Exploring Value Space

By Robin Hanson · January 19, 2022 3:05 pm · Comments (0)

If you have enough of a following, Twitter polls are a great resource for exploring how people think. I’ve just finished asking a 8 polls each regarding 12 different questions that make people choose between the following 16 features, either in themself or in others:

attractiveness, confidence, empathy, excitement, general respect, grandchildren, happiness, improve world, income, intelligence, lifespan, pleasure, productive hrs/day, professional success, serenity, wit.

The questions were, in the order they were asked (links give more detail):

  1. UpSelf: Which feature of you would you most like to increase by 1%?
  2. Advice: For which feature do you most want a respected advisor’s advice?
  3. ToMind: Which feature of yourself came to your mind most recently?
  4. WorkedOn: Which feature did you most try to improve in the last year?
  5. UpOthers: Which feature of your associates would you most like to increase by 1%?
  6. City: To which city would you move, options labeled by the feature that people there are on average better on?
  7. KeepSelf: If all your features are to decline a lot, which feature would you save from declining?
  8. Aliens: What feature would you use to decide which civilization survives?
  9. Voucher: On which feature would you spend $10K to improve?
  10. World: Which feature of yours would you most like to improve to become world class?
  11. Obit: Which feature would you feel proudest to have mentioned in your obituary?
  12. KeepOthers: If all of your closest associates’ features will decline a lot, which feature would you save from declining?

Each poll gives four options, and for each poll I fit the response % to to a simple model where each feature has a positive priority, and each feature is chosen in proportion to its priority. The max priority feature is set to have priority 100. And here are the results:

This shows, for each question, the average number who responded to each poll, the RMS error of the model fit, in percentage points, and then the priorities of each feature for each question. Notice how much variation there is in priorities for different questions. Overall, intelligence is the clear top priority, while grandkids is near the bottom. What would Darwin say?

Here are correlations between these priorities, both for features and for questions:

Darker colors show higher correlations. Credit to Daniel Martin for making these diagrams, and to Anders Sandberg for the idea. We have ordered these by hand to try to put the stronger correlations closer to the diagonal.

Notice that both features and questions divide neatly into self-oriented and other-oriented versions. That seems to be the main way our values vary: we want different internal versus external features, and different features in ourselves versus others. Thus ~3/4 of our values are social.

 

GD Star Rating
loading...
Tagged as: ,

Best Case Contrarians

By Robin Hanson · January 16, 2022 4:10 pm · Comments (0)

Consider opinions distributed over a continuous parameter, like the chance of rain tomorrow. Averaging over many topics, accuracy is highest at the median, and falls away for other percentile ranks. This is bad news for contrarians, who sit at extreme percentile ranks. If you want to think you are right as a contrarian, you have to think your case is an exception to this overall pattern, due to some unusual feature of you or your situation. A feature that suggests you know more than them.

Yet I am often tempted to hold contrarian opinions. In this post I want to describe the best case for being a contrarian. I’m not saying that most contrarians are actually in this best case. I’m saying that this is the case you most want to be in as a contrarian, as it can most justify your position.

I recently posted on how innovation is highest for more fragmented species, as species so often go wrong via conformity traps. For example, peacocks are now going wrong together with overly long tails. To win their local competitions, each peacock needs to have and pick the tails that are sexy to other peacocks, even if that makes them all more vulnerable to predators.

Salmon go wrong by having to swim up hard hazard-filled rivers to get to their mating groups. Only a third of them survive to return from that trip. Now imagine a salmon sitting in the ocean at the mouth of the river, saying to the other salmon:

We are suffering from a conformity trap here. I’m gonna stay and mate here, instead of going up river. If you stay here and mate with me, then we can avoid all those river hazards. We’ll survive, with more energy to help our kids, and win out over the others. Who’s with me?

Now salmon listening to his should wonder if genetic losers are especially likely to make such contrarian speeches. After all, they are the least likely to survive the river, and so the most desperate to avoid it. For all its harms, the river does function to sort out the salmon with the best genes. If you make it to the end, you know your mating partner will also be unusually fit.

