Saturday assorted links

Comments

#5. Cats eye research on their behaviors with contained bemusement, flick their tails in amazement, stride off to nap secretly for an hour or two, announce their hunger in order to salivate near the electric can opener, complain that their litter pan needs cleaning or re-filling, bathe and groom before napping for another hour or two . . . . as long as bipedal mammals can keep feline names straight, life is good.

that is all true and very well expressed,
but my cats often looked at me with honesty and friendship in their eyes, which is almost as good as gratitude

just kidding honesty and friendship are way better than gratitude

Way too complicated.
The way it works:
You like a post, tap your smart card on the screen and transfer a bearer asset penny to the author. Then read the post. No transaction costs.

This is about #6, presumably.

I think this sort of idea is covered. The worry is that if the number of such likes is used to rank what people see, then the system will devolve into advertising.

But maybe that's fine: we already have a system in which both content can be funded by both advertising and donations. A payment-upvote system would simply make the donations more convenient than they are now.

It's complicated the same way any other money system is complicated: cheats are all over the place. It's also the same reason we can't have a simple tax system.

And if, after reading the post, you find that you actually don't like it? You get your money back?

The problem with this "like" system is that commenters are too prone to start acting like politicians and political parties. One hangs out and communicates only with those who already tend to agree with you and your comments are not the result of actual thought followed by the attempt to persuade, but are designed merely to garner favour. This has nothing to do with "collusion" and everything to do with predominant human nature. The social media phenomenon, I'm pretty sure, explains in large part our increasing partisanship and division, not merely in the political, but now also in the social sphere. In fact, a large number of the population seem to have lost the ability for critical thought.

1. "This supports our overall theme that the productivity growth slowdown from the early postwar years to the most recent decade was due to a retardation in technical change that affected the same industries by roughly the same magnitudes on both sides of the Atlantic." I don't know about "retardation of technical change", but I do know that a drop in investment in productive capital will cause productivity to fall. Since the authors of the linked paper have brought up the subject of retardation, does anyone have an explanation for the retardation in economic research?

#1 - Apropos of something, the Soviet Union in the 1970s also had a wage stagnation, just like the West. This is not to be confused with productivity, but it's part and parcel of the same 'slowing down' phenomena.

Could Woodstock really have been that bad for morale, world-wide? I think not. The blame IMO is the rise in local and state governments, worldwide, which mushroomed after WWII and have held back creative juices (and bad patent laws). I mean, if you're a Medicare innovator, would you risk going to jail like that present FL exec just for cutting corners? If you're an energy guru like J. Skilling, would you give up 13 years of your life to risk innovating in energy (note some of Enron's businesses now are hot again)?

Bonus trivia: I own Enbridge, a Canuk pipeline company, gives about 8% dividend, not bad at all, supplies the US NE, captive customers.

I’m not sure that Buterin understands game theory. He references two variants of game theory, non-cooperative and cooperative, and states that only the latter “allows” for collusion. This is wrong in two ways. First, collusion can be sustained in repeated games where actors are non-cooperative; indeed most theoretical and empirical work on collusion over the past several decades has taken this approach. In this setting, collusion is sustainable so long as the short-term profit gain to cheating on the cartel agreement is less than present value of the “punishment” phase once cheating is detected (where punishment might simply be reversion to more competitive and less profitable outcomes). This is all fairly basic IO.

Second, cooperative game theory doesn’t simply allow for collusion, it assumes it. This was originally developed to model bargaining in a simple, axiomatic way. One way to capture this idea is that the bargainers’ incentives are aligned with regard to maximizing the size of the pie created through their mutual efforts. So they act to make the pie as large as possible. So far, no problem, but their incentives are not aligned with regard to how large a slice of the piece each should get. That’s where the axioms come in (one of which is that the bargaining is efficient; i.e., the maximal pie is achieved and fully apportioned to one or the other party). In a simple dynamic, non-cooperative game of bargaining, maximization of joint surplus is likewise assumed at the outset.

