Category Archives: Technology

Facebook’s “freakishly accurate” face recognition technology

Posted on by

Facial recognition software is almost as good at identifying people as humans are, thanks to Facebook. The Facebook AI team published a paper last week highlighting their achievements with DeepFace, the company’s unsettlingly precise facial recognition program.

DeepFace can identify faces at a 97.25 percent accuracy level, just slightly worse than the average human score of 97.53 percent, as Technology Review noted. The DeepFace system reduced facial recognition software errors by 25 percent compared to earlier versions of the software, which is a vast improvement.

‘Freakishly’ is one way of putting it.

Source

Why Snapchat is interesting

Posted on by

As usual, danah boyd nails it:

Snapchat offers a different proposition. Everyone gets hung up on how the disappearance of images may (or may not) afford a new kind of privacy. Adults fret about how teens might be using this affordance to share inappropriate (read: sexy) pictures, projecting their own bad habits onto youth. But this is isn’t what makes Snapchat utterly intriguing. What makes Snapchat matter has to do with how it treats attention.

When someone sends you an image/video via Snapchat, they choose how long you get to view the image/video. The underlying message is simple: You’ve got 7 seconds. PAY ATTENTION. And when people do choose to open a Snap, they actually stop what they’re doing and look.

In a digital world where everyone’s flicking through headshots, images, and text without processing any of it, Snapchat asks you to stand still and pay attention to the gift that someone in your network just gave you. As a result, I watch teens choose not to open a Snap the moment they get it because they want to wait for the moment when they can appreciate whatever is behind that closed door. And when they do, I watch them tune out everything else and just concentrate on what’s in front of them. Rather than serving as yet-another distraction, Snapchat invites focus.

Furthermore, in an ecosystem where people “favorite” or “like” content that is inherently unlikeable just to acknowledge that they’ve consumed it, Snapchat simply notifies the creator when the receiver opens it up. This is such a subtle but beautiful way of embedding recognition into the system. Sometimes, a direct response is necessary. Sometimes, we need nothing more than a simple nod, a way of signaling acknowledgement. And that’s precisely why the small little “opened” note will bring a smile to someone’s face even if the recipient never said a word.

Snapchat is a reminder that constraints have a social purpose, that there is beauty in simplicity, and that the ephemeral is valuable. There aren’t many services out there that fundamentally question the default logic of social media and, for that, I think that we all need to pay attention to and acknowledge Snapchat’s moves in this ecosystem.

My idea of a perfect blog post. It’s insightful, thought-provoking and beautifully written.

Michael Lewis on Lightspeed

Posted on by

Michael Lewis is, IMHO, one of the best long-form journalists around and his new book is well up to his usual standard. In many ways, it adheres to the classic Lewis formula: find a scandalous set-up of which most people are blissfully unaware; locate some smart guys who have detected the systemic scam and figured out a way to profit from their ingenuity; and then tell their story.

In this case, the story is basically about the speed of light – or, to be more precise, about how the time-difference (in millionths of a second) that it takes an electronic share transaction to traverse one fibre-optic connection rather than another can provide an exceedingly lucrative trading advantage to those who have the kit and the know-how to exploit it.

In the video clip he explains the nub of the idea but, as always, it’s not so much the story as the way Lewis tells it — which is why his book is a must-read for anyone who cares about this stuff.

Writing about it in Quartz, Matt Phillips quotes from another part of the TV interview

“If it wasn’t complicated, it wouldn’t be allowed to happen,” he says. ”The complexity disguises what is happening. If it’s so complicated you can’t understand it, then you can’t question it.”

“This problem”, Phillips says, “goes beyond stock markets: The US financial system is awash in unnecessary complexity. And the reasons are simple: Complexity is profitable and it keeps regulators at bay. ”The jargon of bankers and banking experts is deliberately impenetrable,” wrote economists Anat Admati and Martin Hellwig in their indispensible The Bankers’ New Clothes: What’s Wrong with Banking and What to Do about It. “This impenetrability helps them confuse policy makers and the public.”

There are some echoes of the sub-prime/CDO scandal in Lewis’s new book, in that the people who are supposed to understand how the system worked had little or no idea what was going on under their corporate noses. He recounts how the ‘good’ guys in his tale discovered this when they sought to enlighten leading figures in the financial world about flash trading:

The most sophisticated investors didn’t know what was going on in their own market. Not the big mutual funds, Fidelity and Vanguard. Not the big money-management firms like T. Rowe Price and Capital Group. Not even the most sophisticated hedge funds. The legendary investor David Einhorn, for instance, was shocked; so was Dan Loeb, another prominent hedge-fund manager.