So yes, those less likely to pass the river test are more likely to become salmon contrarians. But they aren’t the only ones. Also more likely are:
A) those who can better sort good from bad mates in other ways,
B) those who can better see the conformity traps, and see they are especially big,
C) those who can better see which are the best places to start alternatives to the conformity traps, and
D) those who happen to have invested less in, and thus are less tied to, existing traps. Like the young.

Our world suffers from myriad conformity traps. Like investors who must coordinate with other investors (e.g., via the different levels of venture capital), may feel they must do crypto, as that’s what the others are doing. Even if they don’t think that much of crypto. Like academics in fields that use too much math feel they also need to do too much math if they are to be respected there. Like journalists and think tank pundits feel they must write on the topics on which everyone else is talking, even if other topics are more important.

In all of these cases, it can make sense to try to initiate a contrarian alternative. If many others know about the existing conformity traps, they may also be looking for a chance to escape. The questions are then: when is the right time and place to initiate a contrarian move to escape such a trap. Who is best place to initiate, and how? And, what is the ratio of the gains of success to the costs of failure?

In situations like this, the people who actually try contrarian initiatives may not be at all wrong on their estimates about the truth. They will be different in some ways yes, but not necessarily overall on truth accuracy. In fact, they are likely to be more informed on average in the sense of being better able to judge the overall conformity trap situation, and to evaluate partners in unusual ways.

That is, they can better judge how bad is the overall conformity trap, where are promising alternatives, and who are promising partners. Even if, yes, they are also probably worse on average at winning within the usual conformity-trapped system. Compared to others, contrarians are on average better at being contrarians, and worse at being conformists. Duh.

And that’s the best case for being a contrarian. Not so much because you are just better able to see truth in general. But because you are likely better in particular at seeing when it is time to bail on a collective that is all going wrong together. If the gains from success are high relative to the costs of failure, then most such bids should fail, making the contrarian bid “wrong” most of the time. But not making most bids themselves into mistakes.

GD Star Rating
loading...
Tagged as:

Much Talk Is Sales Patter

By Robin Hanson · January 13, 2022 3:50 pm · Comments (0)

The world is complex and high dimensional. Even so, it sometimes helps to try to identify key axes of variation and their key correlates. This is harder when one cannot precisely define an axis, but merely gesture toward its correlates. Even so, that’s what I’m going to try to do in this post, regarding a key kind of difference in talk. Here are seven axes of talk:

1. The ability to motivate. Some kinds of talk can more move people to action, and fill people with meaning, in ways that other kinds of talk do not. In other kinds of talk, people are already sufficiently moved to act, and so less seek such added motivation.

2. The importance of subtext and non-literal elements of the talk, relative to the literal surface meanings. Particular words used, rhythms, sentence length, images evoked, tone of voice, background music, etc. Who says it, who listens, who overhears. Things not directly or logically connected to the literal claims being made, but that matter nonetheless for that talk

3. Discussion of, reliance on, or connection to, values. While values are always relevant to any discussion, for some topics and context there are stable and well accepted values at issue, so that value discussions are just not very relevant. For other topics value discussion is more relevant, though we only rarely every discuss them directly. We are quite bad at talking directly about values, and are reluctant to do so. This is a puzzle worth explaining.

4. Subjective versus objective view. Some talk can be seen as making sense from a neutral outside point of view, while other talk mainly makes sense from the view of a particular person with a particular history, feelings, connections, and concerns. They say that much is lost in trying to translate from a subjective view to an objective view, though not in the other direction.

5. Precision of language, and ease of abstraction. On some topics we can speak relatively precisely in ways that make it easy for others to understand us very clearly. Because of this, we can reliably build and share precise abstractions of such concepts. We can learn things, and then teach others by telling them what we’ve learned. Our most celebrated peaks of academic understanding are mostly toward this end of this axis.

6. Some talk is riddled with errors, lies, and self-deceptions. If you go through it sentence by sentence, you find a large fraction of misleading or wrong claims. In other kinds of talk, you’d have to look a long time before you found such errors.

7. Talk in the context of a well accepted system of thought. Like physics, game theory, etc. Where concepts are well defined relative to each other, and with standard methods of analysis. As opposed to talk wherein the concept meanings are still up for grabs and there are few accepted ways to combine and work with them.

