LESSWRONG
LW

Comment Permalink

At this point you probably know more about it than I do. The reason I brought N-player PDs up, though, is because I expect that many of them are tough nuts to crack and mere iteration doesn't foster cooperation nearly so easily as in the 2-player version.

See in context

15 Maybe we're not doomed

by Manfred

2nd Aug 2014

2 min read

15

This is prompted by Scott's excellent article, Meditations on Moloch.

I might caricature (grossly unfairly) his post like this:

Map some central problems for humanity onto the tragedy of the commons.
Game theory says we're doomed.

Of course my life is pretty nice right now. But, goes the story, this is just a non-equilibrium starting period. We're inexorably progressing towards a miserable Nash equilibrium, and once we get there we'll be doomed forever. (This forever loses a bit of foreverness if one expects everything to get interrupted by self-improving AI, but let's elide that.)

There are a few ways we might not be doomed. The first and less likely is that people will just decide not to go to their doom, even though it's the Nash equilibrium. To give a totally crazy example, suppose there were two countries playing a game where the first one to launch missiles had a huge advantage. And neither country trusts the other, and there are multiple false alarms - thus pushing the situation to the stable Nash equilibrium of both countries trying to launch first. Except imagine that somehow, through some heroic spasm of insanity, these two countries just decided not to nuke each other. That's the sort of thing it would take.

Of course, people are rarely able to be that insane, so success that way should not be counted on. But on the other hand, if we're doomed forever such events will eventually occur - like a bubble of spontaneous low entropy spawning intelligent life in a steady-state universe.

The second and most already-implemented way is to jump outside the system and change the game to a non-doomed one. If people can't share the commons without defecting, why not portion it up into private property? Or institute government regulations? Or iterate the game to favor tit-for-tat strategies? Each of these changes has costs, but if the wage of the current game is 'doom,' each player has an incentive to change the game.

Scott devotes a sub-argument to why we're still doomed to things be miserable if we solve coordination problems with government:

Incentives for government employees sometimes don't match the needs of the people.
This has costs, and those costs help explain why some things that suck, suck.

I agree with this, but not all governments are equally costly as coordination technologies. Heck, not all governments even are a technology for improving peoples' lives - look at North Korea. My point is that there's no particular reason that costs can't be small, with sufficiently advanced cultural technology.

More interesting to me than government is the idea of iterating a game to to encourage cooperation. In the normal prisoner's dilemma game, the only Nash equilibrium is defect-defect and so the prisoners are doomed. But if you have to play the prisoner's dilemma game repeatedly, with a variety of other players, the best strategy turns out to be a largely cooperative one. This evasion of doom gives every player an incentive to try and replace one-shot dilemmas with iterated ones. Could Scott's post look like this?

Map some central problems for humanity onto the iterated prisoner's dilemma.
Evolutionary game theory says we're not doomed.

In short, I think this idea of "if you know the Nash equilibrium sucks, everyone will help you change the game" is an important one. Though given human irrationality, game-theoretic predictions (whether of eventual doom or non-doom) should be taken less than literally.

Personal Blog

15

New Comment

47 comments, sorted by

top scoring

Click to highlight new comments since: Today at 3:02 PM

[-]ChristianKl11y200

Game theory is not the best way to think about the tragedy of the commons.

Elinor Ostrom got in 2009 the "nobel prize" in economics for her work of studying how people actually deal in the real world with the tragedy of the commons. It makes much more sense to go to her empirically derived work than to think in terms of game theory.

She suggests 8 principles:

Clearly defined boundaries (effective exclusion of external un-entitled parties);

Rules regarding the appropriation and provision of common resources that are adapted to local conditions;

Collective-choice arrangements that allow most resource appropriators to participate in the decision-making process;

Effective monitoring by monitors who are part of or accountable to the appropriators;

A scale of graduated sanctions for resource appropriators who violate community rules;

Mechanisms of conflict resolution that are cheap and of easy access;

Self-determination of the community recognized by higher-level authorities; and

In the case of larger common-pool resources, organization in the form of multiple layers of nested enterprises, with small local CPRs at the base level.

[-]Shmi11y80

But how do these rules emerge and why?

[-]ChristianKl11y70

Empirical study of a lot of cases of the tragedy of the commons and an analysis where people found a way to successfully deal with the issue versus cases where they didn't.

The fact that she actually did study analysed >100 different real world episodes of tragedy was the work that earned her the "nobel prize".

[-][anonymous]11y20

Out of curiosity, why the quotes around "nobel prize"?

