Toph Tucker


The Future

3D printing, drones everywhere, computers and phones that work better faster longer (batteries), medicine, cure the common cold, cure cancer, GM foods that are actually like crazy healthy and delicious, education, monorails, non-disruptive supersonic transportion, automated infrastructure, cheap pervasive sensors, no more paper forms, a new kind of spreadsheet, efficient labor markets, more robust electrical grid, grid-scale storage, metamaterials or whatever crazy materials science stuff, water purifiers, climate control, street thug surveillance, better automatic data backup, durable data storage, parallel processing architectures, search engines for your life, a replacement for passwords, lifelogging, quantified self, natural language interfaces for everything, automatic translation, data portability and APIs, decent DRM or the eradication thereof, better intellectual property, structured data extraction, open scholarship, sense rehabilitation (sight/sound), prosthetics, gene therapy, selective abortions, domesticated pandas, recycle everything, cheap space cargo transport, space debris fix, a replacement for GPS, next-gen bicycles, smarter gyms, a new kind of math, better economic indicators and finance models, better climate models, efficient solar, safe nuclear, NEO detection/deflection, an arsenal of last-resort geoengineering tools, terraforming, hurricane suppressants, tornado dispersers, better radiation cleanup, the successor to democracy.

Thanks to Aaron Wolf for feedback.

Too Few Astronauts Have Died

Twenty-seven years ago yesterday, the Space Shuttle Challenger exploded, seventy-three seconds after takeoff. Evidence suggests that the crew quite possibly survived breakup and was alive until impact with the ocean.* NASA lead accident investigator Robert Overmyer later said, referring to Commander Dick Scobee: “Scob fought for any and every edge to survive. He flew that ship without wings all the way down … they were alive.”*

In his minority report appended to The Presidential Commission on the Space Shuttle Challenger Accident, Richard Feynman argued that NASA had been fooling itself about the risks.* Management had been estimating the fatal failure rate at about 1 in 100,000, while engineers had been estimating more like 1 in 100. History suggests that the engineers were closer to correct. (The final fatal failure rate of the Shuttle program was 2 in 135.*)

“The astronauts,” Feynman wrote, “like test pilots, should know their risks, and we honor them for their courage. Who can doubt that McAuliffe was equally a person of great courage, who was closer to an awareness of the true risk than NASA management would have us believe? … NASA owes it to the citizens from whom it asks support to be frank, honest, and informative, so that these citizens can make the wisest decisions for the use of their limited resources.”

And what are the wisest decisions? With risks revised radically upward, do we move radically toward greater caution? All else being equal, that would be a rational response. But at the same time, I fear we already have a seriously irrational risk-averseness.

I will express a distasteful opinion, which I would likely be too scared to confess if more people were listening to me:

Too few astronauts have died.

Observe that we accept very high mortality rates in war. I am not going to get into a discussion of just and unjust wars; I just mean to point out that human civilizations have long since determined that the conquest of other nations is a good worth dying for, in legions.

If that is worth sending our citizens to die for, why not the conquest of space? Is there anywhere a more threatening enemy than the engulfing vacuum, which sentences us to near-certain extinction? It is a very cold war indeed, and we are falling prey to a very serious implementation of containment.

The outrage and incredulity at any fatal accident is an artifact of a culture with the hubris to think it can afford to be comfortable. Space exploration is not something to be done at our leisure, at whatever risk tolerances we deem politically feasible.

Marshaling the resources to make a serious stab at this is incredibly hard. If not at the height of a superpower, then when? Looking around the universe, we see no evidence that anyone else has ever succeeded.* That suggests that the opportunity to try is rare and precious. It may seem to stretch into the indefinite future. But how much harder will it be to spare the resources when our cities are sinking, crops dying, population stagnating? It is cheaper now than it will ever again be. Debt, healthcare, war — these problems won’t be solved, leaving us free to tackle space, any time soon. They’ll probably get worse. And the best thing we can do for the Earth is get off of it.

There is hope on two fronts.

One: the private space sector, less burdened by politics, may be more likely to find an optimal failure rate.

Two: unmanned space exploration is becoming increasingly fruitful, and there we can evaluate optimal failure rate without particular sentimentality for the sanctity of life.

Thanks to the latter, the marginal returns to increased tolerance of risk to human life is falling dramatically. As in war, the agonizing question of how to risk life is dropping out of the equation, for better and for worse.* This question of life is emotionally provocative, but the more interesting dilemma is more general.

I doubt that the cost of failing to undertake this enterprise is being fully priced-in. It cannot be priced by empirical precedent, because this future is unprecedented; it must be priced by theoretical modeling.* That doesn’t appear to be something our free market, or indeed the free minds that compose it, are good at.

R.I.P. Greg Jarvis, Christa McAuliffe, Ronald McNair, Ellison Onizuka, Judith Resnik, Michael Smith, and Dick Scobee — K.I.A.