It seems to me that these seven axes are all correlated with each other. I want to postulate a single underlying axis as causing a substantial fraction of that shared correlation. And I offer a prototype category to flag one end of this axis: sales patter.

The world is full of people buying and selling, and a big fraction of the cost of many products and services goes to pay for sales patter. Not just documents and analyses that you could read or access to help you figure out which versions are betting quality or better suited to your needs. No, an actual person standing next you being friendly and chatting with you about the product or whatever else you feel like.

You can’t at all trust this person to be giving you neutral advice. Even if you do come to “trust” them. And their sales patter isn’t usually very precise, integrated into systems of analysis, or well documented with supporting evidence. It is chock full of extra padding, subtext, and context that influences without being directly informative. It is even full of lies and invitations to self-deception. Even so, it actually motivates people to buy. And thus it must, and usually does, connect substantially to values. And it is typically oriented to the subjective view of its target.

At the opposite end of the spectrum from sales patter is practical talk in well defined areas where people know well why they are talking about it. And already have accepted systems of analysis. Consider as a prototypical example talk about how to travel from A to B under constraints of cost, time, reliability, and comfort. Or talk about the financial budget of some organization. Or engineering talk about how to make a building, rebuild a car engine, or write software.

In these areas our purposes and meanings are the simplest and clearest, and we can usefully abstract the most. And yet people tend to pick from areas like these when they offer examples of a “meaningless” existence or soul-crushing jobs. Such talk is the most easily painted by non-participants as failing to motivate, and being inhuman, the result of our having been turned into mindless robots by mean capitalists or some other evil force.

The worlds of such talk are said to be “dead”, “empty”, “colorless”, and in need of art. In fact people often justify art as offering a fix for such evils. Art talk, and art itself, is in fact much more like sales patter, being vague, context dependent, value-laden, and yet somehow motivating.

There’s an awful lot of sales talk in the world, and a huge investment goes into creating it. Yet there are very few collected works of the best sales patter ever. Op-eds are a form of sales talk, as is romantic seduction talk, but we don’t try to save the best of those. That’s in part because sales patter tends to be quite context dependent. It also doesn’t generalize very well, and so there are few systems of thought built up around it.

So why does sales patter differ in these ways from practical systematic talk? My best guess is that this is mostly about hidden motives. People don’t just want to buy stuff, they also like to have a relation with a particular human seller. They want sellers to impress them, to connect to them, and to affirm key cherished identities. All from their personal subjective point of view. They also want similar connections to artists.

But these are all hidden motives, not to be explicitly acknowledged. Thus the emphasis on subtext, context, and subjectivity, which make such talk poor candidates for precision and abstraction. And the tolerance for lies and self-deception in the surface text; the subtext matters more. Our being often driven by hidden motives makes it hard for us to talk about values, since we aren’t willing to acknowledge our true motives, even to ourselves. To claim to have some motives while actually acting on others, we can’t allow talk about our decisions to get too precise or clear, especially about key values.

We keep clear precise abstract-able talk limited to areas where we agree enough on, and can be honest enough about, some key relevant values. Such as in traveling plans or financial accounting. But these aren’t usually our main ultimate values. They are instead “values” derived from constraints that our world imposes on us; we can’t spend more money than we have, and we can’t jump from one place to another instantly. Constraints only motivate us when we have other more meaningful goals that they constrain. But goals we can’t acknowledge or look at directly.

If, as I’ve predicted, our descendants will have a simple, conscious, and abstract key value, for reproduction, they will be very different creatures from us.

GD Star Rating
loading...
Tagged as: , ,

My Old Man Rant

By Robin Hanson · January 12, 2022 5:15 pm · Comments (0)

As a 62 year old man, I think I’m entitled to rant once in a while. But instead of “you kids get off my lawn!”, this is my rant:

In principle, economics can help advise most any decisions, like when to wake up, or whether to own a second car. But there are fixed costs to doing explicit econ analysis, and also persuasion costs when you try to influence the decisions of some audience. Thus econ analysis seems most valuable for the biggest decisions whose the audience respects economists for those decisions. Or perhaps many similar but smaller decisions which can all be analyzed at once in the same framework. As we economists are most known for our work evaluating institutions, and as our institutional choices are some of the biggest ones we have, this all suggests our biggest wins come there.

