As it plans for its first stock offering, how will Twitter make more money? One idea aired this morning at EmTech,
MIT Technology Review’s annual conference on emerging technologies,
after a talk by Deb Roy, Twitter’s chief media scientist. His MIT
spinout company, Bluefin Labs (see “A Social Media Decoder”), was bought by Twitter this year.
Many
people tweet about exciting moments during televised sports games, but
people seeting the tweets might have missed the action on TV. In
response to a question fromJason Pontin, MIT Technology Review’s publisher (and conference emcee), Roy replied:“In the example I showed you with ESPN and the basketball clip, a very natural
thing one could do is insert a pre-roll piece of advertising [so that]
along with a clip you just missed on TV, there could be a piece of
sponsorship. That could be revenue shared between ESPN and Twitter.”
Already Twitter has been forging agreements under its Amplify program, which lets content owners distribute programming in Twitter feeds with ads embedded. Last month Reuters reported that the NFL was making a similar-sounding deal with Twitter, providing sponsored Tweets with sports highlights.
Presumably, as this strategy
evolves, Twitter users could start to see very specific sponsored tweets
closely married to a piece of TV content that someone they followed
just tweeted about. The
targeting of such sponsored tweets could be further refined using
insights about the users’ interests, gleaned from his tweets, the
profiles of people he follows, and other sources (see “A Taste of Future Twitter Technologies”). But that’s just an educated guess. Twitter is in its pre-IPO quiet period, and we’ll have to wait and see what happens.
Since last year, Duolingo has offered free online lessons in a smattering of languages—currently, just six of them, including English, Spanish, and French—through
which users learn while also translating texts on the Web. The company
often gets requests for more languages, but it’s time-consuming and
costly to add each new one, and many of the ones users want won’t bring
in much revenue, which comes from companies paying Duolingo for its
text-translation services.
Now Duolingo is hoping to quickly ramp up the number of courses it offers without breaking the bank by rolling out a language incubator
that allows volunteers to add all kinds of new tongues, ranging from
widely-spoken ones like Arabic and Chinese to lesser-known ones like Kichwa to fictional ones like Elvish and Dothraki, that other users can then learn for free.
Duolingo cocreator and CEO Luis von Ahn
says that Duolingo will allow people to apply to serve as volunteer
moderators for each language, and the startup will choose a moderator
who can then choose others to help contribute to the language course.
Based on the company’s user base, which includes 10 million people—26
percent of them in the U.S.—the site is likely to have plenty of
volunteers.
In theory, this could make it much faster to add each language. Von
Ahn says it currently takes one paid person four months to add each
language to Duolingo (a course, he says, is meant to give a user an
intermediate-to-high understanding of a given language). Yet he suspects
one or two moderators, working with five or six volunteers, may be able
to complete a new language course in about a month. He expects about 10
languages, including Chinese and Russian, to be added within the first
four or five months of the incubator’s life.
I’d guess it will be hard to maintain quality, at least initially, though it’s certainly possible that, similar to Wikipedia’s
slow rise as an acceptable knowledge source, Duolingo’s crowd-sourced
language Incubator will become an acceptable way to learn. Von Ahn
obviously can’t know for sure how well it will work at this point, but
he says Duolingo will be able to measure factors like how much of each
course users complete, and if they return to the site.
He says the
moderators will use Duolingo’s existing language-course blueprint,
which starts with some basic words, concepts, and phrases and moves on
from there. He says that while the languages Duolingo staff have added
to the site include exercises where users translate online texts, this
won’t initially be a part of the crowdsourced language courses.
Back
in 2008, Google launched its now famous flu trends website. It works on
the hypothesis that people make more flu-related search queries when
they are suffering from the illness than when they are healthy. So
counting the number of flu-related search queries in a given country
gives a good indication of how the virus is spreading.
The predictions are pretty good. The data generally closely matches
that produced by government organisations such as the Centers for
Disease Control and Prevention (CDC) in the US. Indeed, in some cases,
it has been able to spot an incipient epidemic more than a week before
the CDC.
This year’s Nobel Prize in Chemistry goes to three researchers who developed powerful computer models that have been used, among other things, to model combustion.
Computer
models have become essential to improving the efficiency and emissions
of internal combustion engines, gas turbines, and other equipment that
depends on combustion. For example, they allow researchers to make
subtle changes to the shape of pistons to ensure that fuel and air mix
properly and burn cleanly, and at just the right rate. Or to understand
how it’s possible to enhance efficiency by burning more than one fuel
(see “Exploding Engine Could Reduce Fuel Consumption” and “Swiss Researchers Make an 80-mpg Hybrid”).
The
awardees are Martin Karplus, of Université de Strasbourg and Harvard
University, Michael Levitt from Stanford University, and Arieh Warshel
from the University of Southern California.
Honda
says it has found a way to improve the efficiency of hybrids—and in all
likelihood significantly lower the cost to build them. For its new
Honda Accord Hybrid, which goes on sale across the U.S. at the end of
the month, it’s done away with the car’s transmission.
Hybrids
could help automakers meet fuel economy regulations, but their sales
are limited by high costs, which are as much as $5,000 more than
conventional cars. It may eventually be possible to reduce the price gap
to zero, according to a spokesperson from Honda.
A new antenna design being tested by the U.S. Air Force could make
GPS significantly more reliable and able to function in dense urban
areas where GPS accuracy is weak. It might even allow the technology to
work indoors in some cases.
Good GPS readings are hard to get in
cities because of the multipath phenomenon: signals from positioning
satellites bounce off buildings and other structures. That confuses GPS
receivers, which calculate their location by knowing exactly how long it
took for signals to arrive from satellites overhead.
In 2011, General Michael Hayden, who had earlier been director of
both the National Security Agency and the Central Intelligence Agency,
described the idea of computer hardware with hidden “backdoors” planted
by an enemy as “the problem from hell.” This month, news reports based
on leaked documents said that the NSA itself has used that tactic,
working with U.S. companies to insert secret backdoors into chips and
other hardware to aid its surveillance efforts.
That revelation particularly concerned security experts because
Hayden’s assessment is widely held to be true. Compromised hardware is
difficult, and often impossible, to detect. Hardware can do things such
as access data in ways invisible to the software on a computer, even
security software. The possibility that computer hardware in use around
the world might be littered with NSA backdoors raises the prospect that
other nations’ agencies are doing the same thing, or that groups other
than the NSA might find and exploit the NSA’s backdoors. Critics of the
NSA say the untraceable nature of hardware flaws, and the potential for
building them into many systems, also increases the risk that
intelligence agencies that place them will be tempted to exceed legal
restrictions on surveillance.
