Monday, June 18, 2007

Singularity Institute Blog

I'm currently a guest blogger for the Singularity Institute's new blog. I consider it quite an honor to contribute to such a fine talent pool.

My first two posts can be found here and here.

Thursday, May 04, 2006

Singularity Summit 05-13-06: Big Names, Big Ideas

Have you registered to attend the Singularity Summit yet? Why not? It's free!

Every time I hear about this thing I get more excited. The speaker list is already packed with some of the most important thinkers on the planet where our future as a species is concerned, and the list of VIPs expected to attend (which, alas, I'm not entitled to share) is equally impressive. Trust me on this one: If your geek factor comes anywhere near mine, you're going to meet more than one personal hero at this conference.

This is our Woodstock.

And if you're not a geek, but want to know and have a say about the upcoming end of the world as we know it (and who wouldn't), then you should still definitely go.

And if you just can't make it to Stanford at the appointed hour, take heart: It will be extensively covered by the more future-savvy media, and you will probably be able to find a full podcast or two. Check out the "Coverage of the Summit" section at the bottom right of the Summit website.

Thursday, March 16, 2006

Mobile Text Entry Showdown

I've gotten in the habit of carrying around a Pocket PC.

My biggest complaint about Pocket PCs and their PDA ancestors has always been the text entry problem, so over the last few weeks I set about looking for ways to overcome it. Voice recording and free-hand writing both have their uses, but at the end of the day I need honest digital text.

The standard on-screen options for this are stylus-pecking on a QWERTY-style keyboard, handwriting recognition, and "block recognition" -- a special handwriting style that brings warm fuzzies to long-time Palm users.

My views on these are as follows:

Too slow.
Too slow and inaccurate.
Too slow.

Enter the new contenders:

In one corner, I had a foldable keyboard for the device. Undeployed, it's not much bigger than the Pocket PC itself, but it's still a little too large to *comfortably* carry in a pocket or on a belt.

In another corner, I had heard good things about the FrogPad one-handed keyboard. At the advice of Outlawpoet, I thought I'd try the Lefty USB version. It was out of stock everywhere, so I patiently waited for one to appear on eBay and plucked it.

In a third corner, I stumbled upon MessagEase, an alternative on-screen keyboard from Exideas. This uses large keys for the most common letters, and directional swipes away from said keys for everything else.

The Folding Keyboard

There's not much to be said about this one. They come in many different sizes and configurations, and a decent one will feel pretty much like a laptop keyboard under your fingers. The only question is whether you will bother to carry it and have a place to put use it. They often don't work very well on laps.

The FrogPad

Outlawpoet's review pretty much sums up my impressions. At the risk of repeating it, I will say that the FrogPad's killer ap is on the desktop rather than with mobile devices. If you spend a lot of time in programs where you must heavily use both the mouse and the keyboard, the FrogPad is for you.

Within a week, however, I realized that I was not, in fact, a heavy user of such programs -- at least not enough to turn a blind eye to some of the FrogPad's tragic design quirks: 'm' and ',' are both much harder to hit than 'z', for instance. Dvorak typists like myself just can't tolerate that kind of oversight. More alarming, some very common typos on the Frogpad put you in ALT or CTRL modes where you can swiftly cause all manner of chaos to whatever it is you were working on. It makes you wonder whether today's Froggers are really just the beta testers for what could, and should, be an improved layout.


The MessagEase, in contrast, begs no such questions. It takes a while before you can find everything without a hunt, but it never, at the end of your search, leaves you screaming, "Why!? Why!? Why!?" Being an onscreen keyboard, it's also easier to learn -- your fingers don't block your view of the keys when you're not typing. And unlike standard on-screen keyboards, you can get to every key and character from a single screen. With not-too-much practice I already find it eerily fast. I'm not easily impressed, but I hereby award MessagEase a "damn clever" ranking, and declare it to be worth every penny of the $19.50 registration price (varies with platform).


The FrogPad would surely prove faster than MessagEase in the long run, but as a keyboard for a Pocket PC it is actually less practical than the foldable keyboards it hopes to replace. One of its selling points is supposed to be that it can be used in tight quarters where table space -- or even lap space -- are in short supply. You can hold it in one hand while typing with the other.