The Six Beings Who Have Died In Space

Six beings have died in space (i.e., above 100 km):

1957: Laika (dog), Sputnik 2 — overheating (instead of mid-orbit euthanasia as intended)*
~1965: French astrocat, name unknown (cat) — unknown
1971: Georgi Dobrovolski (human), Soyuz 11 — decompression
1971: Viktor Patsayev (human), Soyuz 11 — decompression
1971: Vladislav Volkov (human), Soyuz 11 — decompression
1972: A-3352 (mouse), Apollo 17 — massive hemorrhage*

“Others will follow. But these were the first, and they will remain the foremost in our hearts.”*

Related anecdotes:

2003: Nematodes were the only known survivors of Columbia’s reëntry disintegration.
1960: Soviet cosmodogs Damka (Дамка, “Queen of checkers”) and Krasavka (Красавка, “Little Beauty”) crash-landed in the snowy, freezing (-45°F) wilderness after their rocket malfunctioned and the ejection system failed to jettison them to safety — and the capsule was set to self-destruct after 60 hours. A rescue team found the capsule on day one, but the window was frosted over, there were no signs of life, and there was insufficient daylight remaining to disarm the self-destruct. On day two the dogs were heard barking as the team disarmed and opened the capsule. Damka and Krasavka were wrapped in sheepskin coats and flown home to Moscow.*


Not comprehensive, but the best I can do for now. Let me know if you find any others.


Facebook’s dark matter: “secret” groups

There is a discrepancy in the apparent pull exhibited by Facebook. Among industry-watchers, 30-somethings, preteens, and social media cynics of all stripes, there’s a strong narrative that Facebook is already passé. Viral tweets routinely capture the sentiment: “Twitter makes me love people I’ve never met; Facebook makes me hate people I know.” The news feed makes people miserable. Matt Haughey speaks for legions when he says Twitter is fun and ephemeral while Facebook is “mired in the past.” But, as recent search data reiterate, the site is going strong.

I suspect that, at least among my demographic of college students and recent graduates, private Facebook groups account for much of the discrepancy. In particular, groups with privacy set to “secret”, so that even their existence is private knowledge. Because of the secrecy, I think onlookers miss how big a thing this is.

ImageA very casual anonymous poll of my Facebook friends showed that respondents belonged to an average of 2 active “secret” groups. And, anecdotally, for those in secret Facebook groups, it accounts for the majority of their time on the site — like, 90%, often.

Secret groups are fun, like little self-contained private social networks. More like hanging out in private with a group of close friends. No noise, no junk, no posturing. (This will be totally obvious to some readers, but I suspect others actually don’t know that this is happening.) And Facebook is a great place for that, because the product is well-designed and we’re all already there. Lingua franca.

The mass observed in galaxies seems insufficient to hold them together, to keep them from spinning apart. So we infer the existence of dark matter, which has a gravitational signature without emitting telltale light. I think secret groups are Facebook’s dark matter.

(Or, to mix cosmic metaphors: a black hole could contain whole universes behind the veil of its event horizon.)

This isn’t just about secret groups, of course. They’re just, for me, a particularly striking example of the sort of behavior that flies under the radar because it’s quiet. Alexis Madrigal (in a similarly melodramatic fashion to all this) calls the likes of email and IM “dark social”. We can get more granular. What he described was more specifically “dark sharing”; Groups are “dark activity”. Where “dark” is just a flashier term for “private”, with some added connotations of skewing inferences from passively observing phenomena in aggregate.

Lord knows what else I’m missing about people’s private social media behavior. Like, I think Google+ hangouts are actually kinda big. Go figure, right?

I’d connect all this to Taleb and all that if I weren’t tired and losing my mind.

METHODOLOGY: The Google Docs survey asked: “How many ‘secret’ active Facebook groups are you a part of?” It was distributed by a Facebook post, which, to mitigate bias, did not mention that I was asking about Facebook groups (“one-question survey about how you use Facebook, just because I’m curious”). But this exposition of “methodology” – and the whole post – is out of proportion to sample size and all-around seriousness.

Update 28 Dec. 21:00: Hacker News discussion here.

Update 28 Dec. 21:22: “as recent search data reiterates” pedantically corrected to “as recent search data reiterate” — thanks Sharif!

The Cost of a Free Internet: The Price on Our Heads

Note: this is partially about what SOPA & PIPA are partially about, but it is definitely not about them. I have not read them but I hear they’re very poorly written and I bet they are super boring.

“Information wants to be free.” Or, at least, a lot of people want information to be free — and not just selfishly! Free universal access to information enriches the world enormously. Because it’s non-rivalrous (my “consumption” does not impede yours, so it’s hardly “consumption” at all), the gains are not zero-sum. Moreover, information in an Information Age world is extraordinarily difficult to keep contained anyway, so it seems profoundly foolish to try to cling to old ways.

