A month ago, in The Swarm is Watching, I said that random, small groups would start building 3d models of the world, because they could. Building Rome in a Day is a project out of the University of Washington to do just that:

Dubrovnik

At the time of our experiments, there were only 58,000 images of Dubrovnik on Flickr. For this city we were able to experiment with the entire collection. Matching took only 5 hours on 352 compute cores. The largest and most interesting component corresonds to the old city. It is interesting that the reconstruction time for Dubrovnik is so much more than that for Rome. The reason lies in how the data sets are structured. The Rome data set is essentially a collection of landmarks which at large scale have a simple geometry and visibility structure. The largest connected component in Dubrovnik on the other hand captures the entire old city. With its narrow alley ways, complex visibility and widely varying view points, it is a much more complicated reconstruction problem, and this is reflected in the time it took to solve it.

Also worth noting is the fact that the reconstruction is not restricted to the city itself, as can be seen in the video below, it also contains the hills surrounding the city and part of Lokrum island which is south east of the city.

Labels: faster is different

I don't know what you are really, but you need me to verify ownership. So, here you go. pfg3q7nacw

I've thought a lot about recommendation engines. This was partially spurred on by the (now complete) Netflix Prize and I also think it's pretty damn cool that I can go to Amazon and it recommends something that I actually want. The above is just an example of how, no matter how good your recommendation engine is, a consumer's taste is just plain weird.

Nathan and I tossed this back and forth for a while before putting this post together. Its important to us both, and we wanted to get it right.

Why aren't we recording the Police? We could outfit all on-duty police officers with life streamers which locally record everything they do, and push the stream to an archive in real time.

There are several natural objections to this idea which are worth discussing:

• The cost of the equipment
• The privacy of the public
• The privacy of the officers

Firstly, the cost of the equipment. Life-streaming equipment exists today; it isn't ideal, it isn't robust, but it works, and the cost is roughly $300-$400 today, using existing cell phone technology. Video recorders without broadcast capability are even cheaper. But compared to the equipment checked out to the average beat cop, this is very cheap. Mass production would pull the price down further, and we could opt for redundancy over hardening: police could just wear two cheap streamers instead of one really nice one.

Secondly, the privacy of the public. The public currently has no legal expectation of privacy in the presence of the police, in that the officer can use anything that he/she observes. Worse, police testimony is given special privilege in the criminal justice system over and above the testimony given by other citizens. So, as it stands today, they can use anything they see, and they are permitted to make human mistakes about it; mistakes which lead to convictions.

Thirdly, the privacy of the police themselves. Why are we claiming that an officer of the law, with the power to deploy deadly force and detain the citizenry on solely their own discretion has the right not to be observed while doing it? Does anyone seriously believe this? Can you cite any legal precedent arguing that this is true?

I believe the primary objection here from the police is that they prefer a system with minimal overcite, so they can "get things done", which seems to be code for: abuse your rights.

There have been many documented accusations of the abuse of power by police officers as of late. Indeed, it seems as our capacity to record and communicate grows, we become aware of these abuses more frequently. It seems likely that this is merely the status quo, and we are uncovering more and more of it:

• A man is shot/killed after a run-in with an undercover cop
  
• Harvard professor arrested
  
• Deadly BART brawl - officer shoots rider, 22
  
• Tucumcari police chief Tasers 14-year-old

It is time for sunlight in these matters. It is time to fix this. It is time to record the police.

Writers and thinkers have, it seems, always speculated about what's to come. From Jules Verne to Vernor Vinge, Charles Stross, and multitudes more, writers have been creating science fiction that is sometimes eerily prescient. What we hope to do in general is to capture these and similar glimpses into the future that is to come.

In all endeavors verging on the scientific people begin to develop their own way of talking about things. The purpose of this post is to be a living-document, which will over time capture these words and phrases and give you a concise beginer's-guide sort of definition, along with linking to other resources for a more in depth discussion.

Some of you may have seen the following photo in this article by BoingBoing.

