Some Quotes

Due to moving and having my computer locked in my advisor’s office for the evening, a short post consisting mostly of the words of Freeman Dyson from Infinite in All Directions. The book is chock full of ideas that are still relevant today, 25 years after it was first published. In that way, it’s like extremely technical science fiction. Time has rendered some subjects less relevant (nuclear weapons) and others untrue (predictions of an awesome outer-solar-system bound mission launched in 2014) but the core has pivoted to still ring extremely true a quarter century later.

On space colonization:

“Some of the first questions which come up in any practical discussion of space colonization are questions of economics…How do we make a living? What can we expect to export in order to pay for necessary imports? If space colonization makes any sense at all, these questions must have sensible answers.”

This is a huge question that bugs me all the time. I think that an answer will only emerge if we just go out and try.

On the limits of AI (actually a quote of a quote from Sir James Lighthill):

“Valuable results flow from the integration of intellectual activity with the capacity to feel and to relate to other people. Until this integration happens, problem solving is no good, because there is no way of seeing which are the right problems.”

Computers have gotten much better at solving problems since 1988, even ones where they don’t know the problem beforehand. However, I don’t think that computers have gotten any better at defining the problem to be solved. I suspect this is an area where humans may always be relevant.

On why genetically engineered crops are not aberrations:

“Farmers have been growing wheat in open fields for thousands of years, and wheat is also a product of human manipulation, just as artificial as genetically engineered Escherichia Coli

(I would also note to the paleo folks that almonds were domesticated after wheat.)

On why “quick is beautiful”:

“The nuclear industry is not the only one which has suffered from a hardening of the arteries and has lost the ability to react quickly to changing conditions and changing needs. The difference betwee n a three-year and a twelve –year reaction time is of crucial importance.”

Not only does this apply almost everywhere, I think three-year vs. twelve-year is a great rule of thumb to keep in your back pocket when evaluating technology predictions or project proposals.

And that is just a small sample of Dyson’s insights. Since the expected duration of ideas into the future is roughly proportional to how long they’ve survived so far, I expect these words to stay relevant for at least another quarter century. You should read the book. Yes, you.

HAL the Engineer

SPOILER ALERT. If you haven’t read or watched 2001: A Space Odyssey, you should probably stop now and reconsider your cultural literacy.

A conversation last weekend ended up focusing on how the character HAL in 2001: ASO took the easy-to-fix spacecraft paradigm to an extreme when he tried to remove the hardest-to-fix parts of the spacecraft: the humans. (Don’t Mechanical Engineers talk about fun things at parties?) In retrospect, I think HAL actually occupies an interesting intersection between the problems of the hard-to-break paradigm and the limits on computer thinking.

In 2001:ASO, HAL, the ship’s AI comes to the conclusion that the Humans on board the spacecraft are actually the biggest liability to the mission. While you could look at this as making the ship easier to fix (a dead human is very hard to fix indeed) I think that HAL was actually falling into the ‘make it as hard to break as possible’ trap. It’s true that in space, humans are vulnerable, especially since so many other failures have the secondary effect of breaking us – life support failure, a window breaking, etc. Thus, from a hard-to-break perspective, getting rid of the humans makes some morbid sense.

Of course HAL sees humans as the weakest link. A robot can deal with anything it can possibly imagine better than a human. Ah, but therein lies the rub. The humans are far better at dealing with the things that they can’t possibly imagine – like the ship’s computer turning murderous. Thus, from an easy-to-fix perspective, humans become extremely valuable.

Perhaps it’s a bit pie-in-the-sky, but I think that a shift towards an easy-to-fix paradigm of spacecraft engineering goes hand in hand with an increase in manned missions. As Julian Simon pointed out, the human mind is the ultimate resource – one that is far easier to leverage if it’s onboard rather than in a control center millions of miles away, to say nothing of the incentives.