I was first exposed to economics and libertarianism at the same time, and what most excited me about both were similarities to science fiction: they let me imagine very different social worlds. One could see how we could have very different institutions from our current versions, ones that would also plausibly be better. Yes, one couldn’t be very sure that those worlds would be better. But they gave us new things to try, to test and see if they might be better.

When I was young, theory was king, and I tried to master theory. But since then data has come to be king (and queen), even in econ and libertarian circles. Yet I hadn’t realized just how far that trend had gone until this pandemic. To me the obvious theory question a pandemic raises is: what are good general institutions for dealing with pandemics? I wrote a bit on that early on, but was told then that we instead needed immediate help in a crisis. Which I also tried to offer, but which many hated.

Yet it is now two years into what is looking more and more like an eternal pandemic, and I still haven’t see economists or libertarians talking about better pandemic institutions. While this pandemic has done great damage to libertarian sympathies, I’ve only seen libertarians argue that in this particular pandemic, doing nothing officially would have been better than doing what we did. And I’ve seen economists argue about particular parameter settings of the usual government-run system: rules, subsidies and direct government management of masks, lockdowns, tests, and vaccines. Mostly via data, not theory, analysis.

But I’ve not seen work on if there are better institutional alternatives to these two categories, if not for this pandemic then for future ones. Which to me feels like a deep betrayal of what I most value in econ: our ability to imagine, test, and argue for big institutional changes. Even my immediate (and beloved) colleagues haven’t been interested.

To me, the obvious other category is: law. We are better off having law to deal with many harms we can each do to each other, such as assault, slander, and reneging on contracts. Better than ignoring them, and better than having government agencies more directly manage such behaviors. Yes, our society runs law centrally, and likely law would be better if offered privately. But even so, for many harms we are better off because we now apply law over the other two main solutions of doing nothing officially or direct government management.

For law to work for assault, slander, theft, or car accidents, we need it to be often feasible to bring sufficient evidence to convince a court that a particular person harmed a particular other person to a particular degree at a particular event. If so, we can then sufficiently discourage such harms merely via the threat of such legal penalties. At least if we can sufficiently punish those we find guilty, and if we make it easy enough for complainants to subpoena the evidence they need to make their case.

Law today often ensures sufficient punishment via jail and criminal law, which works even if not as well as would vouchers. Law usually allows parties to subpoena any info relevant to a live case, and it so happens that evidence needed to prove assaults and car accidents lasts long enough to let them be so subpoenaed. With vouchers and the level of surveillance likely soon, I don’t actually think we’d need most of our traffic laws; the threat of lawsuits would be enough.

The main policy problem with pandemics is that some people hurt other people by infecting them. Just like they do in assault, slander, theft, and auto accidents. So law could deal fine with pandemics if we could meet the same two conditions: (1) sufficiently able to punish those who found guilty, e.g. via jail or vouchers, and (2) often enough able to easily-enough subpoena sufficient info to show who did what to whom. It is on that last point that economists, and lawyers, have traditionally thrown up their hands and concluded that law can’t deal with pandemics.

That is, people have just assumed that it is not possible to tell who infected who in a pandemic. At least not often enough for law to be our main way to deal with severe pandemics. So for something like the flu we subsidize vaccines and little else, while for covid we go crazy with government managing many related details.

But today with smartphone tracking we can actually see who was close enough to whom when to have infected them. And if we have spit samples from two people infected with covid, we can compare the DNA in their viruses to see if they match. By combining these two pieces of information, one could make a sufficiently strong case that a particular person infected another particular person with the virus at a particular time and place.

So the question that remains is: should we actually induce sufficient information collection and subpoena power, and sufficient punishment ability, to let law deal with pandemics? That is, on the one hand we might make infecting others a punishable crime, require everyone to have their phone track their locations, to report their infections, and to save regular spit samples. And then let government police pour over these details. Which does sound like a pretty intrusive police state, though perhaps still better than the actual police state we’ve had during this last pandemic.