This is an indicator of a really serious underlying problem in our networked world — the stupendous power that superior knowledge, IQ and technical understanding confers on some people. We are completely dependent on systems that are so complex that virtually nobody understands how they work — and how they can be manipulated and gamed by those who do understand them. The obvious rejoinder is “twas ever thus”, but I think that’s too complacent. What’s different now is that the level of technical expertise needed is beyond the reach or capacity of almost everyone. Which means that the elites who do ‘get’ it — and those who employ them — have colossal advantages.

LATER The book has made a BIG impact, to judge from the media coverage, and mostly the reactions have been complimentary. But there were a few contrary opinions. And Andrew Ross Sorkin, writing in the New York Times made some good points.

There is only one problem with Mr. Lewis’s tale: He reserves blame for the wrong villains. He points mostly to the hedge funds and investment banks engaged in high-frequency trading.

But Mr. Lewis seemingly glosses over the real black hats: the big stock exchanges, which are enabling — and profiting handsomely — from the extra-fast access they are providing to certain investors.

While the big Wall Street banks may have invented high-speed trading, it has gained widespread use because it has been encouraged by stock markets like the New York Stock Exchange, Nasdaq and Bats, an electronic exchange that was a pioneer in this area. These exchanges don’t just passively allow certain investors to connect to their systems. They have created systems and pricing tiers specifically for high-speed trading. They are charging higher rates for faster speeds and more data for select clients. The more you pay, the faster you trade.

That is the real problem: The exchanges have a financial incentive to create an uneven playing field.

Footnote: Readers on IoS devices may not be able to see the video clip, for reasons best known to the late Steve Jobs.

Snooping is a public health issue

Posted on by

This morning’s Observer column.

One of the things that baffles me is why more people are not alarmed by what Edward Snowden has been telling us about the scale and intrusiveness of internet surveillance. My hunch is that this is partly because – strangely – people can’t relate the revelations to things they personally understand.

In the past two weeks, two perceptive commentators have been trying to break through this barrier. One is Cory Doctorow, the science-fiction novelist, who had a terrific essay in the Guardian arguing that instead of increasing our security, government agencies such as the NSA, GCHQ and others are actually undermining it. The essay is worth reading in full, but one part of it stood out for me. It’s about the thriving, underworld online market in malicious software. Nowadays, if some hacker discovers a previously unknown vulnerability in widely used software, that discovery can be very valuable – and people will pay large sums for such “zero-day” exploits. But here’s the creepy bit: sometimes, the purchasers are government agencies that buy these pieces of malware to use as weapons against their enemies.

To most people, this will seem pretty abstruse. But with the imaginative skill of a good writer, Doctorow nails it: “If you discovered,” he writes, “that your government was more interested in weaponising typhus than they were in curing it, you would demand that your government treat your water supply with the gravitas and seriousness that it is due.”

Read on

LATER: Right on cue, another great blog post by Bruce Schneier, putting this stuff in an everyday context:

Imagine that you hired a private detective to eavesdrop on a subject. That detective would plant a bug in that subject’s home, office, and car. He would eavesdrop on his computer. He would listen in on that subject’s conversations, both face to face and remotely, and you would get a report on what was said in those conversations. (This is what President Obama repeatedly reassures us isn’t happening with our phone calls. But am I the only one who finds it suspicious that he always uses very specific words? “The NSA is not listening in on your phone calls.” This leaves open the possibility that the NSA is recording, transcribing, and analyzing your phone calls — and very occasionally reading them. This is far more likely to be true, and something a pedantically minded president could claim he wasn’t lying about.)

Now imagine that you asked that same private detective to put a subject under constant surveillance. You would get a different report, one that included things like where he went, what he did, who he spoke to — and for how long — who he wrote to, what he read, and what he purchased. This is all metadata, data we know the NSA is collecting. So when the president says that it’s only metadata, what you should really hear is that we’re all under constant and ubiquitous surveillance.

What’s missing from much of the discussion about the NSA’s activities is what they’re doing with all of this surveillance data. The newspapers focus on what’s being collected, not on how it’s being analyzed — with the singular exception of the Washington Post story on cell phone location collection. By their nature, cell phones are tracking devices. For a network to connect calls, it needs to know which cell the phone is located in. In an urban area, this narrows a phone’s location to a few blocks. GPS data, transmitted across the network by far too many apps, locates a phone even more precisely. Collecting this data in bulk, which is what the NSA does, effectively puts everyone under physical surveillance.