[-]ChristianKl11y40

We are talking about the "Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel" and not one of the Nobel Prizes that were established by Alfred Nobel. Alfred Nobel didn't thought that economics was a discipline that needed a prize and his descendants sued over this dispute.

Saying "Sveriges Riksbank Prize in Economic Sciences in Memory of Alfred Nobel" is a bit long so I go for a shorter version and if I would say Sveriges Riksbank Prize nobody would know what I'm talking about. By using the quotes around "nobel prize" I mark that I'm uncomfortable with calling the prize as it were one of the regular nobel prizes but I still have a short phrase that's recognizable by my audience.

[-]A1987dM11y00

IIRC I've seen it referred to as Nobel Memorial Prize, which might be a decent compromise.

[-]ChristianKl11y00

If I remember right that's the Wikipedia consensus compromise solution. I personally prefer the version with quotes I made up myself.

[-]D_Malik11y120

There are a few ways we might not be doomed. The first and less likely is that people will just decide not to go to their doom, even though it's the Nash equilibrium. To give a totally crazy example, suppose there were two countries playing a game where the first one to launch missiles had a huge advantage. And neither country trusts the other, and there are multiple false alarms - thus pushing the situation to the stable Nash equilibrium of both countries trying to launch first. Except imagine that somehow, through some heroic spasm of insanity, these two countries just decided not to nuke each other. That's the sort of thing it would take.

This would be true if the game being played were, say, Prisoner's Dilemma, but in the actual nuclear arms race the game was that if either side launched their weapons the other side would retaliate, resulting in large losses on both sides. (In general, I think LW overuses PD in game theory discussions, when games like Chicken or Stag Hunt would be better.) Wikipedia on mutually assured destruction:

The strategy is a form of Nash equilibrium in which neither side, once armed, has any incentive to initiate a conflict ... neither side will dare to launch a first strike because the other side will launch on warning (also called fail-deadly) or with secondary forces (a second strike), resulting in unacceptable losses for both parties. The payoff of the MAD doctrine is expected to be a tense but stable global peace.

That said, the US didn't pre-emptively obliterate the USSR, or later all the other countries, and this does support your point, since obliterating the USSR would have been positive-expected-utility for the US, for some dubious values of utility. Which may or may not have been the prudent thing to do - we can't judge decisions only on their actual outcomes. Maybe there was a 90% chance of us dying, and we just got lucky, or maybe there are very few people around to discuss this in worlds that weren't lucky. (This is anthropics though, which is confusing.) Incidentally, von Neumann advocated for a pre-emptive strike on the USSR, and later the USSR advocated for a pre-emptive strike on China.

Re your actual point, that cooperation often arises not in spite of but because of self-interest, and that we might be able to cooperate to preserve our value: I agree with you, and Scott seems to agree that it's possible though he considers it unlikely. Our society already does this to some extent - witness the social sanctions imposed on people who achieve enough that their peers look bad by comparison. Essentially we just need to impose enough social costs to make it negative-expected-utility to sacrifice your children to Moloch. (And impose social costs on people who don't impose social costs in accordance with this rule and the preceding rule.) But this might be too hard, especially when there are instabilities, e.g. when the first brain emulations arrive.

[-]Jan_Rzymkowski11y00

PD is not a suitable model for MAD. It would be if a pre-emptive attack on an opponent would guarantee his utter destruction and eliminate a threat. But that's not the case - even in case of a carefully orchestrated attack, there is a great chance of rebuttal. Since military advantage of pre-emptive attack is not preferred over a lack of war, this game doesn't necessarily indicate to defect-defect scenario.

This could probably be better modeled with some form of iterated PD with number of iterations and values of outcomes based on decisions made along the game. Which I guess would be non-linear.

[-]Manfred11y00

Yeah, I was modeling nuclear war as the stag hunt game. Peace is a Nash equilibrium because starting nuclear war is bad even for the aggressor, but it's not as stable under violations of trust, worries about a trembling hand on the big red button, etc.

[-]atucker11y110

The second and most already-implemented way is to jump outside the system and change the game to a non-doomed one. If people can't share the commons without defecting, why not portion it up into private property? Or institute government regulations? Or iterate the game to favor tit-for-tat strategies? Each of these changes has costs, but if the wage of the current game is 'doom,' each player has an incentive to change the game.

This is cooperation. The hard part is in jumping out, and getting the other person to change games with you, not in whether or not better games to play exist.

Moloch has discovered reciprocal altruism since iterated prisoner's dilemmas are a pretty common feature of the environment, but because Moloch creates adaptation-executors rather than utility maximizers, we fail to cooperate across social, spatial, and temporal distance, even if the payoff matrix stays the same.