Or, only during an officially declared severe pandemic we could tell everyone that they must either strictly isolate, or, they can get a “pandemic passport” by agreeing to get a voucher, have their phone track their locations, and regularly save spit samples, all available only to be subpoenaed in case of lawsuits by people who claim to be harmed, but not for general browsing by a police state.

Yes, once a pandemic becomes nearly endemic, frequent infection events could clog up courts. But at such scale vouchers would streamline their processes and settle almost all cases out of court. I also know of ways to greatly cut court costs. And damages awarded might greatly fall once one could credibly argue that the victim would likely have caught it soon from someone else.

This idea of legally requiring people to save info so that it can be available to be subpoenaed for future lawsuits is not a particularly new idea. It is just the application to the case of pandemics that would be new. But in our new world of greatly increased surveillance and info of various sorts, we should in fact be thinking about how all that new info might help us solve problems. Like pandemics. Via new institutional changes

Come on, don’t any economists or libertarians out there want to think about new pandemic institutions?

GD Star Rating
loading...
Tagged as: , ,

My 11 Bets at 10-1 Odds On 10M Covid deaths by 2022

By Robin Hanson · January 10, 2022 7:34 pm · Comments (0)

In February 2020, I made many bets on Covid19, including 11 bets at ten to one odds on if it would cause 10 million deaths worldwide by 2022, as estimated by WHO.

WHO has a Q&A page on Covid excess deaths that includes this section:

Why is excess mortality the preferred measure? … aggregate COVID-19 case and death numbers … being reported to WHO … under-estimate the number of lives lost due to the pandemic … In light of the challenges posed by using reported data on COVID-19 cases and deaths, excess mortality is considered a more objective and comparable measure that accounts for both the direct and indirect impacts of the pandemic.

This WHO page, updated daily, lists reported deaths. This WHO page estimated “The true death toll of COVID-19”, or world covid excess deaths, as of Dec. 31, 2020. I expect them to post a page like it soon with death estimates as of Dec. 31, 2021. But I doubt those estimates will differ much from The Economist, which as of Dec. 30, 2021 said:

The pandemic’s true death toll; Our daily estimate of excess deaths around the world … Although the official number of deaths caused by covid-19 is now 5.4m, our single best estimate is that the actual toll is 18.6m people. We find that there is a 95% chance that the true value lies between 11.6m and 21.6m additional deaths.


For many bets we agreed that if there were two number estimates instead of one, we’d go with a geometric mean of them. The geometric mean of 5.4 and 18.6 is 10.02.

Here is the current status of my 11 bets, with a link to the bets and the amount I’m owed. (I’ll update this as things change.)

These claim to win, say I should pay them:

No response to queries (both msg & email):

No response since 31Dec:

  • A Twitter msg bet that I’m keeping private for now, $5000

Waiting for official WHO 2021 Excess Deaths page:

Paid to me:

Some say that it is rude of me to brag about winning. But I need to make this bet situation public in order to pressure bettors to make good on their promises.

Some say it is immoral to bet on death. But I didn’t cause these deaths, and my public bets helped convince many to take this problem more seriously, for which they’ve thanked me.

Added 12Jan: Many are talking as if the issue is direct vs. indirect deaths, but I’d be very surprised if more than a third of excess deaths are indirect. Most of them were caused directly by covid, but just not caught by official testing and diagnosis systems.

Added 18Jan: Nature article:

Demographers, data scientists and public-health experts are striving to narrow the uncertainties for a global estimate of pandemic deaths. … Among these models, the World Health Organization (WHO) is still working on its first global estimate, but the Institute for Health Metrics and Evaluation in Seattle, Washington, offers daily updates of its own modelled results, as well as projections of how quickly the global toll might rise. And one of the highest-profile attempts to model a global estimate has come from the news media. The Economist magazine in London has used a machine-learning approach to produce an estimate of 12 million to 22 million excess deaths.

That IHME 95% confidence interval is 9 to 18 million deaths.

GD Star Rating
loading...
Tagged as: ,

On What Is Advice Useful?

By Robin Hanson · January 9, 2022 8:50 pm · Comments (0)

Regarding what areas of our life do we think advisors can usefully advise? Some combination of they actually know stuff, plus we can evaluate and incentivize their advice enough to get them to tell us what they know, plus how possible it is to change this feature.