In my tests this proved terribly awkward. The FrogPad is too big to comfortably palm (in my admittedly smallish hands), but, with any other grip, impossible to hold steady while typing. Worst of all, in this situation I have no remaining hands with which to hold the Pocket PC itself. That leaves... my lap? I'm back where I started, and actually worse off than I would've been with a keyboard in my lap, since the LCD screen of the Pocket PC has so many suboptimal viewing angles. And Google help me should I ever need to use my stylus!

So, short of sewing the FrogPad to the side of my leg, I don't see it working as a keyboard for mobile data entry. If I'm packing a bag for a day away from home, a foldable keyboard just makes more sense. It props the Pocket PC up at a useful angle and allows me to bring my full Dvorak speed to the table. (Stowed, both keyboards are about the same size.)

And for anything other than quality time at a quiet table or desk, MessagEase reigns supreme. For one thing, it's discreet. To the untrained eye, I'm just making notes in my "planner," not writing the initial draft of, say, this post. FrogPads and folding keyboards are hard to ignore, and "Conversation Starter" is not a feature I personally enjoy in my gadgetry.

The loser is thus the FrogPad, which saddens me a little. It's pretty nifty, despite the flaws. But in my case the niche it fills is just too small. It's back to the great virtual auction house for this plucky amphibian.


As a sidebar to this little adventure, learning two new keyboarding styles had me looking around for tools to get me up to speed. For the FrogPad, you can get a free trial for their dedicated tutor. It's right-hand only for the trial, but that shouldn't make much difference. Really, any incremental lesson scheme for any keyboard will work for you if you make sure you use the correct fingering; this is equally true for MessagEase or any other keyboarding style. I just ended up using the the page I had used to learn Dvorak back in the day.

There's an awkward competency gap, however, between the time one has a new keyboard memorized and the time one is fast enough to use it in one's daily typing activities. A logical bootstrapping approach is to find a list of the most common words and letter combinations in your language and drill with them intensively. This doesn't get you practicing every key, but it gets your overall speed up to a useful level quickly.

For completely unrelated reasons, I happened to be learning a little JavaScript while I was learning the new keyboards. So, for my first attempt at a useful JavaScript program, I created a "daily speed drills" page. It incrementally drills you on a random selection of letter combinations drawn from a bank of those most common in the English language. Every time the page is refreshed, new selections are made. It's nothing fancy, but I have found it very effective.

Friday, February 24, 2006

Want a Manhattan Project? Take a Number

It seems that everyone these days is calling for new Manhattan Project.

I'll bet you've heard these cries from people looking for a 'new energy economy' or something related -- maybe fuel cell development, or a serious dent in global warming.

If you live in the United States you have almost certainly caught some other politically-charged calls for Manhattan Projects in recent years: to construct defenses against bioterror, bird flu, or SARS; to improve the quality of our nation's schools; to protect against cyberwarfare; to increase broadband internet penetration.

If you Google around for a little bit, you'll find plenty more Manhattan manifestos: to conquer the diseases of the poor; to clean up the waste from the original Manhattan Project; to build the first space elevator; to create the first nanoassembler; to cure aging; to create strong artificial intelligence; to create "global peace and prosperity"; to develop something called "intimate surveillance"; and, perhaps most esoterically of all, to "do away with the conventional divisions between the natural and social sciences and humanities."

But if you're waiting in line for your own Genuine Manhattan Project, be warned: A Manhattan Project ain't what it used to be, if it ever was.

The Numbers

The total cost of the Manhattan Project is usually estimated at 2 billion war-time dollars. Depending on how you draw the lines, this total was run up over four to six years, starting as early as 1940 and continuing to the first bombs dropped in 1945. 2 billion dollars sounds like a lot of money, and it is. But to put it in perspective, this was only about 6 hundredths of a percent of the 3.3 trillion spent by the US on the war.