Even if you have an incentive to switch, you need to notice the incentive before it can get you to change your mind. Since many switches require all the players to cooperate and switch at the same time, it's unlikely that groups will accidentally start playing the better game.

Convincing people that the other game is indeed better is hard when evaluating incentives is difficult. Add too much complexity and it's easy to imagine that you're hiding something. This is hard to get past since moving past it requires trust, in a context where we maybe are correct to distrust people -- i.e. if only lawyers know enough law to write contracts, they should probably add loopholes that lawyers can find, or at least make it complicated enough that only lawyers can understand it, so that you need to continue to hire lawyers to use your contracts. In fact contracts are generally complicated and full of loopholes and basically require lawyers to deal with.

Also, most people don't know about Nash equilibria, economics, game theory, etc., and it would be nice to be able to do things in a world with sub-utopian levels of understanding incentives. Also, trying to explain game theory to people as a substep of getting them to switch to another game runs into the same kind of justified mistrust as the lawyer example -- if they don't know game theory and you're saying that game theory says you're right, and evaluating arguments is costly and noisy, and they don't trust you at the start of the interaction, it's reasonable to distrust you even after the explanation, and not switch games.

[-]buybuydandavis11y80

Incentives for government employees sometimes don't match the needs of the people.

Sometimes?

Financial security, status, power, dominance. The inherent lack of knowledge and accountability of bureaucracies. The Road to Serfdom. The Iron Law of Bureaucracy. Controlling other people's lives , and other people's money. When the problem you're supposed to fix gets worse, you don't get fired, you get a bigger budget.

Thomas Sowell, arch conservative, was a communist when he started graduate school at the University of Chicago. He was a communist after a year under Milton Friedman. He repudiated communism after one summer as an intern at the Department of Labor, and seeing how the government bureaucracy actually worked.

This year, the IRS targeting political opponents. The NSA spying on all citizens. The CIA spying on on the Senate Intelligence Committee's investigation of the CIA. That's pretty sweet, spying on your supposed overseers in their attempts to oversee you. The CIA is also spying on internal whistle blowers when they contact Congress.

And when "called to account" by Congress, they show up, crack jokes, lie through their teeth, and piss in the face of the supposed representatives of the people, while telling them it's raining.

The apparatchiks are in power, and increasingly don't feel the need to even pretend otherwise. My impression of the EU from afar is similar.

[-]Manfred11y160

Clean water, power plants, the clean air act, civil rights, CFCs, disaster relief, fire departments, research in basic science, banning lead in gasoline, interstate highways. Sometimes.

But sure, "the road to serfdom." That reminds me of another government act - the thirteenth amendment.

[-]buybuydandavis11y70

I'm not an anarchist. I never said government can't do good things. The particular issue I addressed was:

Incentives for government employees sometimes don't match the needs of the people.

Which I don't really see you addressing at all.

Your one example of an actual government agency was the fire department, which has many of the characteristics of what the government would best focus on.

Low probability, high risk, emergency situation, with potential for large externalities, on an agreed upon problem (Burning Building Bad) with an agreed upon solution (put out the fire).

Note that people managed to have fire departments without government employees forever, and still do in many places. I've got a friend living in the burbs in New Jersey who is a volunteer firefighter.

That reminds me of another government act - the thirteenth amendment.

When the government banned what it had been previously enforcing. If you're trying to show the unbridled wonders of government power, that really isn't the best example.

The various environmental regulatory regimes are appropriate areas of government action, since they are largely dealing with issues of commons, but even there, there are large problems with regulatory capture and perverse institutional and personal incentives, leading to waste and outright destruction.

Take the ethanol mandates, which have had huge downsides in environmental and humanitarian costs, and no real upside except for the ethanol manufacturers. Though it's probably fairer to attribute that to congressional incentives rather than the apparatchik incentives.

But the apparatchik incentives generally diverge from sensible and consistent cost benefit analysis. Sure, if you spend enough money, you can produce some good.

But at what cost? Thalidomide, thalidomide! Yeah, if the regulatory process is obstructive enough, eventually it will obstruct something that would have been harmful. But what we don't see are the tens of thousands of yearly deaths from the regulatory obstruction.

(And what people should see but don't: a captive population shaken down for the benefit of the entire protected medical/insurance/regulatory industrial complex.)

The FDA just forced 23andme to shut down the medical information they were giving me, tailored to my genome. What's the cost in lives of delaying the genomic revolution in health care 5 to 10 years?