Yesterday I had an idea for how to find this out via polls. Ask people which feature of them they’d most like to get advice on how to improve it from a respected advisor, and also ask them on these same features which ones they’d most like to increase by 1%. The ratio of their priorities to get advice, relative to just increasing the feature, should say how effective they think advice is regarding each feature.

So I picked these 16 features: attractiveness, confidence, empathy, excitement, general respect, grandchildren, happiness, improve world, income, intelligence, lifespan, pleasure, productive hrs/day, professional success, serenity, wit.

Then on Twitter I did two sets of eight (four answer) polls, one asking “Which feature of you would you most like to increase by 1%?”, and the other asking “For which feature do you most want a respected advisor’s advice?” I fit the responses to estimate relative priorities for each feature on each kind of question. And here are the answers (max priority = 100):

According to the interpretation I had in mind in creating these polls, advisors are very effective on income and professional success, pretty good at general respect and time productivity, terrible at grandchildren, and relatively bad at happiness, wit, pleasure, intelligence, and excitement.

However, staring at the result I suspect people are being less honest on what they want to increase than on what they want advice. Getting advice is a more practical choice which puts them in more of a near mode, where they are less focused on what choice makes them look good.

However, I don’t believe people really care zero about grandchildren either. So, alas, these results are a messy mix of these effects. But interesting, nonetheless.

Added 11am: The advice results might be summarize by my grand narrative that industry moved us toward more forager like attitudes in general, but to hyper farmer attitudes regarding work, where we accept more domination and conformity pressures.

GD Star Rating
loading...
Tagged as: ,

To Innovate, Unify or Fragment?

By Robin Hanson · January 1, 2022 10:30 am · Comments (1)

In the world around us, innovation seems to increase with the size of an integrated region of activity. For example, human and computer languages with more users acquire more words and tools at a faster rate. Tech ecosystems, such as those collected around Microsoft, Apple, or Google operating systems, innovate faster when they have more participating suppliers and users. And there is more innovation per capita in larger cities, firms, and economies. (All else equal, of course.)

We have decent theories to explain all this: larger communities try more things, and each trial has more previous things to combine and build on. The obvious implication is that innovation will increase as our world gets larger, more integrated, and adopts more wider-shared standards and tech ecosystems. More unification will induce more innovation.

Simple theory also predicts that species evolve faster when they have larger populations. And this seems to have applied across human history. But if this were generally true across species, then we should expect most biological innovation to happen in the largest species, which would live in the largest most integrated environmental niches. Like big common ocean areas. And most other species to have descended from these big ones.

But in fact, more biological innovation happens where the species are the smallest, which happens where mobility is higher and environments are more fragmented and changing. For example, over the last half billion years, we’ve seen a lot more innovation on land than in the sea, more on the coasts than on the interiors of land or sea, and more closer to rivers. All more mobile and fragmented places. How can that be?

Maybe big things tend to be older, and old things rot. Maybe the simple theory mentioned above focuses on many small innovations, but doesn’t apply as well to the few biggest innovations, that require coordinating many supporting innovations. Or maybe phenomena like sexual selection, as illustrated by the peacock’s tail, show how conformity and related collective traps can bedevil species, as well as larger more unified tech ecosystems. It seems to require selection between species to overcome such traps; individual species can’t fix them on their own.

If so, why hasn’t the human species fallen into such traps yet? Maybe the current fertility decline is evidence of such a trap, or maybe such problems just take a long time to arise. Humans fragmenting into competing cultures may have saved us for a while. Individual cultures do seem to have often fallen into such traps. Relatively isolated empires consistently rise and then fall. So maybe cultural competition is mostly what has saved us from cultures falling into traps.

While one might guess that collective traps are a rare problem for species and cultures, the consistent collapse of human empires and our huge dataset on bio innovation suggest that such problems are in fact quite common. So common that we really need larger scale competition, such as between cultures or species, to weed it out. To innovate, the key to growth, we need to fragment, not unify.

Which seems a big red loud warning sign about our current trend toward an integrated world culture, prey to integrated world collective traps, such as via world mobs. They might take some time to reveal themselves, but then be quite hard to eradicate. This seems to me the most likely future great filter step that we face.