That makes it very hard to “monetize” — for “content creators” to get paid for creation. They don’t need to be paid in compensation for goods transferred, because piracy does not remove their content from them. But they need to be paid for labor, and if we want to efficiently allocate resources to those who use it “best” (as judged by the semi-free market).

Yes, musicians can freely release music as a loss leader to draw people to scarce priceful concerts. Publications can bring speakers to conferences. But I don’t think those sorts of things can scale to bear the weight.

When information is post-scarcity, what can be sold?

Well, network resources are scarce. Perhaps we should abolish these all-you-can-eat plans and pay proportionate to the resources we consume. Then service providers would want you to consume — not information, which is inconsumable, but bandwidth. And then there could be revenue-sharing with the content sites, who are, after all, the ones on the front lines trying to push you kilobytes. But it would be easy for content providers to take advantage of (by, say, being lazy about serving content to you indiscriminately and ringing up a big bill), and, in any case, it hasn’t happened.

So, when information cannot be kept scarce, the only scarce thing left in the system is eyeballs. When you can’t charge for content, what you get is what we see now: an internet almost entirely paid for by advertising. And, as they say, when you use ad-supported products, you are not the customer; you are the product being sold. If they can’t sell content to you, they’ll sell you to content.

One long-term problem with that is that industry growth potential in eyeballs is limited by birthrate. We will hit peak population and peak attention. In the interim, huge successes like Facebook are successful almost entirely at the expense of others. As Chris Dixon tweeted, “The stunning speed with which new websites get traction and take share from other sites actually makes them all less valuable.” Since scaling is so cheap, it seems to be the nature of any successful web service to achieve natural monopoly, and the market will just bounce from monopoly to monopoly, indefinitely — serial winner-take-all contests, highly volatile.

Secondly, and of more immediate concern: I’m kind of sick of the ad-powered web. Pity our irrational reptile minds, played like fiddles by persuasion algorithms. It will get worse. So, I would gladly pay content/service providers as much as they’re currently making off me in advertising (i.e., as much as they can charge for my eyeballs). Jeff DeChambeau wrote today that, for Facebook, it comes out to about $4 per user per year. I’d pay that.

But payment systems are incredibly difficult and high-friction on every side, and have to be awkwardly bolted atop the web. There’s no way to just automatically and universally post a microbill to all outbound links and receive micropayment from every http referrer. But — can you imagine?

What about if I could frictionlessly pay the photographer of that lovely and apt photo I found on Google Images and want to use in a blog post? Pay something like a fraction of a cent per page view? If I could in turn benefit from frictionless revenue-sharing for works derivative of my IP, I’d be down. Not because it’s “fair”, or because anything is being “stolen” — I’m not worried about fairness, and nothing is being taken away from its owner. No, I just like the idea because I think introducing some sort of price system could more efficiently allocate resources in the ecosystem to original content creators, and not parasites.

But that’s not possible, no matter how far you go — short of wiring a secure HDCP line straight into the brain (which will probably happen).

Suppose you started a new walled-off Paid Internet, where every IP address was instead something like an International Bank Account Number (IBAN), and the network itself enforced payment to the registered cryptographically-signed owner of every packet. Well, users could still move information in and out of the technically-closed system by analog means. They could take a photo of the screen, or transcribe information by hand and distribute it across pirate peer-to-peer mesh internets. You could make the free flow of information very tedious and expensive, but you could not stop it.

And so we humans must, over the coming century, resign ourselves to being among the only scarce things left in existence, and thus the chief commodity that will be bought and sold.

A silly little thought: choices as worthless

If it’s an easy choice, it can hardly be called a choice. Choices are bifurcation points in the mind, where dependence on your mental processes is especially sensitive. “Choice” is spontaneous symmetry breaking.

So the real choices are all hard. Choice happens at the margin, on the knife’s edge. The scale is balanced. If you have two choices, you will be wrong 50% of the time — if three, 33%, etc. — or else it was not a real choice. In that sense, choices are worthless.

We smile on those whose choices turn out well, and on their choices. That doesn’t mean we smile on those who make good choices. There’s no such thing.

There are people with good circumstances. People with good bodies. Good hearts, good biceps, good brains. But there are not good choices.

Going further in redesigning Uber’s surge pricing UX

Brenden Mulligan had a good post today about Uber surge pricing UX. But I think we can go a lot further in revamping it. Here’s my meticulously rendered mockup:

(Note that I don’t have a clue what normal prices are. And trip price is not actually a linear function of as-the-crow-flies distance. Not insurmountable.)

I think it’s especially important to get this right because “fast pricing” (all pricing is variable, so don’t call it that) is so important, if we intend to take full advantage of the opportunities in market efficiency offered by the internet.

I imagine that one day, users will be able to place limit orders (“I want a ride so long as the rate doesn’t jump above this threshold in the interim”), or buy & trade options to get an Uber ride at a certain time, place, and rate. And so on. Ayup.


Get every new post delivered to your Inbox.

Join 1,363 other followers