But delay they do, because they deal in power, and exercising that power swells their testicles and benefits the medical industrial mafia.

[-]AlexSchell11y60

The FDA just forced 23andme to shut down the medical information they were giving me, tailored to my genome. What's the cost in lives of delaying the genomic revolution in health care 5 to 10 years?

Interestingly, FDA will have achieved this obstruction by merely adding a trivial inconvenience to personal genomics. A consumer can still get essentially the same health info through free tools (which interpret your data file with info from sites like SNPedia), except they have to spend an extra 1-2 hours on this extra step. Of course, this more than suffices to decimate 23andme's sales.

[-]buybuydandavis11y20

I've been using SNPedia for years.

It's a trivial inconvenience, if you know about SNPedia, and if you're willing to trust them with your information.

I found out about SNPedia after I had purchased my 23andme kit, and poked around with the information they gave. Right now, 23andme gives you your genome, and some ancestry information. Unlikely I would have purchased it in the first place without the promise of associated health care info.

Amusingly, the NIH just gave those criminals at 23andme a grant for a research front end to their system.
http://www.ibtimes.com/23andme-back-dead-thanks-million-dollar-grant-your-genetic-data-1643884

That's the way the apparatchiks roll. Whatever is good for them is ok. Good for you and me, not so important.

Incentives for government employees sometimes don't match the needs of the people.

It often doesn't even match the needs of government employees.

They want the data, and they want to make it impossible for 23andme to collect it.

Of course, the first order of business is always protecting the power of entrenched interests. Can't ever threaten that. Once that is secure, then they'll try to do something with the data.

I see that 23andme is now busy sucking up to the government.
http://www.columbusceo.com/content/stories/apexchange/2014/07/06/fda-setback-alters-course-but-not-mission-for-wojcicki-and-her-23andme.html

I'm sure that once the NSA gets onboard with dna collection, it will be mandatory in schools. In case of abduction, don'tcha know. For the children.

[-]AlexSchell11y20

Good point. My guess is that with some googling, the median 23andme customer would find out about SNPedia and find some tool they trust. Of course, requiring even a tiny bit of open-ended research on the part of the prospective consumer is what constitutes the trivial inconvenience.

Re: 23andme supplicating USG, that seems like a reasonable strategy for advancing personal genomics. Across-the-board liberalization of drugs and medical devices won't happen soon.

[-]ChristianKl11y00

The FDA just forced 23andme to shut down the medical information they were giving me, tailored to my genome. What's the cost in lives of delaying the genomic revolution in health care 5 to 10 years?

It's not clear that the results of 23andMe where much better than chance:

For the seven diseases analyzed by the researchers, only about half of the risk factors provided by 23andme and Navigenics agreed for the five patients. For instance, for lupus and type 2 diabetes, three of the five subjects received conflicting results.

23andme market their product in a way that suggested that the numbers they provided were bunch more informative than they were in practice.

The FDA is tasked with stopped people from selling snake oil and if you get random results on your risk type based on whether you send your DNA to 23andme or it's competitor Navigenics, they are selling snake oil. How badly should 23aneMe be allowed in your opinion to misbehave?

A customer who knows what he's doing can still use the service and use free tools to analyse it's data. People who don't know what they are doing won't anymore be mislead by 23aneMe. To me that doesn't seem like bad regulatory action.

[-]buybuydandavis11y20

It's not clear that the results of 23andMe where much better than chance:

Not clear, to whom?

23andme market their product in a way that suggested that the numbers they provided were bunch more informative than they were in practice.

That's a claim about about the accuracy of the information that 23andme provides, and how that compares to the accuracy of their claims of accuracy. Where's your evidence for that claim?

The FDA is tasked with stopped people from selling snake oil ...

Not by me. If they disappeared into a black hole tomorrow, I would sing "Yippie! Yippie! Yippie!"

And what's their evidence that 23andme is selling snake oil? What's yours?

if you get random results on your risk type based on whether you send your DNA to 23andme or it's competitor Navigenics

I suggest you read Jaynes on probability. The results aren't "random", they're based on different information. You use different information, you assign different probabilities.

How badly should 23aneMe be allowed in your opinion to misbehave?

definition: Misbehaving - offering me information that I want, that those in power don't want me to have, and will use guns to prevent me from having.

I like it when people "misbehave".

People who don't know what they are doing won't anymore be mislead by 23andme.

Yes, people "misled" by their own incompetence about what 23andme says will now be "protected" from such confusion. They, and everyone else, will also be "protected" from the relief of their ignorance of any and all of the accurate information 23andme might have put out as well.

