending on the sub-sector. Healthcare isn’t. Insurance has big conglomerates on the backend, but lots of small independent salespeople on the frontend. Internet infrastructure is dominated by large companies, but the volume of small e-companies operating on that infrastructure is massive. Oil and gas is more concentrated than most of these sectors, but I’d still guess that you’ve never looked at the numbers enough to notice the long tail of medium-to-small oil producers (I have seen some of those numbers). Pharma is also relatively concentrated, and there the pressure toward aggregation is real. Retail has a long tail. Telecoms is concentrated, though it’s become less concentrated in recent decades.
Note that, in each of these sectors there do exist large conglomerates. That does not mean that the industry is dominated by conglomerates. The big conglomerates are highly visible and salient; the small companies are not.
So out of these, you’re right on maybe half of them if I’m generous. Overall, it sounds like you’re making a really strong claim without ever having looked at the data, and the data does not back the claim.
At some point I need to write a post on purely Bayesian statistical mechanics, in a general enough form that it’s not tied to the specifics of physics.
I can probably write a not-too-long explanation of how abstraction works in this context. I’ll see what I can do.
Ooh, good one. If I remember the trick to the algorithm correctly, it can indeed be cast as abstraction.
This plays well with impact measures, too. I can definitely include it.
Of course there have been particular cases where an industry consolidated during a particular period. You made a much stronger claim: that industries in general tend toward consolidation. Pointing to two or three examples where industries consolidated does not provide much evidence for such a claim. On the other hand, pointing to examples where industries did not consolidate provides significant evidence against such a claim.
[Question] Problems Involving Abstraction?
Restaurants, car dealerships, spas and hair salons, construction, plumbers and electricians, doctors and lawyers. Every industry dominated by small businesses.
Even in some of the sectors you list—like automotive manufacturing—we haven’t seen much net consolidation. We haven’t seen a lot of new entrants, but’s it’s not like the number of car manufacturers is rapidly decreasing either. It’s at an equilibrium, and that equilibrium has a lot more than just one company—which is not something you’d see if economic forces generally favored consolidation.
A review of The Design of Everyday Things, ideally with some discussion of how the ideas there intersect with rationality-adjacent topics.
Should be working now.
If anyone tries this, I’d be interested to hear about the results. I’d be surprised if something that simple worked reliably, and it would likely update my thinking on the topic.
Interesting question.
I don’t think that shelter in place has had that large an effect on most of the kinds of interaction relevant here. People still have largely the same economic interactions—we’re still buying mostly the same products from mostly the same sources. There’s maybe some shift in middlemen—more Amazon, less brick-and-mortar—but our interactions with the brick-and-mortar Walmart store weren’t really any more personal than the interactions with Amazon anyway. There might have been a time when everyone was on a first-name basis with their grocer, but those days were long past even before COVID.
To the extent that COVID has had an effect, I expect it’s probably increased the proportion of repeated interactions, rather than decreased that proportion. There just aren’t as many opportunities to go out and interact with strangers. We’re interacting less with cashiers or waiters or bartenders or hotel staff. Conversely, we’re still interacting about as much with family or coworkers—those close ties are relatively less sensitive to the barriers created by COVID.
Your thoughts on abstraction here are exactly on the mark.
For the cluster-intersection thing, I think the biggest barrier is that “biggest cluster intersection” implicitly assumes some space in which the cluster intersection size is measured—some features, or some distance metric, or something along those lines. A more complex model will likely be using a different feature space, a different implicit metric, etc.
I feel like photography (and in particular, photographs showing interesting physical systems) is a particularly good fit for “rationalist” artwork. For me, one of the top two or three most fundamental principles of rationality is entangling oneself with the world—living in the real world, and appreciating the real world, rather than escaping into one’s own imagination. That’s not to say imagination is bad, but rather that the role of imagination is to act on the real world, rather than to live on a separate plane of existence. Let the real world inform our beliefs, and use that knowledge to bring our ideas into physical reality.
Photography is a great fit for that, because it portrays physical reality. In some cases, a photograph can help us entangle our own beliefs with reality—as in scientific photographs, like an image of a cell with certain organelles stained. In other cases, a photograph can display the form of some brilliant idea made real—as in the photograph above.
/https://public-media.si-cdn.com/filer/15/97/15972379-a704-4c12-8848-196d954896dc/42-55992923.jpg)
Nicola Tesla
I love the ambience of dark fascination in this one. Tension between horror and curiosity.
I have an out-of-print national geographic book called “Inventors and Discoverers” with a bunch of great pieces in it, but I’m having trouble finding good versions of them on the web. A few particular favorites:
Ceramic which won’t crack under large temperature gradients—originally a Corning ad Charles Goodyear, experimenting in his kitchen Stephenson’s Locomotion, racing (and winning) against all comers, on the opening day of the Stockton & Darlington Railway in 1825 Why someone designed a particle accelerator to have discharges like this I do not know, but the visual effect is great. A young James Watt plays with the steam from the kettle
You still seem to be missing the key point: if you want to claim that industries tend toward concentration in general, citing particular concentrated industries isn’t going to cut it. You need to argue that industries which don’t concentrate don’t exist (or are at least rare), not merely that industries which do concentrate do exist. That requires a more comprehensive view. You cannot get there just by picking particular industries and arguing that they’re concentrated.