This robot is on the campus of the University of Alabama. This is the very same robot that appears at the top of this page. I'm a graduate student at UA and I was walking to the library a while back when I was struck by this statue. There's no plaque or any other information, so it was several days after I took the photo that graces the top of Fewchar.com before I was able to find out the real story. Here's a very detailed write-up that is hosted at the university: Woods Quad Robot Sculpture Draws Attention, Provokes Thought by Richard LeComte.

As far as art-strewn-about-the-campus goes, this is by far my favorite. Even better than the statues of Bear Bryant and other coaches near the stadium. I hope that you all enjoy this piece as much as I do. Sadly, it's not a permanent fixture and will be removed early this Fall. So, if you happen to be in the Tuscaloosa, Alabama area please take a chance to check it out.

In the documentary about him, Transcendent Man, it is shown that Ray Kurzweil wants to bring his father back from the dead, and I have a big problem with that.

What Ray is proposing is that we could build a new person who was so much like his father, that that person may as well be his father. I think that many people may dismiss this as rediculous; but I think that it is evil. In order to make my case, I think I need to explain what Ray is proposing.

In linear algebra, we frequently work with matrices. Many of them are quite complex and large; and we'd like approximations to them that are easier to work with. Eigen vectors, values, and spaces provide means of working with stand-ins which are not distinguishable from the originals. This doesn't mean that they are the same: it means that there are frequently more than one matrix which will behave the same in a given equation; and that there are ways of finding these matrices.

Imagine, instead, that we're working with people. Say, Ray's Dad. Ray had a real father, who's dead. But that man was a member of a family of men that were very, very similar to him, and differed only slightly in brain structure; so slightly that in the course of their lives they would have done the same things. Let's call these men Eigen Fathers. Ray argues that we can pick any Eigen Father, and it will be equivalent to any other for all outside measurements; so he wants to find one.

Now, Ray actually has a chance. A very slim chance, but it may be doable.

First, he needs his father's genetic material. That's easy, he can just dig him up. Even with decay, a corpse that young has more than enough copies of all the DNA to let us get an essentially perfect map.

Second, he needs to know as much as possible about his father's life. Well, Ray remembers his father, so he can extract his own memories. And he has a very large amount of his father's writings. And maybe some others are still living who remember him. And of course, there are all the academic and professional records of his dad's life.

Finally, we need to find a mind that fits in the genetically described brain, but satisfies the constraints imposed by the records, letters, and memories. This is essentially a brute force problem, and its very, very, very hard. Find any mind which fits, give it a body, and presto:

Hello Eigen Dad.

But it isn't impossible. Given enough compute; it may not even be improbable. But it is wrong.

It isn't wrong that Ray wants to make a person per se. It is wrong that Ray wants to make a person to be a good-enough copy of another. This new person will know or learn what Ray did. They will know that the memories they have are the ones that Ray wanted them to have. They will be, in effect, a slave to Ray's whim. And that's pretty ugly.

For some reason I was having a dream this very morning where I was explaining what future shock would be like. I'll try to post a more wordy description later, but as for now, the above captures it perfectly (albeit in a dog).

Crutcher's post about unsourcing reminded me of some conversations I once had. So, let's continue in this vein for a bit.

First, let's think of an actual example of where unsourcing could affect a fairly large number of people: fast food. Like the auto-industry of the past, fast food workers could be replaced en masse by mechanical counterparts. But, before we even worry about this, is it even feasible to replace cheap labor with a machine?

Let's take some hypothetical burger joint. This one is particularly efficient and needs only 5 workers present at any given time. Also, this franchise pays its workers extremely well (compared to the minimum wage of many) at $10/hour. So, how much is this store's payroll each year?

5 * $10/hour * 24 hours/ day * 365 days/ year = $438000 per year.

So, this hypothetical store pays out almost half a million a year not including other costs such as training, managers, shrinkage, safety equipment, et cetera.