Couldn't they be equally "misled" by anything anyone anywhere ever says to them?

[-]ChristianKl11y-20

And what's their evidence that 23andme is selling snake oil?

From the FDA perspective 23andme has the burden of proof.

From my perspective: If SNP based genetic testing produces useful risk diagnoses that the diagnosis of different SNP based genetic testing companies should agree. As far as I know those people who used multiple of those services get different results from each one.

If you get completely different answer based on the company that you ask about your risk, do you really think that's no issue?

The results aren't "random", they're based on different information.

They are both based on the DNA and there are both based on the published literature. We aren't even sure whether all the published literature is true.

Not by me. If they disappeared into a black hole tomorrow, I would sing "Yippie! Yippie! Yippie!"

A lot of drugs fail in stage 4 trials. Would you really want to have all those drugs on the market?

[-]buybuydandavis11y40

From the FDA perspective 23andme has the burden of proof.

It's nice to have the guns.

If you get completely different answer based on the company that you ask about your risk, do you really think that's no issue?

Discrepant diagnoses are hardly a new issue in medicine.

I'd note that the article you referred to is almost 5 years old now. How much more data does 23andMe have to make accurate predictions? A good order of magnitude at least.

How much more will they have next year? Not so much more, as our benevolent protectors have stopped their data collection efforts.

Yes, it would be nice if medicine got a lot better. They'd make predictions, they'd predict the same things, and they'd be right.

It doesn't get better by preventing data collection efforts. It doesn't get better by preventing prediction efforts. Get the FDA out of the way, and we'd have free GoogleCare in a couple of years for most routine doctor visits.

In the article you link, Ventner mentioned

a call to use more markers that provide information on risk factors for taking medication

This is actually the prime use for 23andMe info right now, and was useful 5 years ago. You get info on your particular genotype for some enzymes, which largely determines how well you metabolize different drugs. Turns out I'm a poor metabolizer for the enzyme that metabolizes a drug I was given years ago, and had a bad reaction to. Whaddya know, this crazy science stuff actually works.

A lot of drugs fail in stage 4 trials. Would you really want to have all those drugs on the market?

Yes. And that's even if I don't get the FDA incinerated in the bargain.

Having the drugs on the market doesn't require anyone to use them, it only allows them to use them. Many drugs with failed trials would likely be very useful for a lot of people. And they still have to get permission from a duly deputized agent of the state to get a prescription. Bah! What do doctors know about medicine? About your health? It's not like they're the all knowing, all seeing FDA apparatchiks.

How charming it is that we have to get permission to try to heal ourselves.

By the way, Stage 4 trials? You're talking Phase 4 monitoring?

[-]ChristianKl11y-20

Get the FDA out of the way, and we'd have free GoogleCare in a couple of years for most routine doctor visits.

I'm not aware that the FDA was responsible for Google ending Google Health.

By the way, Stage 4 trials? You're talking Phase 4 monitoring?

Sorry I was mentally starting to count with 1 and not 0.

Having the drugs on the market doesn't require anyone to use them, it only allows them to use them.

The question would be whether the companies would still run the trials if they weren't legally mandated. Trials are expensive and if the government wouldn't force big pharma to pay for those trials, do you think Big Pharma would pay?

[-]buybuydandavis11y30

I'm not aware that the FDA was responsible for Google ending Google Health.

This seems like and debating point more than a good faith evaluation of the situation.

Really, it's a mystery to you that Google would want to create a health service, and that the regulatory environment would have a significant effect on their ability and desire to create such a service?

Google is in the business of getting eyeballs, largely by aggregating data to be able to make better predictions than other people can.

People's records are already stored, and they flow through the medical system. The value add from the point of view of most patients view comes from making diagnoses and recommendations. Exactly what the FDA is prohibiting 23andme from doing.

Of course there would be much greater value add if we were able to act on those recommendations, and purchase medications, tests, treatments, and devices. You know, what would happen if we were free.

Google said they closed it for lack of demand.

Do you find the link between what functionality they're allowed to provide and the associated demand for that functionality mysterious?

The question would be whether the companies would still run the trials if they weren't legally mandated.

"The" trials?

They almost certainly wouldn't run trials in the form the FDA requires today. They'd still want to convince people to buy their products. They'd still want to avoid getting sued. People would still want convincing.

We want to be well, we want to find medicines that will make us well, and companies want to sell us stuff, but you can't imagine us arranging that without threats of violence from the government beyond enforcing contracts?

Presumably, then, we couldn't get together on this list and find a solution to some shared medical problem without the government stepping in and threatening us all with violence?