Adjusted for inflation, that 2 billion would be about 22.5 billion in today's dollars, which also sounds like a lot, but is only about the size of our annual Department of Energy budget. That's only 4.5 billion a year for five years, about what we currently spend on the combined pleasures of bagged salads and used books. It's about $4.50 for every VISA card in circulation.

This doesn't seem like such a huge deal to me. I'm certainly not seeing visions of a nation with radically redirected priorities at such a low price -- to the dismay of many pundits, I'm sure.

Perhaps it's better, then, to look at the price of the Manhattan Project as a percentage of the Gross Domestic Product of the time. This would be a more realistic measure of national sacrifice. The US GDP skyrocketed during the war years, but so did Project expenditures. My calculations suggest that we can average out the Manhattan Project as costing a quarter of one percent of the national GDP for five years.

A New Unit

So, let me now propose a new functional unit, the MP, that will do for 'cost to the nation' what the LOC* (Library of Congress) does for 'quantity of data'*.

Let one MP equal one quarter of one percent of the US GDP for five years.

*One LOC is said to equal 20 terabytes (but not without debate).

One MP is far more than our earlier 25 billion dollar figure in today's much larger economy. A quarter of a percent of our 12 trillion dollar GDP comes out to 30 billion dollars; projecting for rises in GDP this would total 175 billion for the next five year period.

But still, we need perspective.

30 billion a year is less than a tenth of what we spend servicing the national debt. It's less than six percent of our defense budget. It is about one 75th of the total US budget.

The five-year total for one MP today would likely be about 600 dollars for every man, woman, and child in the country-- a figure which finally brings the number home enough to feel. But, in practice, when the government goes looking to fund a new program, Manhattan or otherwise, politics necessitates hiding the cost in some combination of sources that includes corporate taxes, royalty assessments, re jiggering of existing budgets, and new treasury bonds (more debt). So even though you pay it, you don't quite feel it.

So, is an MP a lot of money or not?

Well, it is, and it isn't.

The MP, applied

Many talk about the Manhattan Project as a research project, but it was really more of a massive industrial engineering project that happened to depended on new research. At its peak in 1945, the Manhattan Project employed some 130,000 people, but only a small fraction of these were researchers. Most were employed in the many stages of the nuclear fuel processing and enrichment chain, an entirely new industry that had to be built from scratch in total secrecy -- with all the extra expenses one would expect from this.

Not only would it have been overkill, it would have been impossible to spend an MP on just nuclear research during the war . There were only so many scientists to go around, and too many other areas needed their attention.

It would be similarly impossible, or at least foolhardy, to spend an entire MP on research for any specific project today. Let's try a scenario.

An MP for Research

Consider what would happen if, as some have suggested, we spent an MP to develop a fuel cell that was both efficient and affordable enough for widespread use in automobiles and other products.

If the government chose to start or grow an in-house program, the first thing they would do would be to hire the most qualified people they could, by offering them more money than they are currently making. Those most qualified would be those already working on fuel cell technology. All we will have done at this point is increase the average salary of fuel cell scientists and changed the bank account numbers on the checks.

Next, the project would begin hiring researchers from related fields and bringing them up to speed on the fuel cell problem. This increases the total number of fuel cell researchers, but at the expense of other projects that must now, incidentally, pay more for their researchers. The expanding fuel cell project inflates the cost of research everywhere.

(This might just encourage more people to become scientists instead of, say, lawyers or executives. This might be a good thing. But since the fuel-cell quest is billed as a Manhattan Project -- a five-year crash program -- it's unclear as to whether students would see rosy long-term prospects in science.)

Our in-house project is now swollen with huge numbers of scientists working as closely as they can. On the one-hand, they are sharing much more information than they were while working in the R&D labs of competing corporations. On the other hand, they are losing productivity to ego battles, too-many-cooks syndrome, and an intellectual monoculture. Risky alternative approaches to fuel cells are discouraged in favor of the organization's orthodox angle.

