Leopold Aschenbrenner just wrote a really interesting blog post about Burkean Longtermism.

A heuristic for distinguishing Burkean longtermism from EA style longterism maybe this: Could you get a longtermist from a 100 years ago to sympathize with your priorities? You could probably convince such a person that we must increase economic growth, sustain the population level, protect the environment, and fight wars, plagues, and tyrannies. But you would have a hard time convincing him that artificial intelligence was the greatest risk to ever face humanity.

Historical maturity

Leopold writes:

the Burkean longtermist reverts to a historicist perspective in thinking about the future. We should prize those forces that have been robustly good over long timespans in the past, like economic growth. And we should guard against those threats that have bedeviled humanity for time immemorial, like wars and plagues and tyranny. “In history a great volume is unrolled for our instructions, drawing the materials of future wisdom from the past errors and infirmities of mankind” (Burke, p. 157).

The concept of mathematical maturity comes up a lot in my computer science classes. Many CS subjects like algorithms and complexity theory won’t make sense to you unless you come into them with some preliminary mathematical intuitions and context.

I wonder if we longtermists lack historical maturity. To talk about the long term - to consider the fate of empires, the consequences of technologies, and the trajectory of progress - without this maturity will be as frustrating and unenlightening an experience as walking into Scott Aaronson’s quantum computing class without really understanding linear algebra.

The longermists of history had a far greater understanding of their past than I have of my own. For example, here is a portion of the great books Thomas Jefferson (a longtermist by any meaningful definition) recommended to his young nephew:

For the present, I advise you to begin a course of antient history, reading every thing in the original and not in translations. First read Goldsmith’s history of Greece. This will give you a digested view of that field. Then take up antient history in the detail, reading the following books, in the following order: Herodotus, Thucydides, Xenophontis Hellenica, Xenophontis Anabasis, Arrian, Quintus Curtius, Diodorus Siculus, Justin. This shall form the first stage of your historical reading, and is all I need mention to you now. The next, will be of Roman history (*). From that, we will come down to modern history. In Greek and Latin poetry, you have read or will read at school, Virgil, Terence, Horace, Anacreon, Theocritus, Homer, Euripides, Sophocles. Read also Milton’s Paradise Lost, Shakspeare, Ossian, Pope’s and Swift’s works, in order to form your style in your own language. In morality, read Epictetus, Xenophontis Memorabilia, Plato’s Socratic dialogues, Cicero’s philosophies, Antoninus, and Seneca….

Or read about the self-educations of people like Churchill and Napoleon (who were, if not consciously longtermists, certainly influential into the long term):

Churchill’s reading programme began with Edward Gibbon’s 4,000-page The Decline and Fall of the Roman Empire – which he was to re-read twice more over the course of his life, and parts of which he could quote from memory. He followed it with Gibbon’s autobiography and then read Macaulay’s six-volume History of England, which he loved (except for the attacks on the 1st Duke of Marlborough) and the Lays of Ancient Rome. After that he read Jowett’s translation of Plato’s Republic, and the key texts of Schopenhauer, Malthus, Darwin, Adam Smith, Henry Hallam, Samuel Laing, William Lecky, the Marquis de Rochefort and very many others

Andrew Roberts, Churchill: Walking with Destiny

It would be one thing if only I were ignorant of the facts and trends of history but other longtermists could make up for my deficiencies. But so far as I can tell, very few longtermists have integrated into their worldview the lessons of the bronze age collapse or the impact of the printing press. I have heard us debate about whether Bayes or Popper explains how an AGI would work, but not whether famine or plague was proximally responsible for the rapid depopulation of the late Roman empire.

Perhaps it is because we lack an understanding of the problems the ancients faced that we assume we live in a uniquely dangerous and promising century. More dangerous than when mankind numbered less than 10,000, more promising than when two halves of globe were at long last reunited.

Our lack of context gives us this Whig view of history where the human story only began in earnest about 200 years ago. It has since been a tale of accelerating ethical and scientific improvement which has placed us conveniently in the position of absolute moral superiority. We have thus been freed from quaint problems like foreign invasion, societal collapse, scientific stagnation, and technological decay.

The past is not some prelude which explains the present without resembling it. Our problems are new in the same sense that the iPhone 12 is new - better than previous iterations but not fundamentally original.

Radical inside-view schemes

In Zero to One, Peter Thiel identifies two axes along which ideas can be placed: definite vs indefinite and optimistic vs pessimistic. Here’s where I would place some prominent ideas from the longtermism community.

• Indefinite pessimism: Bostrom’s Vulnerable World Hypothesis

• Definite pessimism: Yudowsky and Bostrom’s concerns about AI

• Indefinite optimism: Ben Franklin’s long term charitable fund

• Definite optimism: Elon Musk building successful companies whose ong term plan is to colonize the solar system and revolutionize the electric grid

Burkean longtermism falls into the indefinite pessimism quadrant. A Burkean, in the political sense of the word, is someone who recognizes that our civilization has been developed by unthinking cultural evolution. We impose our rational schemes on it at our peril. A Burkean thinks that most changes, especially those motivated by appealing abstractions, are likely to be harmful. This pessimism is not motivated by a lengthy argument about how any one specific movement or technology is worrying (such as Yudowsky on AI). Rather, it is a general sensibility about changes motivated by big new ideas.

As Leopold writes:

We should be skeptical of any radical inside-view schemes to positively steer the long-run future, given the froth of uncertainty about the consequences of our actions. A “longtermist” in the Middle Ages might have been convinced that the best thing to do was to violently spread fundamentalist Christianity. So too will future generations think of whatever plans we have hatched.

This is odd, since elsewhere Leopold has argued that a big problem with Effective Altruism is that it doesn’t fund or encourage enough ambitious projects.

