Scott Alexander wrote in Book Review: History of the Fabian Society about how the Fabian society had eight political thesis that they agreed on.

I want to propose eight points that I think we can agree as a rationality movement about:

1. We want public policy that's backed up by empiric evidence. We want a government that runs controlled trials to find out what policies work.

2. We want that government bureaucrats who make predictions about the future makes them in a way where the accuracy of their predictions gets measured.

3. We want a policy that funds AI safety research and that's a bit cautious about AI.

4. We want government money to go into anti-aging research of the kind that SENS does.

5. We want less bureaucracy around running scientific studies so that nightmares don't have to happen

6. We want the right not to have to endure what amounts to torture at end of our lives

7. We want to end the War on Drugs

8. We want that the FDA switching from using p-values to using better statistical models

Do you think one of those points isn't consensus in our community? Do you think there are other political demands where we have consensus?

(Please avoid talking about political issues where you don't think there's consensus in our community in this thread)

New to LessWrong?

New Comment
35 comments, sorted by Click to highlight new comments since: Today at 8:15 PM

Which drugs? As a baseline, "anything not obviously worse than tobacco cigarettes" might as well be legal. On the other hand, it seems to me that things like methamphetamine and many opioids aren't things that you should just be able to walk into a store and buy...

As worded the point doesn't call for legalization. Countries like Germany also ban a lot of drugs but don't have a war on drugs that incarcerates massive numbers of people.

Concretely, the war on drugs is a policy of the federal government in the US. I would count getting the federal government out of the topic and leaving the states to do what they currently do as ending the war.

Additionally, "anything not obviously worse than tobacco cigarettes" would be a decent standard for legalization.

Incidentally, I would love to be able to ban tobacco cigarettes if it were at all practical. Are e-cigarettes a good enough substitute that banning them wouldn't create a huge black market?

Not IME.

Incidentally, tobacco products aren't an unqualified vice the way alcohol is sometimes argued to be. (I also disagree with that assessment WRT alcohol, but the benefits are smaller and the harms are larger there than they are with tobacco.) They're better seen as general-purpose OTC psych meds -- they're surprisingly good at ameliorating a wide variety of flavors of things being a bit shit -- that have the unfortunate side-effect of dramatically increasing the likelihood that you get cancer.

Absent alternatives, this is probably a worthwhile tradeoff for many people, most of whom are not upper-middle-class sorts who, if neuroatypical, are so in upper-middle-class ways, because things are much less likely to be a bit shit for said sorts; so those sorts (who, if neuroatypical, are so in etc.) keep failing to pick up on this and deciding cigarettes should be banned.

Taxing cigarettes already create a black market in cigarettes today.

I would prefer many of these things to the current state. I do not identify with political movements, and my rationalism is not about others.

I am a rationalist because I want to understand the universe. Separately, I may have political, personal, and social goals that align with others who study and think about truth and knowledge. They are not the same.

Wondering how controversial this would be:

Copyright terms on software should be closer to 30 years than to 100 years.

#8 seems oddly specific compared to the others (why just the FDA? are p-values really the main problem?), and #7 and #8 are both about the US which feels a bit weird for an international movement.

Given the United States is the primary country among rationalists; successes in moving the dial on any policy in that country as one of the biggest, richest and most powerful in the world based on EV grounds, even given the small chance of success, would be a big win for rationality and the proof of concept for mass sanity waterline-raising; practice in policy optimization among rationalists everywhere on behalf of rationalists in one country will teach us lessons transferable to other countries or domains of action; and how American rationalists can benefit from the resource of extra cognitive power as the community focuses on one country at a time, I think the worldwide community endorsing a country-specific policy platform now is consistent with the community's long-term eye towards be(com)ing a more global movement.

Lots of other countries rely indirectly on the FDA for deciding what medical treatments to allow...

I don't know if I count as part of the "movement", but I can't agree on these demands, because they all assume that notions such as "public policy" and "government" are valid and legitimate.

Suppose we were to turn them round and write them as negative demands; 5, 6, and 8 all reduce to freedom of contract. 7 is covered by "no crimes, only torts" since the concept of a "victimless tort" is obviously meaningless. 1 and 2 are fundamentally just questions of governance procedure, and become a lot less odious when you're not forced to live under / pay for a policy you don't support chosen by broken systems of governance. 3 and 4 are demands that other people's money be spent, and allowing such things is a gift to Moloch (see also: Olsonian scramble, rent-seeking); if we rule that out, we're left with "we want to be able to give our money to $CAUSE instead of having it taxed to fund causes we don't support" (causes like #7, come to think of it).

