🔮 Augmenting humans, securing the planet; exposing biased algorithms; clever birds, children’s books and Kremlin tactics++ #148
|Azeem Azhar||Jan 14, 2018|
IoT needs regulation. Algorithms and stereotypes. Pseudoscience. Augmenting humans, work, and social services. Cross country in a Tesla 3. Birds and fire.
Best enjoyed with a lovely mug of tea or a hot cup of chai.
**🙌 **Supported by our partner OnePlus.
DEPT OF THE NEAR FUTURE
👾 The internet of things is a planet-sized robot. Can we control it, asks the cyber-security expert, Bruce Schneier?
Markets alone can’t solve our security problems. Markets are motivated by profit and short-term goals at the expense of society. They can’t solve collective-action problems. […] And we need a counterbalancing force to corporate power.
Any policy changes to secure this world-size robot will mean significant government regulation.
🌈 Blaise Aguera y Arcas: Do algorithms reveal sexual orientation or just expose our stereotypes? **EXCELLENT ESSAY **(Also, Google ‘fixed' its ‘racist photo algorithm’ with a kludge: deleting the label ‘gorilla’.)
🔨 Early Facebook adviser, Roger McNamee: How to fix Facebook before it fixes us. (Facebook just announced it would prioritise stories from family and friends ahead of that from brands and publishers. Jeff Jarvis is concerned that “Facebook has become the primary distributor of news and information in society, it cannot abrogate its responsibility.” Indeed, Joshua Benton is concerned that readers won't look for news if they can't find it on Facebook, and has data to back it up.)
🖇️ The new union? Ride-hailing drivers are banding together to better understand the plans, policies and behaviour of their gig platforms.
🚴 As the line between public and private transportation blurs, how will city planners get access to the travel data they need to understand city life? (A good example of where regulatory-led intervention has taken tentative steps towards a 'commons' that will benefit consumers is OpenBanking, which launched in the UK this week.)
DEPT OF FAKE
If people don't trust experts (or have had 'enough of them') and Facebook's turn away from news distribution further limits the distribution of information, where will well-critiqued expertise find its audience? Do we run a risk of the fake becoming the benchmark? Could our faith in the process of science be replaced by something worse? My sense is yes. And I'm increasingly concerned that we need to get in front of this issue.
Philosopher Thomas Kuhn grasped the realpolitik of science. Steven Fuller, writing in Kuhn vs Popper, explains how Kuhn understood that:
scientific revolutions succeed not because the same people are persuaded of a new way of seeing things (à la Popper) but because different people’s views start to count [...] The sheer fact that newcomers have not yet personally invested in the old paradigm may be enough to make them open to a radical change in direction.
Fake news, identity-driven beliefs, take aim at the notion of falsifiability, which Karl Popper argues was a core ethic of science.
Let’s guard against this, not to prevent a new and better paradigm emerging—like the Copernican replacing the Ptolemaic—but to prevent a turn away from science back towards superstition, the unevidenced and inexplicable.
Fake news and distrust of science could lead to global epidemics.
Oprah Winfrey has long dabbled in junk, pseudoscience.
Raw water is probably another absurd example.
Autism pseudoscience crept into Apple's Webby-nominated ad.
DEPT OF AUGMENTING HUMANS
🥑 Great feature on Ocado’s collaborative robot, which helps engineers repair machines.
🕸 Emil Wallner has trained a neural net that can turn design mockups (basically PNGs) into working code. This could obviate the need for an HTML jockey, but give designers who don’t grok HTML new powers.
How do you become a great surgeon? By being a junior surgeon for years. But what happens when a robosurgeon deprives you of the training? Early evidence suggests that junior surgeons don’t acquire the skills they need to run procedures themselves. (Academic paper.)
Laundroid is a $16,000 laundry-folding robot.
The reality of tweaking social services with algorithms.
Corti.AI helps emergency dispatchers figure out whether someone is having a heart attack or stroke.
Interesting analysis of the growing divide between software and hardware companies tackling machine learning problems. In a nutshell: edge-based inference is going to be hardware limited but hasn’t attracted as much attention as breakthroughs in training models (which is the playground of software engineers wallowing in oodles of cloud capacity.) Illuminating, if nerdy.
Reconstructing images by applying deep neural networks to human brain activity as measured by fMRI. (Video | article)
Psychedelic stickers that confuse image recognition systems.
SHORT MORSELS TO APPEAR SMART AT DINNER PARTIES
🇺🇸 EV reader Alex Roy just drove a Tesla 3 across the US of A. Here is his review.
The Kremlin’s 20 political warfare tactics.
IBM announced a 49 qubit chip at CES. What is the status of quantum hardware?
🤳 Facebook tracks you by the pattern of dust on your camera lens.
Two large investors have challenged Apple over iPhone addiction.
📚 What is the difference between kids books in China and those in America? (Guess before you click.)
Is self-improvement enough to achieve the singularity?
Hydro-Quebec is looking to attract cryptocurrency miners to use its excess clean energy. (Free electricity is a powerful motivator, as this photo essay of dorm room crypto miners shows.)
🐬 Dolphins show self-recognition earlier than humans
EV reader Michael Keating's electric scooter-sharing company is adding new types of electric vehicles in San Francisco and Barcelona. It's an exciting time to join a company like Scoot, so I hope you don't hesitate to apply to one of the open positions:
*Community Careers is a paid space within the newsletter, reserved for long-term readers.
I'm enjoying AlphaGo documentary on Netflix. It deserves its 100% Rotten Tomatoes rating. It's really compelling.
And if you have three hours to spare, Anil Seth, a long-time friend of Exponential View, gets into a superb conversation with Sam Harris on consciousness and the self. Anil is a fantastic communicator on the very complex questions of consciousness. I'm 90 minutes through this discussion and it hasn't disappointed.
Hope 2018 is treating you well!
P.S. There's a small, but meaningful, thing you can do this Sunday to fight pseudo-truth. Please take a moment to share EV with friends on Twitter, Facebook & LinkedIn! (Or, better still, forward this by email.)
This week's issue is brought to you with the support of our partner, OnePlus.
"The OnePlus 5T is the best Android phone you can buy at any price."
- Business Insider