By the end of the five years, the team would probably succeed in their mission. But at what cost? The inefficiencies of cramming combined with the opportunity cost of research deferred. Who knows how many lives might have been saved by projects left waiting by researchers who had been sucked into the 175 billion dollar cash cow that was the Fuel Cell Manhattan Project? Sure, you might save more lives in the long run by getting fuel cells out there sooner, but it's awfully hard to say. And it will be impossible to say that a different but equally viable cell -- or even some non fuel-cell alternative --would not have been developed in the same time period, in the absence of the Project.

The closest thing to a research MP that I can think of right now is ITER, the international project to test the feasibility of a commercial-scale power generation via tokamak-style fusion reactor. Despite having a large physical engineering requirement on top of the heavy research budget, ITER is currently projected to cost a total of only 10 billion dollars. Even if overruns take it to 20 billion (a not unlikely scenario), that will only total about .11 MP.

The lesson? Unless your pet research project is very broad in scope, and can be distributed among many different fields -- something like SENS, the project to defeat physical aging -- an MP is way too much money, and would ultimately carry too many hidden costs. Asking for such an absurd sum is surely counterproductive to your cause.

But what if yours is also a massive civil engineering project? Is an MP still too much?

An MP for Civil Engineering

Let's suppose we've decided to spend an MP to create a 'hydrogen economy'. Advanced fuel cells are an obvious starting place, and we've already concluded that an MP is way too much to spend on just this, so we'll have plenty of money left over. We're also going to need fuel tanks for dense, safe hydrogen storage. That's another big research project, but probably no larger than the fuel cell one, so we still have plenty of our MP left to spend -- let's say 80 percent.

So we'll move on to the hydrogen infrastructure. I'm guessing that .8 MP will be more than enough to retrofit our nation's approximately 170000 fueling stations to safely pump hydrogen. It might even be enough to start on the network of tanks and pipelines needed to handle the new fuel. But I doubt it would be enough to finish it.

And we haven't even started on the hydrogen production infrastructure. We have to figure out what combination of facilities to make, then build them, whether these be be next-generation high-temperature nuclear plants, genetically-tweaked algae, solar 'cracking' of water molecules, the reprocessing of fossil fuels, something we've not yet thought of, or simply industrial-scale water electrolysis using our existing power grid. This production infrastructure could easily cost a few MP all by itself.

(We won't, but at this point we are in a position to consider just how many tens or hundreds of MP it would cost to make a real dent in global warming by spreading this new energy infrastructure around the world...)

The moral here is that for your project with a large civil engineering component, you are likely to need much more than an MP to get the job done. You might, instead, ask for one or more PAs (A Project Apollo is equal to three tenths of one percent of GDP for ten years). More realistically, consider breaking your cause down into the smallest sub-projects that can be seen as independently valuable, and seek funding for some of those instead.

(I should note that there are some "smaller" civil engineering projects that may, in fact, fit easily within the bounds of a single MP. These include the Space Elevator, estimated to cost 0.23 MP for the first useful 91,000 km tether, with subsequent cables costing closer to 0.08 MP. )

The Bottom Line

The original Manhattan Project was expensive, but not nearly as expensive as our gut reaction today would suggest. Pundits use "Manhattan Project" as a clarion call for new national priorities and collective sacrifice, but a single MP may not actually qualify as such.

Nevertheless, it would not take too many simultaneous MPs to knock the wind out of our economy. With so many different calls for Manhattan Projects, we must be judicious in the asking and granting of funds. It helps to know that research projects probably overstate their needs when they cry for an MP, while projects with civil engineering objectives may tend to understate -- even wildly so.

As a unit of cost, the MP may prove useful for helping people wrap their heads around really large expenditures, letting them more intuitively tell the difference between a merely very expensive project and an astronomically pricey one.

In any case, if people are going to keep asking for Manhattan Projects, it makes sense to quantify the term.

Saturday, February 18, 2006

Timmy and the Dragon

Februrary 11, 2006. Mexico. The services were in Spanish, a language I spoke for two years in Spain as a missionary in a time that now seemed as impossibly distant as my religious past. The trembling convictions spoken by Timmy's siblings, my wife among them, were thus understood, but foreign. Familiar, but empty. The contradictions of a heavy dream.