[-]ChristianKl11y-10

People's records are already stored, and they flow through the medical system. The value add from the point of view of most patients view comes from making diagnoses and recommendations. Exactly what the FDA is prohibiting 23andme from doing.

The FDA seems to say that 23andMe is supposed to provide some evidence that the diagnosis and recommendations they make are better than rolling a dice. The FDA also gave them years to do that it didn't shut them down immediately to give them some time to actually provide evidence.

As far as the timing goes Google Health got shut down before the FDA got 23andMe to stop giving diagnoses.

They almost certainly wouldn't run trials in the form the FDA requires today. They'd still want to convince people to buy their products. They'd still want to avoid getting sued. People would still want convincing.

There are many ways to convince people to buy products that don't have much to do with running expensive medical trials. You can bring the doctor to a fancy all expensive paid vacation where he get's lectured on all the great benefits.

They'd still want to avoid getting sued.

I rather have a group of experts in a government department examine whether or not a drug does what is claimed than push that task to courts and let juries make the corresponding decisions.

Do you think this case is handled better if someone sues 23andMe in court because they got cancer and 23andMe told them they had a low cancer risk? It's somehow less violence if the whole process goes through a jury than when it's government regulators doing their job?

I don't think the libertarian position of passing all issues to the courts is a wise move. Having regulations where a diagnostic company can tell a jury: "We follow the rules laid out by government agency X" is valuable. Otherwise the jury has to make up rules on the spot.

[-]buybuydandavis11y10

The FDA seems to say that 23andMe is supposed to provide some evidence...

No, that's not what they say. They're not asking for "some evidence", they're asking 23andme to jump through their hoops. Dance for me, peasant.

23andme provides plenty of evidence in terms of peer reviewed papers as citations to their customers. Evidence isn't lacking, freedom is.

As far as the timing goes Google Health got shut down before the FDA got 23andMe to stop giving diagnoses.

This strikes me as yet another dishonest debating point. Are you unaware that the regulatory regime controlling the right to "practice medicine", and thereby the right to give a diagnosis, predated the latest FDA attack on 23andme and Google Health by decades?

You can bring the doctor to a fancy all expensive paid vacation where he get's lectured on all the great benefits.

How horrible, allowing doctors to hear information. We're much better off having the Ministry of Truth vetting all communications on medical information.

I rather have a group of experts in a government department...

That is indeed the bottom line for most all these questions.

Ronald Reagan

This is the issue of this election. Whether we believe in our capacity for self government or whether we abandon the American revolution and confess that a little intellectual elite, in a far distant capital can plan our lives for us better than we can plan them ourselves.

I'd note for people lacking the historical context, that self government meant what it said - governing your self.

I'm aware that the view expressed here is a minority one, certainly in the world, and even in the US. You, and people who think like you on this question have the numbers, and have the guns.

It's somehow less violence if the whole process goes through a jury than when it's government regulators doing their job?

Enormously less.

Most market transactions are perfectly satisfactory to the market participants involved. Free people exchange because they perceive mutual benefit in the exchange.

I'd prefer threats of violence to be the exception, and not the rule. I'd prefer to be free to choose. I understand your preference goes the other way.

[-]ChristianKl11y20

This strikes me as yet another dishonest debating point. Are you unaware that the regulatory regime controlling the right to "practice medicine", and thereby the right to give a diagnosis, predated the latest FDA attack on 23andme and Google Health by decades?

Look, don't make me out as a defender of the status quo. I did sit down with someone making government regulations in this space because of my QS-media credentials and told him that we QS folks don't really need his help.

Plenty of people do manage to create products that don't get shut down by the FDA.

23andme provides plenty of evidence in terms of peer reviewed papers as citations to their customers.

Other testing companies seem to provide nearly the same papers as evidence for their diagnosis and then a different diagnosis. That suggest that the process has issues.

Unfortunately the full exchange between 23andme and the FDA isn't public so we don't know what they FDA specifically asked from 23andme. To me it's not clear that they asked something that's completely unreasonable in the sense that asking the same thing of Google Health would prevent it from working.

How horrible, allowing doctors to hear information. We're much better off having the Ministry of Truth vetting all communications on medical information.

Do you believe that pharma representatives bribing doctors should be completely legal?

Crazy people like me would prefer threats of violence to be the exception, and not the rule. I understand your preference goes the other way.

If I make a product than I want to know whether not I have to expect violence. I want clear rules so that I don't have to explain myself to some jury who's driven by emotions.

