David C Denkenberger on Food Production after a Sun Obscuring Disaster

Having paid a moderate amount of attention to threats to the human species for over a decade, I’ve run across an unusually good thinker with expertise unusually suited to helping with many threats to the human species, that I didn’t know about until quite recently.

I think he warrants more attention from people thinking seriously about X-risks.

David C Denkenberger’s CV is online and presumably has a list of all his X-risks relevant material mixed into a larger career that seems to have been focused on energy engineering.

He has two technical patents (one for a microchannel heat exchanger and another for a compound parabolic concentrator) and interests that appear to span the gamut of energy technologies and uses.

Since about 2013 he has been working seriously on the problem of food production after a sun obscuring disaster, and he is in Lesswrong’s orbit basically right now.

This article is about opportunities for intellectual cross-pollination!

Appearances On Or Near Lesswrong In The Past

On 2016-05-10 RyanCarey posted Improving long-run civilisational robustness and mentioned Denkenberger as one of the main authors in the literature of shelter construction after really major disasters, with a special interest in the food production that would happen in such facilities.

On 2017-08-24 ChristianKl posted [Link] Nasas ambitious plan to save earth from a supervolcano in whose comments turchin mentioned Denkenberger as a relevant expert.

Slightly farther afield but very recent and still nearby, on 2017-09-14 Robin Hanson posted Prepare for Nuclear Winter which was a very abstract and formal exhortation to care about global food production in the event of a sun obscuring disaster, until the final sentence where he called attention to “ALLFED” which is very dense with citations to papers on the subject, and basically every paper has Denkenberger as a co-author.

In the last few hours I saw Denkenberger working his way through comments on Hanson’s couple-day-old posts, correcting the factual mistakes in people’s comments with links pointing to papers that contain the correct information.

There is probably an intellectual opportunity to get Denkenberger’s attention and help Lesswrong get smarter about an important sub area related to the mitigation of existential risks.

A Generic Solution To Many Classes Of Risk

One of the long term deep insights on X-risks that is somewhat unique to Lesswrong is the idea that specific disaster scenarios often seem more plausible and more likely to people when additional details are added, but logically speaking each burdensome detail actually makes the scenario LESS likely.

Also, once a detailed scenario is accepted as worryingly plausible by an audience, the natural tendency is to find solutions that address that single scenario...

“We will solve the asteroid problem by shooting lasers at the asteroid before it hits!”

“We will solve the AI problem by making the AI intrinsically motivationally safe!”

“We will solve the nanotech problem by building a planetary immune system out of better nanotech!”

Once you really feel the problems of burdensome details in your bones, it becomes clear that the cost effective solution to many such problems is plausibly the construction of a SINGLE safety measure that addresses all or almost all of the problems in a single move.

(Then perhaps build a second such solution that is orthogonal to the first. And so on, with a stack of redundant and highly orthogonal highly generic solutions, any one of which might be the only thing that works in any given disaster, and which does the job all by itself.)

One obvious candidate for such a generic cost effective safety intervention is a small but fully autonomous city on mars, or antarctica, or the moon, or under the ocean (or perhaps four such cities, just in case) that could produce food independently of the food production system traditionally used on the easily habitable parts of Earth.

The more buffered and self sufficient such a city was, the better it would be from a generic safety perspective.

It appears to me that Denkenberger’s work is highly relevant to such a project, and for this reason deserves our attention.


I’m thinking it might be interesting to start a bunch of comment threads below, one for each of Denkenberger’s papers that can tracked down, that could be discussed and voted on independently.

Also, if Denkenberger himself is interested in having me correct errors in this article or put a prominent message written by himself somewhere here at the bottom or the top, I’m open to that.

Another thought would be to try to schedule an AMA for some day in the future, and link to that from here?