So rather than 8 demands, one principle would seem to suffice: "The support of a plurality of irrational humans is no license to trespass upon individual liberties". Then again, nothing provides such a license; a superintelligent AGI doesn't have the ethical right to do so either. Both the elected government and the AGI, in practice, have the power to do so (in the former case it's because most people believe democracy legitimises government and will support it), but if might makes right then we may as well not bother with AI alignment research ;-)

#6 is really "we want legal euthanasia" right? Might as well say it like it is.

I think legal prostitution belongs on the list as well.

And maybe an end to tax advantages for churches? Because that's direct state funding for irrationality.

Upvoted for the suggestion to reword the euthanasia point.

1. We want public policy that's backed up by empiric evidence. We want a government that runs controlled trials to find out what policies work.

This seems either empty (because no policy has zero empirical backing), throttling (because you can't possibly have an adequate controlled trial on every proposal), or pointless (because most political disputes are not good-faith disagreements over empirical support).

Second, as this list seems specific to one country, I wonder how rationalists who don't follow its politics can inform this consensus.

Third, did you choose eight demands only to mimic the Fabians? Does that mean you omitted some other plausible demands, or that you stretched a few that perhaps should not have made the cut?

The public gets the government it deserves and demands. Demanding better government is effectively a demand for a better selectorate. These are all results, not causes.


If you choose to spend time on politics, I'd advise https://www.lesswrong.com/posts/XqmjdBKa4ZaXJtNmf/raising-the-sanity-waterline over pushing for fine-tuning of a busted system.

Does demand #1 naturally lead to demands #2 - #8? If the other demands are shown to lead to better social outcomes then they may be redundant. To what degree are we certain that points #2 - #8 lead to better social outcomes?

Not everything that leads to better outcomes has evidence that proves that it leads to better outcomes. More importantly an issue such as whether or not we allow legal euthanasia is a question of values.

How much support is there for promotion of prediction markets? I see three levels:

1. Legalization of real-money markets (they are legal in some places, but their illegality or legal dubiousness in the US--combined with the centrality of the US companies in global finance--makes it hard to run a big one without US permission)

2. Subsidies for real-money markets in policy-relevant issues, as advocated by Robin Hanson

3. Use of prediction markets to determine policy (futarchy), as envisioned by Robin Hanson

I don't think the rationality community should become a political movement; there are ways to achieve things outside the government. I think putting controversial items like euthanasia and ending the war on drugs on the list risks alienating more conservative rationalists.

I agree, but what ways to achieve things outside the government do you support for the rationality movement ?

::insert something about anthropogenic global warming here::

Are there any political parties anywhere in the world, other than the US Republican Party, that endorse climate change denial?

Most of the debate about climate change is about whether or not we support policies to reduce emissions. If you believe in short AGI timelines and the ability of an AGI to simply fix the issue, it makes sense to oppose those policies because they don't matter in the relevant timeframe.

2. We want that government bureaucrats who make predictions about the future makes them in a way where the accuracy of their predictions gets measured.

This sounds like it might be a good idea in principle but it might not be a good idea in practice, depending on exactly how the new system ends up incentivizing the bureaucrats. E.g. if their accuracy just gets measured but nobody pays attention to the measurement then this is useless; or if they can somehow Goodhart the metric then that could cause worse outcomes than having no metric.

We want that government bureaucrats who make predictions about the future makes them in a way where the accuracy of their predictions gets measured.

How about having decisions made by people with skin in the game?

  1. Is not as simple as you think. Evidence is hard. If the evidence were easy and simple it would already be in use.

A couple of years ago, former Obama and Bush officials estimated that only 1 percent of government spending is backed by any evidence at all. 1 percent. Perhaps unsurprisingly, then, evaluations of government-sponsored school and work programs have found that some three-quarters of those have no effect.

Source

Evidence requires running good studies. Running controlled studies requires treating some people differently then other people based on arbitrary criteria. Treating everybody the same in front of the law is a basic principle in many countries which discourages the running of controlled studies.

We want the right not to have to endure what amounts to torture at end of our lives

Whatever happened to "live forever or die trying"?

1. Don't we already? At least a little bit. There are various studies about, e.g. the effects of minimum wage. Maybe you want more, but at some point the trials might be problematic. I think we should leave testing to academia (and it's fine to want it to do that testing "better", but it's not quite clear what that means).

2. I'm not sure I care that much about how well the politicians are calibrated. Maybe you're talking about promises? But then I'm also not sure whether that's a good idea or a naive bandaid.

3. "A bit", sure, but I'm not sure we all agree on how much "a bit" is, or what that policy would even look like.

4. I know nothing about the quality of SENS research, and I would question how funding for aging should compare to other diseases, but, sure, why not.