Of course, in the real world there are myriad ways that cooperation can break down, whether that cooperation is illicit (price-fixing collusion among substitute providers) or serves a salutary business purpose, such as creating a high quality product. Much of IO is working through such issues. Assuming commitment power (which is my understanding of what Buterin means by collusion within a coalition of cooperative play) is a dead end. Analyzing the sources of commitment power, how such power can be supported or undermined, is more useful.

I would suggest to Mr. Buterin that “collusion” is not as helpful an organizing principle for cataloguing the problems that platforms face as “principal-agent problems,” of which collusion among agents is one variety, but far from the only one. There is a large mechanism design literature on the topic. I’d go further to suggest that platforms such as Buterin’s may be the next great laboratory for studying principal-agent problems and their mitigation or solution.

I'm well aware of iterated game equilibrium, folk theorem results, etc. I apologize if I used the wrong language; but there is definitely a very real sense in which large classes of games give optimal results in one-shot non-cooperative models, but suboptimal results where collusion is permitted. Consider second price auctions. The standard textbook reasoning is that they are optimal because (i) your bid never influences the price you have to pay, as the only case the bid affects the price is when you're second place, but in that case you don't win, and (ii) you get the item if and only if your valuation is highest. But (i) breaks down when collusion between the top two bidders is possible, because the highest bidder could pay the second highest bidder to drop their bid to zero, splitting the surplus (namely, the second highest valuation minus the third highest valuation). Implicitly assuming impossibility of collusion especially in public contexts where everyone can see everyone else's behavior is a very common error in practice.

Thanks for your thoughtful reply, Mr. Buterin. I didn’t mean to quibble about semantics or deny that collusion can be a serious problem. It can. My point is simply that some tools already exist for analyzing this problem. Also that coordination among agents is not universally a bad thing. On a broad platform, some agents perform complementary tasks. Here coordination typically benefits platform users.

You sound like a smart guy with your long sentences trying to outpedant Buterin. But your braininess in game theory is also why you didn't get rich with bitcoin. Game theory is no better than astrology in the real world bro.

#5) Before being made aware of this article, I would have assumed that domesticated cats recognized their names. After reading the article, it's not obvious to me that cats do recognize their names. The cited studies seem at best inconclusive.

Some cats may not. Their owners might not have spoken their names in a manner that habituated them into a response.

Also, as the research suggests, they might not care I'd someone calls their name. The trick then is to determine whether the responses were actual or false positives.

You should be proud, hun.

4. As we have learned in earlier posts, corporations became entitled to the rights of "persons" under the constitution at the same time as African Americans (pursuant to the 14th amendment). Hanson seems to say that, since corporations are "persons", don't expect any more from them than one would expect from the worst person you know.

Don't you chaps differentiate between legal persons and natural persons?

If not why not?

#6...Way above my pay grade.

#2 sprawling! But interesting.

Not sure about his overall point. The article seems to hit the following points: governments are useless at producing things, perhaps by intentional design, sometimes a few Mavericks come in losen up rules and magic occurs (likely there were spectacular failures as well, but they don’t get as much press), then a side discussion about our nuclear stockpiles and the lack of security, the lack of focus in the west on operational excellence vs. China which maybe is getting it right, followed by the impressive gains in computation over the last 75 years, followed by the potential rise of AI and the alignment problem, followed by maybe getting the government involved in AI?

I’m not sure what to get out of it. Maybe he wants to get government into AI to slow it down? Maybe he is worried that China will get to AI 1st? Somehow maybe he is worried about AI launching nukes before it become true AGI?

Maybe a problem statement? Maybe this is just background for future articles? Maybe he is simply bucking for a job running some AI safety program?

Did anyone get a cohesive narrative from that interesting article?

I got out of it "western governments are woefully incompetent and will screw up tomorrow's war in a fundamental way... but that might not be true of China." The piece is either insane or terrifying.

More people here need to read Dominic Cummings. If only to learn about risks in the pipeline (i.e., drones and pandemics).

7# fortnite is creating the next generation of incels they start them so young

#2 The author brings to mind the Tommy Lee Jones character in "Under Siege."

#4 is possibly the most absurd post in the history of the internet. Making the case that we should act on benign fictions about "shark-like legal entities devoted to commercial profit" deems sociopathic at best.

Comments for this post are closed