I was seated near the front center, my eyes left to idly study the casket, a handsome cloud-gray vessel in tasteful, fully rectangular proportions. Sturdy shafts ran along the sides. The pallbearers, a group including myself, would use these to heft the too-light assemblage away to a place where it would never be seen again.

The mounting brackets for these shafts were decoratively embellished. Each seemed, to me, to resemble the mask of a ravenous reptile. I recognized this for what it was: The mark of the Dragon-Tyrant.

For Timmy, it was no mere cliche to say his life was over before it had begun. Born severely disabled, he was never able to whisper a single word, never able to give his parents a hug. Had he not been adopted from birth, he would almost certainly have been left to die, either tacitly or explicitly, in the cage of extreme poverty into which his was born.

Every day of his seventeen-year life, his very existence was thus a living reflection of his parents' indefatigable charity. Theirs was a love so powerful and pure that the entire community basked in it. (That so many people were now crowding the church for the funeral of a drooling, twisted mute was final evidence that I do not exaggerate.)

It was not enough. Under the rules in which humanity came into existence, the house always wins. The Dragon-Tyrant of physical aging and death has a particularly loathsome modus operandi, typically tearing first into your grandparents and parents before moving on to your friends and perhaps some of your children. As he then spirals in to finish you off, you may well feel that life is no longer worth living. The kill of the Dragon is thus total, both in body and spirit.

I am probably luckier than most. I did not lose so much as a grandparent until my early twenties. Then I lost my Grandpa. After a pause of several years, the Dragon has returned to assert his dominion. In the span of about a year, he has consumed my Grandad, my Grandma and now, my brother-in-law, Timmy.

The Dragon's orbit draws tighter and faster.

In such times, a natural instinct is to reach for a belief in a better life after this one. To accept death as the will of god. There is comfort in these ideas that I will not begrudge on anyone. Submitting served humanity well for many thousands of years, when revolt against the Dragon-Tyrant was clearly impossible. But in the end, when we submit, we still collaborate with the enemy -- especially now, when the possibility of defeating him is real.

Yes, he is the enemy. And I will not submit.

This makes me a cornered animal. An ice floe drifting in the sun. A sack of thermite left next to a furnace at the onset of winter. For inevitably, the Dragon must devour someone I have loved very closely, and my fury may well incinerate me from within. For even in the exquisite, faithless pain of total mourning, I will not submit.

It is customary in many circles to make charitable donations in memory of loved ones at the time of their passing. This is a good tradition. Timmy's parents, for example, run a school for disabled children in an impoverished corner of Mexico. I admire and respect anyone who contributes to this or any other worthy cause.

I definitely participate in this giving tradition, but have nothing left to donate on the occasion of Timmy's passing, because I have not been postponing my charity for such occasions. It would be wrong of me to wait until my loved ones are gone, for my charity of choice intends to save them.

So it is that I share my tears with Timmy's many loved ones who, because of their beliefs, cannot fully endorse my cause. But I will not submit.

In the name of Timmy, I will fight the Dragon.

Join me. The Singularity Institute seeks to create benevolent artificial intelligence as a stepping stone to ending involuntary death and suffering in our time.

Monday, January 16, 2006

Why Google Wants Your Data

I was configuring my Gmail account for POP access today when the thought occurred to me: "Why does Google let me do this? Gmail makes the company money only when users click on ads. When I access my email via POP, I never even visit the site."

Perhaps they think POP access will draw enough new users to make up for the lost revenue. After all, even the POP people will probably use the web client at times when they are away from their desks.

Personally, I think Google is just looking ahead -- to the era of personally-tailored search results.

You can't get there by merely indexing every web page and every book ever written, as useful as these habits are. No, to make the next leap in search, Google must get to know you personally. Google therefore hopes to learn your tastes and your tendencies, or at least have the tools in place to do so later. What better way to do this than with products useful in their own right, like Google Toolbar, Google Desktop, and Gmail​?

If you ask a good friend whether you should see "Disappointing Sequel III" this weekend, she'll give you an intelligent answer. She won't just tell you if she liked it, but will use her knowledge of your tastes to venture whether you would like it. "I didn't care for it," she'll say, "but you will probably love it."