Added 10Jan: There are papers on how to design a population structure to maximize the rate of biological evolution.

GD Star Rating
loading...
Tagged as: , ,

Innovation Liability Nightmare

By Robin Hanson · December 26, 2021 1:10 pm · Comments (1)

When I try to imagine how our civilization might rot and decline over the coming millennia, my thoughts first go to innovation, as that has long been our main engine of growth. And while over the years I’ve often struggled to think of ways to raise the rate of innovation, it seems much easier to find ways to cut it; in general, it is easier to break things than improve them.

For example, we might press on one of our legal system’s key flaws. Today, law does far more to discourage A from harming B than to encourage A to help B. B can often sue A for compensation when A harms B, but A can rarely sue B for compensation when A helped B. Law. Today is mostly a system of brakes, not of engines or accelerators.

This is less of a problem for auto accidents or pandemics, where the most important effects of the most important actions are indeed harms. But it is a much bigger problem in innovation, where the main problem is too little incentive to help. In general, society gains far more from innovations than do the people who push for them. So innovation needs engines, not brakes.

The problem is that even events whose effects are overall beneficial will still have some harmful effects. For example, if you invent a new better mousetrap, you may displace previous mousetrap makers. Or by introducing cars, you may hurt people who supplied or managed horses. So what if our legal system makes it easier to sue people for the harms caused by their innovations?

For example, many have complained lately of negative effects of social media, such as increasing anxiety, decreasing privacy, and passing on “fake” news. And just as legal liability has been a big weapon in recent campaigns against harms from tobacco and pain-killers, liability may well also become a big weapon against social media. Wielded especially strongly against those who have most innovated and developed social media.

Imagine that holding innovators liable for the negative effects of their innovations became more widespread. But without increasing the rewards we allow to innovators for the benefits that they bestow. Together with the trend to increased regulation, this might just become enough to kill the innovation goose that lays our golden egg of growth.

GD Star Rating
loading...
Tagged as: ,

Karnataka Hospital Insurance Experiment

By Robin Hanson · December 20, 2021 8:15 pm · Comments (1)

In 2008 I posted on the famous RAND Health Insurance Experiment:

1974 to 1982 the US government spent $50 million to randomly assign 7700 people in six US cities to three to five years each of either free or not free medicine, provided by the same set of doctors. … people randomly given free medicine in the late 1970s consumed 30-40% more medical services, paid one more “restricted activity day” per year to deal with the medical system, but were not noticeably healthier! (More, see also)

I got 60 signatures on a petition then for the “US to publicly conduct a similar experiment again soon, this time with at least ten thousand subjects treated for at least ten years”.

In 2011 I posted on the Oregon Health Insurance Experiment:

Oregon assigned a limited number of available Medicaid slots by lottery. … 8,704 (~30%) [very sick and poor US adults] were enrolled in Medicaid medical insurance. … at most see two years worth of data. … had substantially and significantly better self-reported health. … over two thirds of the health gains … appeared on the very first survey, done before lottery winners got additional medical treatment. (More)

No statistically significant effect on measures of blood pressure, cholesterol, or blood sugar. … did not reduce the predicted risk of a cardiovascular event within ten years and did not significantly change the probability that a person was a smoker or obese. … it reduced observed rates of depression by 30 percent. (More)

Today I report on the new Karnataka Hospital Insurance Experiment:

This study … is amongst the largest health insurance experiments ever conducted … in Karnataka, which spans south to central India. The sample included 10,879 households (comprising 52,292 members) in 435 villages. Sample households were above the poverty line … and lacked other [hospital] insurance. … randomized to one of 4 treatments: free RSBY [= govt hospital] insurance, the opportunity to buy RSBY insurance, the opportunity to buy plus an unconditional cash transfer equal to the RSBY premium, and no intervention. … intervention lasted from May 2015 to August 2018. …

Opportunity to purchase insurance led to 59.91% uptake and access to free insurance to 78.71% uptake. … Across a range of health measures, we estimate no significant impacts on health. … We conducted a baseline survey involving multiple members of each household 18 months before the intervention. We measured outcomes two times, at 18 months and at 3.5 years post intervention. … only 3 (0.46% of all estimated coefficients concerning health outcomes) were significant after multiple-testing adjustments. We cannot reject the hypothesis that the distribution of p-values from these estimates is consistent with no differences (P=0.31). (more)

So a new randomized experiment on ordinary health residents of India had 6.8x as many subjects as the RAND experiment, and also found no net effect on health. It only looked at the effects of hospital treatment, but to many that is the crown jewel of medicine.