This is new. Police could always tail a suspect, but now they can tail everyone — suspect or not. And once they’re able to do that, they can perform analyses that weren’t otherwise possible. The Washington Post reported two examples. One, you can look for pairs of phones that move toward each other, turn off for an hour or so, and then turn themselves back on while moving away from each other. In other words, you can look for secret meetings. Two, you can locate specific phones of interest and then look for other phones that move geographically in synch with those phones. In other words, you can look for someone physically tailing someone else. I’m sure there are dozens of other clever analyses you can perform with a database like this. We need more researchers thinking about the possibilities. I can assure you that the world’s intelligence agencies are conducting this research.

Schneier is one of the very best commentators on this stuff. Everything he writes about it is worth reading.

Why the obsession with “coding” misses the point

Posted on by

My relatively mild column about the Year of Code fiasco has generated a fair amount of comment, and a good many emails, including some from friends who think I was too willing to give the YoC crowd the benefit of the doubt, and citing Andrew Orlowski’s characteristically caustic take on the matter.

Leaving aside the motives of those involved in the ‘initiative’, a bigger concern (for me at any rate) is that the obsession with “coding” has two significant downsides:

  • it misses the point of the new school curriculum (or which more in a minute); and,
  • it risks alienating the audience that the initiative urgently needs to convince — schoolteachers who are not techies and are probably very nervous about what lies ahead for them as they come to terms with the new computing curriculum.

In her disastrous Newsnight interview, Lottie Dexter (and indeed her tormentor, Paxman) both seemed to think that the only motivations for the ‘coding’ initiative are utilitarian and economic: it was, they seemed to think, about kids being able to get jobs, start companies and thereby boost the prospects of UK Plc.

It’s nothing of the kind. This is first and foremost about citizenship. Today’s schoolchildren will inherit a world that is largely controlled by computers and software. The choice that faces them is “Program or Be Programmed”, as Douglas Rushkoff puts it in his book of the same title. If we don’t educate them about this stuff, then they will wind up as passive users of powerful black boxes that are designed and controlled by small elites, most of them located abroad.

Preparation for citizenship in this new world requires an understanding of how software works, how it is created and controlled, and how it can be changed. We don’t want them to grow up as technologically clueless as the parliamentarians who are supposed to oversee GCHQ; or indeed as Paxman, who at one stage fell back on the old trope about not having to understand electricity in order to replace a light bulb. (The obvious riposte — that light bulbs don’t decide whether you get a mortgage, monitor your private communications or count your vote – was obviously beyond poor Dexter.)

The other aspect of this is that, while learning to program is desirable, it’s not the most important part — which is about having a good critical understanding of the technology. And much though I love Raspberry Pi, teachers can achieve a lot of what I would like to achieve without ever touching a piece of hardware — as the wonderful Computer Science Unplugged project in New Zealand demonstrates.

Finished that ebook yet? Hang on…

Posted on by

This morning’s Observer column.

A few weeks ago I bought a copy of The Second Machine Age by two MIT researchers, Erik Brynjolfsson and Andrew McAfee, who are two of the most insightful commentators currently writing about the likely impact on employment of advanced robotics, machine learning and big-data analytics. Since I already own more physical books than my house and office can hold, I tend now to buy the Kindle version of texts that are relevant to my work, and so it was with the Brynjolfsson and McAfee volume.

Yesterday, I received a cheery email from Amazon. “Hello John Naughton,” it read. “An updated version of your past Kindle purchase of The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies by Erik Brynjolfsson is now available. The updated version contains the following changes: Improved formatting for readability. Significant editorial changes have been made. You can receive the improved versions of all your books by opting in to receive book updates automatically.”

Note the phrase, “significant editorial changes have been made”…

Read on

G2Z is out in the US

Posted on by

Quercus_ad

My book has just come out in the US. There’s a generous review of it by Cory Doctorow on BoingBoing in which he says that the book

sums up the big, important effects that the Internet has in a very quick read, placing them in historical perspective, projecting to their plausible futures, warning of their imminent dangers. From copyright to collective action, from governance to ecommerce, Naughton’s book sets out, in reasonable, measured tones, the systemic underpinnings of the net’s disruptive power, and promises attentive readers the theoretical and practical grounding they need to separate hype from hope.