Now, imagine your friend is a search engine. If you type 'fedora', your friend would know you are, in fact, one of those few twisted souls more interested in hats than operating systems.

Google, on the other hand, will talk about Linux until about the 20th link. That's because, right now, it just plays the percentages by looking at the links between pages. If you mean what most web page makers do when you search for 'fedora', you are probably satisfied with the results.

As you may know, Google already records every search query you enter and links it to your IP address. They don't yet use your search history to improve your search success. There are privacy concerns, after all. Imagine your horror if your geeky friend came over to your desk, and, googling 'fedora' on your computer, saw a bunch of links about early twentieth-century head coverings come up.

Google could certainly do more to get to know you. Suppose, for a moment, that you're ok with this.

If you let the search engine monitor your surfing habits, it would know which sites you visit and for how long. This would not only give it a better picture of your overall habits, but would let it know what you're working on right now. If you're uncharacteristically looking at pages about Linux, it might correctly assume that a search for 'fedora' should ignore hats, just this once.

The engine gets similar benefits by reading your mail, your instant messenger transcripts and your half-forgotten, half-baked novel.

Now, give it a little more access. Let it measure how fast your computer is and how big your hard drives are. Let it watch which applications you use, and how you use them. Let it hear what music you listen to. Let it tally up how fast you type, and which words you misspell most often. Let it see your digital photos.

How is it supposed to make use of this information, you might ask? By finding patterns among users.

Some of these patterns might make a kind of sleuthy sense to us: Perhaps it guesses you like hats -- not because of your previous searches, but because it has discovered that users who, like you, listen to country music and take pictures of horses, tend to spend more time on hat pages than on Linux pages.

Some patterns it uncovers might seem bizarre: Maybe it finds that people who listen to country music and like hats express the most love online to Mac users who read the Onion and take photos with greater-than-average red eye. Google Date, here we come.

Pattern-seeking software has another name: expert system.

Medical expert systems help doctors diagnose illnesses that fit complicated patterns of symptoms and test results. Outside their area of expertise, expert systems are useless. But, within it, they are savants. Sifting through mountains of data, they look for, and find, patterns that humans never would. Sometimes these are statistical artifacts that don't hold up against future data. Sometimes the conclusions hold up surprisingly well, for reasons we simply haven't figured out yet.

People like me refer to expert systems as a type of narrow AI. Google, or any aspiring rival, must see their products as examples of narrow AI in the very wide domain of answering questions. If a search engine or any other expert system can widen itself sufficiently, it ceases to be narrow and becomes a general AI. Better hope it knows how to 'not be evil'.

With this longer view in mind, we should not be surprised by observations that Google is collecting or retaining more data than would seem necessary -- or even profitable. Likewise, we can expect Google to continue churning out products that give users immediate benefits while giving their clusters more data to crunch. We can even know when Google is about ready to serve us the Next Big Thing: when it provides, behind a personal log-in screen, a beta search tool that asks if it can access your other Google tools to give you better results.

Whether you use the improved search will be up to you. Technically, Google should be able to do all this without violating your privacy. But if you're the kind of person who doesn't even want a program looking at your honeymoon photos, you will probably pass.

What were you two doing with that fedora, anyway?

Monday, January 02, 2006

Conversation with an Atheist

So you're religious.

You've probably read hundreds of pages of scripture and commentary regarding a single faith -- not because you shopped around, but because you were born into it. Why shouldn't you cuddle up for a bit with some friendly rebuttal?

I've even done my best to make it fun.

Conversation with an Atheist

Seriously, though... I'm not out to attack anyone. If you're totally content with your religion, move along. There's nothing for you here.

And if you're not religious? I had thought of subtitling this piece "an introduction to rational thought." It may still make you smile.

New Home

I had been itching to buy a new domain and switch web space providers for some time. I also had a number of longer written pieces near the end of the production pipeline without any great place to put them.

A new site seemed like the best way to pull everything together.

As mentioned there, I'm not getting rid of this blog. But when I have longer pieces I will host them on my site and link to them from here.