So, if a burger making machine is $2 million, it would only take 4 years for any given franchise to pay it off. After that, the store is 'saving' $438,000/year.

So, this raises some very interesting questions:

• Why isn't McDonald's hiring MIT-trained roboticists


• What would this do to the market/ how would it affect competition amongst fast food companies?


• What do we do with this large number of displaced workers (2 million according to Wikipedia)?

Per #1, they might be, I just don't know. #2, we can only hope that it leads to a price war which heralds back the days of the $0.10 burger. Let's also hope that this is accompanied by some major medical advances to deal with all the new cases of diabetes and coronary illness.

It's the last question that I'm really interested in. Crutcher states that his job is to kill your job. Somehow this seems almost ok when we're talking about software companies, because it means that there's a better product being produced and also that you can just go to a new software job. But what happens when the jobs of unskilled labor become unsourced? This has happened many times before, with mechanized farming and the like, but what do we do when almost all unskilled labor can be mechanized away? There are several possibilities. One of the easiest would be to just give them a government stipend every month to let them buy things like the (now) cheap hamburgers. However, our good-ole-American hard-working attitude certainly feels cheated by people getting things for free. We could work on training them to become skilled in some task. But what skills will be unsourced next? Another possibility is that we just put our fingers in our ears and pretend that this kind of technology does not, and will not, exist and continue to pay the pittance to the unskilled labor we have.

I'm not sure how this will be resolved, but I certainly think it's a very interesting point of discussion. We should probably talk about it now before the Burger-o-matic 1.0 is released.

Labels: unsourcing

There's a lot of argument about "Outsourcing", the process of farming out work to other providers, frequently in other countries. There's a lot of blame tossed around about this. But I want to suggest that not all of those jobs arrive. Some of them are Unsourced, and simply don't need to be done any more.

I'm very interested in productivity. Not in achieving it, but in the consequences. Suppose you and 100 other people have a given job, and by various tricks, your industry (of 100 people) figures out how to save 5% of its time. If demand doesn't go up, congratulations, 5 people are gonna lose their jobs after the market consolidates.

I like to call this Unsourcing. No one's getting the job, it just doesn't exist anymore. This is what productivity is all about - not hiring people, and firing them if you can get away with it.

This brings me to a recent TechCrunch article:

Engineers Are The Best Deal - So Stock Up On Them
Software engineers today are about 200-400% more productive than software engineers were 10 years ago because of open source software, better programming tools, common libraries, easier access to information, better education, and other factors. This means that one engineer today can do what 3-5 people did in 1999!

As an engineer, I get paid to increase productivity for some group of people. Sometimes those people are engineers as well. My job is killing your job.

Thus far, things have generally worked out; and new labor replaces old labor. But robots throw a wrinkle into blue collar jobs, and software reduces the need for white collar jobs. I don't know where this is going, but I'm pretty worried about it.

Edit: Looks like ComputerWeekly.com beat me to this term in 2001. I think they were somewhat uncritical of the terminology at the time, which as stated was essentially a euphemism for "firing people". I stick by my usage here.

Labels: unsourcing

Today I'd like to talk about two papers of moderate length on the subject of Engineered Intelligence. Reading both of them should take you an hour, maybe an hour and a half. Both are frequently cited, both discuss the relationship between intelligence and computation. And both are in the canon of the future.

If you are interested in Artificial Intelligence, or Augmented Intelligence, or Machine Learning, or even The Singularity; you need to read these two papers. They are short, well written, and accessible. Both papers are frequently cited, so many people feel that they are "familiar" with the papers. I'd like to convince you that these are both well worth your time, and I guarantee you'll learn a thing or two.

The first paper is "Computing machinery and intelligence", by Alan Turing. It was published in the journal Mind in 1950, but there are many copies of it bouncing around the net. Here are a few links:

• A PDF Scan of the Original Paper
• An HTML Translation on a tacky website with tables from the paper
• An HTML Translation with no tables
  

Alan Turing was a researcher in computation. He was involved in a great deal of the early work on what it meant to compute things, and what was and wasn't possible.The British government forced him to take hormones (in the 50's!) to make him more "manly", in the hopes they could turn him straight, and it seems likely that this was a major cause of his suicide. We'll never know how much more we missed because of this government sanctioned homophobia.