Sci-fi Startup

I was was originally going to discuss the intersection of easy-to-fix spacecraft and human-robot task differentiation: the story of HAL 9000.  But then it went somewhere entirely different:

I just realized that science fiction has a striking similarity to startups. Most Sci-Fi predictions look ridiculous and dated in retrospect. ‘How could anybody have thought that is what the future would be like’ we think. Just look at most of the images from the 50’s of the year 2000. The exact same situation holds true for new companies – most of them end up failing, leaving the question ‘how could anybody ever have thought that would make money?’ But those crazy ideas – companies or fiction – that survive the test of time bring a lot of value to the world.

Successful companies add value in an obvious way – making our lives better through improved products or services.   Science fiction does it more subtly by providing a framework for thinking about a world that does not yet exist. Just as startups often ‘pivot’ away from their original market while maintaining their core idea, sci-fi never predicts the future exactly.  Instead, the ‘successful’ sci-fi (for this axis of success) predicts the essence of a future technological or societal change even if it doesn’t happen exactly the way the author imagined.

Of course, sometimes science fiction is just successful because it’s a damn good story. Nothing wrong with that.


The economist Arnold Kling inadvertently did a great job of describing how we need to shift the paradigms of spacecraft design. I’ve replaced only one word in this quote:

 “We tend to think of the task of engineering as one of making systems hard to break. An alternative to consider is making systems easy to fix. Think of a computer. You can try to use firewalls and anti-virus software to make your computer hard to break. But it still pays to back up your data to make it easy to fix.“

The substitution changes the meaning entirely, but doesn’t alter the truthiness at all.

I look at it this way: both things that are easy to fix and those that are hard to break are robust. ‘Robust’ spans the gap between fragile and antifragile. Hard-to-break is on ‘fragile’ end of the spectrum while easy-to-fix is closer to ‘antifragile.’

I’ve used this example before, but what are two of the coolest fictional spacecraft? The Millennium Falcon and The Serenity. What do they have in common besides a roguish, gun-slinging captain, a crew avoiding the authorities, smuggling compartments and frequent comparisons to various trash receptacles?

They’re both really easy to fix. On top of being a great plot device, I think in the end it allows earned confidence, rather than assumed confidence.

A lot of technology has trended towards ‘hard to break’ – cars, phones, and appliances come to mind. In these cases, hard-to-break may be preferable because the replacement costs have dropped dramatically, both in terms of money and effort.

On the other hand, if you’re millions of miles from the nearest mailbox, you can’t exactly a replacement ship on Amazon. Given the remote hostility of space, which category would you rather your ship fall into?   I know which one I’d choose.

Links of the Week 10-26

Saturday Stories!


TFHL Musings

Two tinfoil-hat-line-skirting thoughts.

 Uranium In Space – I haven’t heard it discussed, but it seems like one of the most valuable things to mine in space would be uranium or other fissile material. Nuclear reactors are incredibly valuable in space. They are basically the only option for long-term power to meet requirements greater than the amount the sun can provide. And since the sun doesn’t provide much power past the asteroid belt, or during a dust storm on Mars, nuclear power will be important for significant exploration of the solar system.

Unfortunately, it’s an epic regulatory adventure to launch any sort of radioactive material from earth. If a rocket with a radioactive payload were to explode during launch, it could potentially contaminate a huge swatch of land.  So, instead of launching huge chunks of fissile material, why not aggregate it in space? One option is to mine uranium in space (although I have no idea if mineable asteroids or the moon have anywhere near enough to consider.) Launch a little bit on each rocket that goes up and then bring together the pieces in orbit. This would effectively divide the risk into pieces so small they are zero.  You are more likely to suffer adverse effects from old watches than a tiny bit of radioactive material smeared over several hundred miles.

Stop Worrying about Martian Life – We need to stop hobbling space exploration (especially on Mars) by worrying about ‘biological contamination’ – introducing earth bacteria that could mess up tests for present or past Martian life. There will never be a point where it will be possible to say ‘AHA we can conclusively show once and for all that there is/was/never was life on Mars.’

Honestly, I think the upside of unbridled expansion of humanity and earth-originated life generally is huge. The advantages I see far outweigh the disadvantages of ‘biological contamination.’ Finding conclusive answers, even without ‘contamination’ has low probability combined with low impact (the most significant thing we’ll find by now are fossilized microbes.) Of course, I’m also glad that Europeans made their way to North America, so take it all with a grain of salt.