One useful way to think about the future is to ask what key future dates are coming, and then to think about in what order they may come, in what order we want them, and how we might influence that order. Such key dates include extinction, theory of everything found, innovation runs out, exponential growth slows down, and most bio humans unemployed. Many key dates are firsts: alien life or civilization found, world government founded, off-Earth self-sufficient colony, big nuclear war, immortal born, time machine made, cheap emulations, and robots that can cheaply replace most all human workers. In this post, I want to highlight another key date, one that is arguably as important as any of the above: the day when the dominant actors take a long view.
So far history can be seen as a fierce competition by various kinds of units (including organisms, genes, and cultures) to control the distant future. Yet while this has resulted in very subtle and sophisticated behavior, almost all this behavior is focused on the short term. We see this in machine learning systems; even when they are selected to achieve end-of-game outcomes, they much prefer to do this via current behaviors that react to current stimuli. It seems to just be much harder to successfully plan on longer timescales.
Animal predators and prey developed brains to plan over short sections of a chase or fight. Human foragers didn’t plan much longer than that, and it took a lot of cultural selection to get human farmers to plan on the scale of a year, e.g., to save grain for winter eating and spring seeds. Today human organizations can consistently manage modest plans on the scale of a few years, but we fail badly when we try much larger or longer plans.
Arguably, competition and evolution will continue to select for units capable of taking longer views. And so if competition continues for long enough, eventually our world should contain units that do care about the distant future, and are capable of planning effectively over long timescales. And eventually these units should have enough of a competitive advantage to dominate.
And this seems a very good thing! Arguably, the biggest thing that goes wrong in the world today is that we fail to take a long view. Because we fail to much consider the long run in our choices, we put a vast great future at risk, such as by tolerating avoidable existential risks. This will end once the dominant units take a long view. At that point there may be fights on which direction the future should take, and coordination failures may lead to bad outcomes, but at least the future will not be neglected.
The future not being neglected seems such a wonderfully good outcome that I’m tempted to call the “Long View Day” when this starts one of the most important future dates. And working to hasten that day could be one of the most important things we can do to help the future. So I hereby call to action those who (say they) care about the distant future to help in this task.
A great feature of this task is that it doesn’t require great coordination; it is a “race to the top”. That is, it is in the interest of each cultural unit (nation, language, ethnicity, religion, city, firm, family, etc.) to figure out how to take effective long term views. So you can help the world by allying with a particular unit and helping it learn to take an effective long term view. You don’t have to choose between “selfishly” helping your unit, or helping the world as a whole.
One way to try to promote longer term views is to promote longer human lifespans. Its not that clear to me this works, however, as even immortals can prioritize the short run. And extending lifespans is very hard. But it is a fine goal in any case.
A bad way to encourage long views is to just encourage the adoption of plans that advocates now claim are effective ways to help in the long run. After all, it seems that one of the main obstacles so far to taking long views is the typical low quality of long-term plans offered. Instead, we must work to make long term planning processes more reliable.
My guess is that a key problem is worse incentives and accountability for those who make long term predictions, and who propose and implement longer term plans. If your five year plan goes wrong, that could wreck your career, but you might make a nice long comfy career our of a fifty year plan that will later go wrong. So we need to devise and test new ways to create better incentives for long term predictions and plans.
You won’t be surprised to hear me say I think prediction markets have promise as a way to create better incentives and accountability. But we haven’t experimented that much with long-term prediction markets, and they have some potential special issues, so there’s a lot of work to do to explore this approach.
Once we find ways to make more reliable long term plans, we will still face the problem that organizations are typically under the control of humans, who seem to consistently act on short term views. In my Age of Em scenario, this could be solved by having slower ems control long term choices, as they would naturally have longer term views. Absent ems, we may want to experiment with different cultural contexts for familiar ordinary humans, to see which can induce such humans to prioritize the long term.
If we can’t find contexts that make ordinary humans take long term views, we may want to instead create organizations with longer term views. One approach would be to release enough of them from tight human controls, and subject them to selection pressures that reward better long term views. For example, evolutionary finance studies what investment organizations free to reinvest all their assets would look like if they were selected for their ability to grow assets well.
Some will object to the creation of powerful entities whose preferences disagree with those of familiar humans alive at the time. And I admit that gives me pause. But if taken strictly that attitude seems to require that the future always remain neglected, if ordinary humans discount the future. I fear that may be too high a price to pay.
loading...