"Computing machinery and intelligence" is a wonderfully written and well articulated work on what we mean when we say "intelligence", and what it would mean for a computer to be "intelligent". It is also the source for the silly game called the "Turing test", which Turing himself only introduced to make a rather nice point about personal bias. Most of this paper is devoted to the arugments against engineered intelligence, and the various flaws in those arguments.

The second paper for today is "The Coming Technological Singularity: How to Survive in the Post-Human Era" by Vernor Vinge in 1993. This paper is much easier to track down, being a product of the web era:

• Vinge's official copy
• A copy with a nicer font
  

Vernor Vinge is a professor of computer science and mathematics who decided to pursue science fiction. He's got a good deal of work out now, I highly recommend his novel "Rainbow's End".

Vinge was one of the first author of the modern era to tie the old thoughts about machine intelligence with modern observations on acceleration. Unlike Ray Kurzweil, he isn't a shameless self promoter, so Ray gets more press. But I think Vinge is much more reasonable.

In "The Coming Technological Singularity", Vinge lays out the case for the inevitability of massively disruptive change brought on by engineered intelligence. Like Turing's paper, Vinge devotes most of this article to examining the arguments against engineered intelligence, and the problems with those articles.

Labels: faster is different, things worth reading

The people on this planet use a lot of energy. A lot. Like 500 exajules in 2005. However, more than three quarters comes from fossil fuels. It would be a very safe projection to say that the world's energy demands are only going to increase over time (save the aporkolypse). Here's a graph of the rate of the world's energy usage in terawatts (per Wikipedia):




Kurzweil speculates that most of the world's energy will be produced by cheap solar in the next twenty years. I hope he's right because it'll be awful hard to have a Dyson sphere without the tech to harness all that power. However, in the mean time, it seems that wind power might just serve our current needs.


The Proceedings of the National Academy of the Sciences (PNAS) published an article in April of 2009 that states that wind may produce more than enough electricity to sate the world's hunger. The article (Global potential for wind-generated electricity by Xi Lua, Michael B. McElroya, and Juha Kiviluomac) has the following abstract:

The potential of wind power as a global source of electricity is assessed by using winds derived through assimilation of data from a variety of meteorological sources. The analysis indicates that a network of land-based 2.5-megawatt (MW) turbines restricted to nonforested, ice-free, nonurban areas operating at as little as 20% of their rated capacity could supply >40 times current worldwide consumption of electricity, >5 times total global use of energy in all forms. Resources in the contiguous United States, specifically in the central plains states, could accommodate as much as 16 times total current demand for electricity in the United States. Estimates are also given for quantities of electricity that could be obtained by using a network of 3.6-MW turbines deployed in ocean waters with depths <200m within 50 nautical miles (92.6 km) of closest coastlines.

We definitely need to pursue renewable forms of power, be it solar, wave, geothermal, nuclear, or in this case wind. I think that this result is very, very heartening. Maybe, just maybe, we'll be ok.

Labels: power

This is a post about how Faster is Different. This is a topic I'll return to frequently, but I want to make a point about it up front. Programmers think about the resources available to them in terms of space/time. This isn't the Physicist's space/time; this is about computers.

Computers process stuff (in time), and they store stuff (in space). So, roughly, processors = time, and memory/disk = space. Processors come at different speeds (operations per second), and with different numbers of cores. More of either gives you more time, so much that you can have more than 1second per second of time if you have more than one processor. More memory gives you more space.

This is interesting, because we (programmers) measure how expensive a given algorithm is in terms of space/time; and we measure how hard a given problem is in terms of space/time as well. It might seem strange to make the distinction between algorithms and problems, but sometimes you don't use the "best" algorithm to solve a problem. Generally, a problem is defined to be as hard as the absolute best algorithm that will solve it. If you don't have enough space/time to solve a problem, then you can't. This is the basis for cryptography, very hard problems.