5. Sure, but we do want some bureaucracy, and it's not quite clear how much.

6. This is weirdly worded, it might refer to a few things, but, sure, some of those are ok by me.

7. This is weirdly specific and weirdly uncontroversial. I haven't seen anyone say "hey, war on drugs is great!" for a while, maybe I'm poorly informed. Though, as always, the specific actions of ending that war are less clear.

8. I don't think anyone would object to doing anything "better". I don't really know much about it, but this point does smell like a naive bandaid a little.

To sum up, most of these are only consensus if you don't think too much about them. To be clear, they mostly feel right. It's only that many are either too vague to pass for policy, or it's not clear that they would have the intended effect.

Also, they get increasingly specific to your local context, as you go down the list. Is that intentional?

At least a little bit. There are various studies about, e.g. the effects of minimum wage.

There are no controlled studies about the subject and the studies that exist have relatively little effect on policy making.

I think we should leave testing to academia

It's not possible to run a business like Facebook without running A/B tests in today's landscape. Even if you are ideologically opposed to it leaving the testing to academia is not an option given how useful it's to private enterprise.

Currently, you have a principle according to which every citizen is supposed to be equally affected by a given policy with discourages testing of policies the way Facebook doesn't have a principle that every user can use Facebook equally.

I'm not sure I care that much about how well the politicians are calibrated. Maybe you're talking about promises?

I'm talking about government bureaucrats and not about politicians. Various government bureaucrats are payed to predict things about the future from changing in the rate of inflation to the amount of people who will leave in ten years in a given city.

This is weirdly specific and weirdly uncontroversial.

The current US government has a policy of stepping up the war on drugs.

Also, they get increasingly specific to your local context, as you go down the list. Is that intentional?

Given that I don't live in the US, I don't see how it's specific to my local context. In general I focus the points on agreements and after writing the first down I didn't think of broad points.

Regarding 1, I remain unconvinced. Nobody here is against understanding of policy effects better, but I think you're underestimating the downsides and overestimating the usefulness of such tests. Regardless, I'm saying that we don't have a consensus, I'm not sure if trying to develop one is within the scope of your post.

I'm talking about government bureaucrats and not about politicians.

Ok, but if they're not elected politicians, then why do I need to know their accuracy? I entirely support the idea that bureaucrats should be hired to be effective, but it seems dubious if measuring predictions is the right solution.

The current US government has a policy of stepping up the war on drugs.

I'm aware to some extent. But I seem to remember him campaigning on ending that war, same as everyone else (maybe "end" meant "win" in his mind)? Anyway, to the extent that it is worth saying, I agree with the position, although, like all other points, it's lacking in details.

I don't live in the US

Well, then I'm just confused. Does your country have the same political problems as US? Or did you just tailor your suggestions to US, since most users are presumably from there? I'm not from US either, so talking to you about what the FDA should do is weird.

Ok, but if they're not elected politicians, then why do I need to know their accuracy?

That's not what I argued. I do understand that this might be the first time many people thought of this and it might take more explanation to rally people around it.

I agree with the position, although, like all other points, it's lacking in details.

The goals of the Fabians also hadn't much details. That's not the point.

That's not what I argued. I do understand that this might be the first time many people thought of this and it might take more explanation to rally people around it.

On one hand, if it's the first time many people have heard of this, then it is definitely not consensus. On the other hand, the idea that we should monitor how well various workers perform they jobs (including when their jobs involve making predictions), is not exactly novel. Ultimately though, I don't know what your idea is, you didn't explain for what purpose the predictions should be measured, I've only make a few guesses of my own.

The goals of the Fabians also hadn't much details. That's not the point.

If you don't need the demands to be detailed, then I propose "we want more good things, and fewer bad things" - everyone can support that. Seriously though, I don't know what the value of demands without details is. When you lack them, different people can fill them in very differently, and a lot of consensus is going to be fake.

On the other hand, there might be some other goal. I haven't read Scott's post yet (I expect many people haven't, I'd encourage you to add some explanation to your post, why having such a list would be good).

I think LW exists to be a place for complex arguments and there's no need to make arguments shorter.

I think that referring to other popular writing is okay and there's no reason to explain things anew. It gives people who want to engage with the argument the opportunity to do so.

I think that referring to other popular writing is okay and there's no reason to explain things anew. It gives people who want to engage with the argument the opportunity to do so.

Yes, it's ok. But I've read Scott's book review now, and I still don't know why you think having such a list is good, or why it's ok for the points to be woefully vague. In general, it's good to explain things.

I think LW exists to be a place for complex arguments and there's no need to make arguments shorter.

I'm not parsing this. What are you replying to? Did I say something should be shorter? In fact, I'm saying that your post is too short.