Do you think that doctors being forced to practice defensive medicine to avoid lawsuits is no issue or somehow an absence of violence? Mob rule is a form of violence and it's not efficient.

[-]Azathoth12311y20

Other testing companies seem to provide nearly the same papers as evidence for their diagnosis and then a different diagnosis. That suggest that the process has issues.

Given how common it is for people to get different diagnoses after going to different doctors, by that logic it seems the FDA should shut down nearly every doctor.

[-]Lumifer11y00

Unfortunately the full exchange between 23andme and the FDA isn't public so we don't know what they FDA specifically asked from 23andme.

You know that you could have checked Google, right? X-/

The letter.

[-]ChristianKl11y00

There were 100 of emails going back and forth and dozens of face to face meetings over five years. Then the FDA said concluded that this isn't working and sent that letter.

Reading the letter it's interesting that the letter doesn't even forbid 23andMe from diagnosing people. It forbids them from marketing that they are diagnosing people.

The letter for example says:

The Office of In Vitro Diagnostics and Radiological Health (OIR) has a long history of working with companies to help them come into compliance with the FD&C Act. Since July of 2009, we have been diligently working to help you comply with regulatory requirements regarding safety and effectiveness and obtain marketing authorization for your PGS device.

[...]

You have not worked with us toward de novo classification, did not provide the additional information we requested necessary to complete review of your 510(k)s, and FDA has not received any communication from 23andMe since May. Instead, we have become aware that you have initiated new marketing campaigns, including television commercials that, together with an increasing list of indications, show that you plan to expand the PGS’s uses and consumer base without obtaining marketing authorization from FDA.

If I want to judge whether or not the FDA is reasonable in this instance than I don't have enough data.

[-]Lumifer11y00

It's not clear that the results of 23andMe where much better than chance:

It is, actually, quite clear. Chance doesn't have much to do with different weightings and different baselines. Your link, e.g., says:

The Venter study notes that, in some cases, companies define the average population disease risk differently. “Navigenics distinguishes population disease risk between men and women (for example, men are more likely to have heart attacks than women), whereas 23andMe primarily takes into account age (for example, incidence of rheumatoid arthritis increases with age),”

Also, have you looked at studies examining how much different doctors agree with each other about same patients? Hint: not much. Here is one such study.

[-]buybuydandavis11y20

Also, have you looked at studies examining how much different doctors agree with each other about same patients

But they're a powerful entrenched interest, therefore their mistakes are a-ok.

I used to work for a company that made a machine that scanned pap smear slides for cancer, so I saw some of the data you describe.

In this case, the problem wasn't as much in the inter observer rates as the false negative rates. A significant fraction of slides with abnormalities, as detected by exhaustive examination by experts, were missed by the large commercial processing centers.

Ladies, have your pap smears done at clinics associated with med schools. I think their false negative rate was approaching an order of magnitude better than HMO Lab Inc.

Our magic machine, of course, was much better than HMO Lab Inc. But the FDA "protected" women from more accurate pap smear diagnoses by a post data look at one rare diagnostic category with a handful of samples, where the detection rate was poor. Horrible decision theory, and a failure to follow even the flawed hypothesis testing that they purport to use.

But they've got the guns, so fuck us.

[-]torekp11y70

For 'prisoner's dilemma', substitute 'N-person prisoners' dilemma' and see how the terrain looks. In appropriate places of course. But also, on a more optimistic note, beware over use of the PD model. (Edited to add link)

[-]Manfred11y10

Huh, maybe we should hold a tournament for iterated multi-person PD rules. The most interesting paper I found was this one and it just scratches the surface. Do you know if anyone's already done it?

It seems like cooperation should still work well, but I'm not sure how well different enforcement mechanisms would work, which might lead to a sizeable niche for defectors.

[-]torekp11y20

[-]pianoforte61111y60

Scott's post proves too much. In particular it proves that early multicellular organisms would inevitably revert back to micro-organisms.

Edit: This is a bit too opaque. Scott argues that agents that sacrifice the common good for individual gains in reproduction and survival sill survival will outcompete agents that do not make these sacrifices. However, cells within early multicellular organisms could have done this, yet multicellular organisms nevertheless exist.

[-]Viliam_Bur11y100

I understood it to mean that multicellular organisms don't maximize the values of their cells, if we would anthropomorphize them.

In other words, in this scenario you are the (potential) cancer. And the system evolves in a way that you either lose your battles (most of the time), or your victory against the system would result in your death anyway. (Because your survival depends on the system in too many ways.)