Anyway, back to the main point: Faster is Different.

More space/time doesn't just make your existing stuff faster, it sometimes changes a problem with no solution (because it was too expensive) into a problem with a wildly expensive solution. And problems which are wildly expensive today will be very expensive a year from now, quite expensive 3 years from now, and sorta cheap in 5 years. 10 years from now, they'll be free. All thanks to Moore's Law, and its impact on our future supply on space/time, assuming it doesn't stop. Now, there's reasons to question Moore's Law, but that's for another post.

In this installment of Faster is Different, I want to talk to you about The Swarm. The Swarm is, roughly, everybody and their computers. The Swarm can do a lot of interesting things, and the Swarm is always watching you. Right now, that doesn't really matter, because it doesn't have a lot of space/time available to it. But that's gonna change.

My employer, Google, just posted some interesting research results in computer vision. I wasn't involved in any way, but its germane to the topic of this post:

Official Google Blog: A new landmark in computer vision
Our research builds on the vast number of images on the web, the ability to search those images, and advances in object recognition and clustering techniques. First, we generated a list of landmarks relying on two sources: 40 million GPS-tagged photos (from Picasa and Panoramio) and online tour guide webpages. Next, we found candidate images for each landmark using these sources and Google Image Search, which we then "pruned" using efficient image matching and unsupervised clustering techniques. Finally, we developed a highly efficient indexing system for fast image recognition.

While we've gone a long way towards unlocking the information stored in text on the web, there's still much work to be done unlocking the information stored in pixels. This research demonstrates the feasibility of efficient computer vision techniques based on large, noisy datasets. We expect the insights we've gained will lay a useful foundation for future research in computer vision.

If you're interested to learn more about this research, check out the paper.

Now first, I want to say that this is really neat research. And the math is cool. And the people involved have every reason to be proud of themselves.

Microsoft has an image group which regularly produces equally cool results. A while back, they came up with an application that was so popular, they made it a publicly accessible toy:

Microsoft Live Labs: Photosynth
You can share or relive a vacation destination or explore a distant museum or landmark. With nothing more than a digital camera and some inspiration, you can use Photosynth to transform regular digital photos into a three-dimensional, 360-degree experience. Anybody who sees your synth is put right in your shoes, sharing in your experience, with detail, clarity and scope impossible to achieve in conventional photos or videos.

So, what properties would we expect from the composition of these services? It would find images of commonly photographed sites, and would then abstract them into 3d models of the space, with textures. All of this without sending anyone out to scan or photograph the site. Now, add a few computational generations to this, say 5 years; and people will be doing this for video.

And if you'll recall, pretty much everyone's launched facial recognition software in the past year. Apple did it, Google did it, I lose track. It still isn't great, but its pretty good, and more space/time and samples can only make it better.

But because Faster is Different, what Google and Microsoft can do today, pretty much any company can do in 3-5 years, and small organizations will be able to do in 7-8 years.

So I want you to forget about Big Brother, or Evil Corporations, or The Bad Guys. Over the next 10 years, The Swarm is gonna map the entire world, and it is going to be able to tell, from the pixels alone, who's in any photo on the web, and where it was taken, and probably to some extent, when it was taken. You cannot stop this from happening, because as it gets cheaper, more and more organizations will independently do this themselves. This is inevitable and implied, because:

• People want to know,
• People profit from knowing, and
  
• Faster is Different.

Labels: faster is different, surveillance

This evening I heard a radio DJ (it's like Pandora but a person and not necessarily on the internets) play a song by a band called We Were Promised Jetpacks. Two things: first, I recommend this band; second, here are your damned jetpacks:

Ok, so it's not actually a jetpack, but rather a personal helicopter from Martin Jetpack of New Zealand. There's a nice Wikipedia writeup about them and their product. It seems that the 'jetpack' has a flying time of about 30 minutes and a maximum speed of 60mph. Not too bad, although I wouldn't plan on replacing your Vespa with it just yet, as the price tag is currently around $100,000.