Elon Musk is arguably the most effective longtermist alive. But my Effective Altruist friends don’t seem to think that starting an innovative longtermist company is the most important thing they could be doing with their life. Why is that? Perhaps because Musk’s entire raison d’etre is “radical inside-view schemes to positively steer the long-run future.” What happens if we tell a generation of smart, ambitious, and altruistic kids to be vary of making definite plans which seem improbable from the outside? I doubt we produce lots of people like Elon Musk and Jeff Bezos who are EA aligned.

You might say that’s okay, since these people are outliers anyways. We want to make the average effective altruist more effective, not optimize for that 0.01% who can start space companies. This sounds a lot like saying that it’s okay that my investment advice would have told you to sell all your Apple stock in 2000. On the other stocks, I would have slightly beat the market.

I agree there is a risk that being so committed to your scheme that you don’t notice the harm it’s doing. But frankly, EAs don’t seem like the kind of crowd we need to keep reminding not to start a violent conversion campaign. It seems to me that we would benefit if the marginal effective altruist had more confidence in his grand plan to remake a sector of the economy. Let’s not keep telling the depressed patient about the dangers of toxic positivity.

Yes, the crusades and the French Revolution were radical inside-view schemes, but so were the Constitutional Convention and the post war Bretton Woods system. Longtermism itself is a radical inside-view scheme.

We should be more open to improbable crazy sounding plans which don’t undermine economic growth, human rights, and the possibility of criticism. This excludes plans like, Rob everyone and donate the proceeds to GiveWell, but not schemes like, Start a nuclear fusion company and see if you can crack the technical challenges.

This criterion also excludes Nick Bostrom’s suggestion that we might need a global panopticon to stop the development of destructive technologies. Such a proposal is probably the kind of radical inside-view scheme which Leopold is criticizing anyways, not only because it undermines rights, but because it seeks to monitor and control divergent groups. Leopold writes:

Given epistemic humility, I see our world’s increasing cultural homogeneity with great worry. We must preserve the ability for our views and understandings to evolve. For example, we should attach great import to preserving American-style religious liberties: in no other country can religious groups live in “parallel societies,” experimenting with truly different ways of life. Even “liberal” places in Europe make it an aim of state policy to crack down on “parallel societies,” e.g. forcing everyone to go to secular public school; the forces of modernity are too overwhelming and crush all difference.

I don’t mean to suggest that Leopold would disagree with this section, only that we should explicitly encourage radical innovations new ideas which are compatible with liberalism. The marginal nerdy programmer who joins EA should feel more encouraged to start a startup which directly helps longtermist goals rather than donate a tenth of his six-figure salary to the Against Malaria Foundation.

Appendix

Thomas Paine contra Edmund Burke in The Rights of Man

On intergenerational obligations

Every age and generation must be as free to act for itself, in all cases, as the ages and generations which preceded it. The vanity and presumption of governing beyond the grave, is the most ridiculous and insolent of all tyrannies … When man ceases to be, his power and his wants cease with him; and having no longer any participation in the concerns of this world, he has no longer any authority in directing who shall be its governors, or how its government shall be organized, or how administered …

I am contending for the rights of the living, and against their being willed away, and controuled and contracted for, by the manuscript assumed authority of the dead; and Mr. Burke is contending for the authority of the dead over the rights and freedom of the living …

Those who have quitted the world, and those who are not yet arrived at it, are as remote from each other, as the utmost stretch of mortal imagination can conceive: What possible obligation, then, can exist between them; what rule or principle can be laid down, that two non-entities, the one out of existence, and the other not in, and who never can meet in this world, that the one should controul the other to the end of time?

On historical authority

The error of those who reason by precedents drawn from antiquity, respecting the rights of man, is, that they do not go far enough into antiquity. They do not go the whole way. They stop in some of the intermediate stages of an hundred or a thousand years, and produce what was then done, as a rule for the present day. This is no authority at all. If we travel still farther into antiquity, we shall find a direct contrary opinion and practice prevailing; and if antiquity is to be authority, a thousand such authorities may be produced, successively contradicting each other …

Those who lived a hundred or a thousand years ago, were then moderns, as we are now. They had their ancients, and those ancients had others, and we also shall be ancients in our turn. If the mere name of antiquiry is to govern in the affairs of life, the people who are to live an hundred or a thousand years hence, may as well take us for a precedent, as we make a precedent of those who lived an hundred or a thousand years ago. The fact is, that portions of antiquity, by proving everything, establish nothing. It is authority against authority all the way, till we come to the divine origin of the rights of man at the creation. Here our enquiries find a resting-place, and our reason finds a home.

On principles and power

What Mr. Burke considers as a reproach to the French Revolution (that of bringing it forward under a reign more mild than the preceding ones), is one of its highest honours. The revolutions that have taken place in other European countries, have been excited by personal hatred. The rage was against the man, and he became the victim. But, in the instance of France, we see a revolution generated in the rational contemplation of the rights of man, and distinguishing from the beginning between persons and principles.

But Mr. Burke appears to have no idea of principles, when he is contemplating governments. “Ten years ago (says he) I could have felicitated France on her having a government, without enquiring what the nature of that government was, or how it was administered.” Is this the language of a rationable man? Is it the language of a heart feeling as it ought to feel for the rights and happiness of the human race? On this ground, Mr. Burke must compliment every government in the world, while the victims who suffer under them, whether sold into slavery, or tortured out of existence, are wholly forgotten. It is power, and not principles, that Mr. Burke venerates; and under this abominable depravity, he is disqualified to judge between them.—Thus much for his opinion as to the occasions of the French Revolution. I now proceed to other considerations.

Thanks to Leopold Aschenbrenner for comments.