[-]roystgnr11y90

Look at example 9: cancer. That basically is a multicellular organism reverting back to micro-organisms. And in some sense it's as depressing as Scott says: billions of years, and evolution still hasn't figured out an immune system that can reliably put an end to that garbage?

In another sense it's outright encouraging: billions of years of animals afflicted by cancer, and here we still are, and that wasn't just good luck. The winning strategy seems to be "group selection", where we think of organisms as a group of cooperating cells. Cancer may eventually take my group of cells down, but if by the time that happens there's two or three independent groups running around preserving many of the same values (in this case genes, memes, etc) as my group had, those values live on.

In a third sense it's much more depressing. Scott's only identified solution category is to prevent runaway competition from destroying our values by governing all that competition with a singleton: a single agent or coordinated group so powerful as to be able to squelch any outsider or subgroup's attempt to subvert human values, even if that villain is specifically making itself as effective as possible by sacrificing those values on the altar of efficiency. But in the cancer analogy, this is the sort of idea that would have ended up with the death of all multicellular life: even if a single megaorganism might be vastly better than run-of-the-mill organisms at defeating cancer (better immune systems via economies of scale, less energy wasted in zero-sum games against competing organisms, whatever), it only needs to lose once and it's lost for all time. Back outside the analogy, either Scott makes his singleton perfectly stable, or it gets replaced by whatever manages to subvert it, and now whatever managed to subvert it is a singleton. FAI slips up and we get a universe tiled with smiley faces. Archipelago slips up and we get a boot stamping on a human face forever.

[-][anonymous]11y-10

Very good points:

Cancer may eventually take my group of cells down, but if by the time that happens there's two or three independent groups running around preserving many of the same values (in this case genes, memes, etc) as my group had, those values live on.

This opens up a different means of stopping societal cancers - closed borders. Prevent the infections from spreading to all of civilization.

[This comment is no longer endorsed by its author]Reply

[-][anonymous]11y20

Additionally, if the history of life on Earth should show you anything its that nothing ever 'wins'.

[-]roystgnr11y30

This is true, at least loosely speaking. I'm not sure why you were downvoted. Inducting from "nothing has ever permanently won" to "nothing ever will" would be overconfident, but noticing that nothing has ever permanently won and examining the reasons why might be very instructive.

[-]Pentashagon11y00

Even bacteria? The specific genome that caused the black death is potentially extinct but Yersinia pestis is still around. Divine agents of Moloch if I ever saw one.

[-][anonymous]11y20

Its primary hosts are doing great. And it's got nothing on bacillus subtilis or whatever that cyanobacteria with hundreds of billions per cubic meter of seawater is. And even those haven't 'won' in the sense that sometimes gets discussed around here. They're one form among many. Even bacteria are not the main primary producers in all environments - the land-plants take that up over a third of the earth's surface (in a constantly shifting ecological arrangement with other things).

[-]Shmi11y50

I suspect that there is some mathematical formalism which describes how the tragedy of the commons turns into an iterated PD, where virtues such as fairness and honesty emerge naturally, but I cannot find anything relevant online.

[-]Manfred11y20

Well, I was just sort of making stuff up, so I'm not surprised there no literature :P

But what was going through my head was that a one-shot game can be turned into an iterated game just by breaking the decisions and payoff into pieces and setting it up so that people share information at regular intervals unless they pay a prohibitive cost. Depending on the real-world situation, this may just be a different framing of the same thing - e.g. the rats on a Malthusian island.

Many players vs. two players seems like the bigger gulf between tragedy of the commons and iterated PD (which I'm just bringing up because we understand well why cooperation arises in it). Worst case the players would have to set up a government-like body to effect the right payoffs, and then there's no point to trying to be clever with getting people to play tit-for-tat. I'd guess that there are still good cooperative / reciprocal strategies in iterated multiplayer PD, but I'd need to see a tournament and have someone smart figure it out for me :)

[-]Cicero11y-30

Does anyone have any good recommendations for a book on the subject of: Quantum Mechanics/Physics discoveries and what it tells us about Spirituality, if anything. I am curious to learn more on how these subjects could potentially be connected. Please, does anyone know of a good book you'd recommend? Thank you a lot.

[-]Manfred11y10

1) This sort of question would be better asked in the open thread. You can find the latest one by going to the discussion section, then looking for "latest open thread" on the right sidebar.

2) Personally, I can't recommend you any such books, since all the attempts I've seen to connect spirituality and quantum mechanics have been very dubious.

[-]Richard_Kennaway11y10

My recommendation would be to avoid any book claiming to draw any connection between them at all. It will contain nothing of substance about either.

Moderation Log

Curated and popular this week

47Comments