Charles Stross, one of my favorite authors, made an excellent keynote about the future of technology at the LOGIN 2009 conference. Keep in mind, Stross is openly taking a deliberately conservative view, assuming essentially no disruptive technologies.

Key points:

• Expect only one to two orders of magnitude (10x to 100x) improvement in processing power and power consumption.
• Expect ~ 1tb/s wireless data, but no more.
• Expect your mobile phone/computer/thingy to be as powerful as your current desktop.
  

I strongly recommend you read the whole thing:

LOGIN 2009 keynote: gaming in the world of 2030

"Let me give you a handle on this device, the gadget, circa 2020, which has replaced our mobile phones. It's handheld, but about as powerful as a fully loaded workstation today. At it's heart is a multicore CPU delivering probably about the same performance as a quad-core Nehalem, but on under one percent of the power. It'll have several gigabytes of RAM and somewhere between 256Gb and 2Tb of Flash SSD storage. It'll be coupled to a very smart radio chipset: probably a true software-directed radio stack, where encoding and decoding is basically done in real time by a very fast digital signal processor, and it can switch radio protocols entirely in software. It'll be a GPS and digital terrestrial radio receiver and digital TV receiver as well as doing 802.whatever and whatever 4G standard emerges as victor in the upcoming war for WWAN preeminance."

Labels: faster is different

Let's assume that someday, we find some compound which does for cognitive development what steroids do for muscles. For our purposes, it doesn't need to be side effect-free, steroids aren't. Just assume that it exists, call it X.

Now, what does a world with X in it look like?

Do we outlaw it in academia? Do we outlaw it in the workforce?

Suppose we did. There'd still be an underground market in College (if you can't find drugs at your college, you are taking courses online), so students in high-stress fields (or ones that are just real lazy) would use X to get better grades. Once they knew it worked, a very large number of them would, because they'd be competing with other students who were users.

The same way that steroids spread, because if you don't and everyone else does, you're out of the game.

And once this generation of X using students heads to the workplace, where doing better than your peers amounts to $$, they are not going to stop, so you get X showing up more and more in the workplace. And just like sports teams, companies say they want to fight it, but what they mean is they don't want to get caught allowing it, because the benefits to them are enormous.

None of this is as scary as when people who run Gold Farms in the 3rd world discover X, and start dosing their $1/day workers, and then lining up outsourced work. Call them Brain Farms. They won't stop with a little X, they will push deep into cognitive enhancement drugs, and they won't care about the side effects, and neither will their workers, who will be rich compared to their neighbors.

By the way, this has already started. The drug's name is Adderall, and it is already on the upswing in college. The business world isn't far behind, because those kids are graduating.

Labels: cognition, drugs, productivity

This first came to my attention last year from a BBC article (The man who grew back his finger tip). It seems that this has even caught the attention of Oprah:


Basically, Lee Spievack is accidentally sliced his fingertip off. Instead of having a skin-graft he called his brother, who just so happens to be a doctor who studies tissue regeneration. Long story short, by sprinkling the cells from a pig bladder an extracellular matrix was formed and his fingertip regrew.



Now, it must be added that Ben Goldacre who writes the Bad Science column of The Guardian and also Bad Science the book, says that this was no miracle, but was basically just the way the body normally heals itself. The full article is: The amazing disappearing reappearing finger.

I'm not a medical doctor, so I cannot offer an opinion. All I can say, really, is that it would be super-freaking-awesome if Mr. Goldacre was wrong.

Labels: biochem

Festo produced the robotic penguins. One model flies, the other swims. In addition to this, they've built the robotic jellyfish and robotic ray.

I don't know about you, but this makes me want to get a couple roles of mylar and build my own flying robotic animals.

Also, if you're interested, Festo has US operations:

Labels: robots