Bottom line: we now have more stronger data that on average, more medicine doesn’t improve health. Though of course for people committed to buying useless medicine insurance can cut financial stress. Update your beliefs accordingly.

GD Star Rating
loading...
Tagged as:

Three Types of General Thinkers

By Robin Hanson · December 19, 2021 5:50 pm · Comments (1)

Ours is an era of rising ideological fervor, moving toward something like the Chinese cultural revolution, with elements of both religious revival and witch hunt repression. While good things may come of this, we risk exaggeration races, wherein people try to outdo themselves to show loyalty via ever more extreme and implausible claims, policies, and witch indicators.

One robust check on such exaggeration races could be a healthy community of intellectual generalists. Smart thoughtful people who are widely respected on many topics, who can clearly see the exaggerations, see that others of their calibre also see them, and who crave such associates’ respect enough to then call out those exaggerations. Like the child who said the emperor wore no clothes.

So are our generalists up to this challenge? As such communities matter to us for this and many other reasons, let us consider more who they are and how they are organized. I see three kinds of intellectual generalists: philosophers, polymaths, and public intellectuals.

Public intellectuals seem easiest to analyze. Compared to other intellectuals, these mix with and are selected more by a wider public and a wider world of elites, and thus pander more to such groups. They less use specialized intellectual tools or language, their arguments are shorter and simpler, they impress more via status, eloquent language, and cultural references, and they must speak primarily to the topics currently in public talk fashion.

Professional philosophers, in contrast, focus more on pleasing each other than a wider world. Compared to public intellectuals, they are more willing to use specialized language for particular topics, to develop intricate arguments, and to participate in back and forth debates. As the habits and tools that they learn can be applied to a pretty wide range of topics, philosophers are in that sense generalists.

But philosophers are also very tied to their particular history. More so than in other disciplines, particular historical philosophers are revered as heroes and models. Frequent readings and discussions of their classic texts pushes philosophers to try to retain their words, concepts, positions, arguments, and analysis styles.

As I use the term, polymaths are intellectuals who meet the usual qualifications to be seen as expert in many different intellectual disciplines. For example, they may publish in discipline-specific venues for many disciplines. More points for a wider range of disciplines, and for intellectual projects that combine expertise from multiple disciplines. Learning and integrating many diverse disciplines can force them to generalize from discipline specific insights.

Such polymaths tend less to write off topics as beyond the scope of their expertise. But they also just write less about everything, as our society offers far fewer homes to polymaths than to philosophers or public intellectuals. They must mostly survive on the edge of particular disciplines, or as unusually-expert public intellectuals.

If the disciplines that specialize in thinking about X tend to have the best tools and analysis styles for thinking about X, then we should prefer to support and listen to polymaths, compared to other types of generalist intellectuals. But until we manage to fund them better, they are rarely available to hear from.

Public intellectuals have the big advantage that they can better get the larger world to listen to their advice. And while philosophers suffer their historical baggage, they have the big advantage of stable funding and freedoms to think about non-fashionable topics, to consider complex arguments, and to pander less to the public or elites.

Aside from more support for polymaths, I’d prefer public intellectuals to focus more on impressing each other, instead of wider publics or elites. And I’d rather they tried to impress each other more with arguments, than with their eliteness and culture references. As for philosophers, I’d rather that they paid less homage to their heritage, and instead more adopted the intellectual styles and habits that are now common across most other disciples. The way polymaths do. I don’t want to cut all differences, but some cuts seem wise.

As to whether any of these groups will effectively call out the exaggerations of the coming era of ideological fervor, I alas have grave doubts.

I wrote this post as my Christmas present to Tyler Cowen; this topic was the closest I could manage to the topic he requested.

GD Star Rating
loading...
Tagged as: ,