Monthly Archives: November 2019
Linkblog
- What’s behind a scary number? Nice analysis by Quentin Stafford-Fraser of alarming estimates about the electricity-generation implications of electric cars.
- The Market for Bulletproof Vehicles Is Skyrocketing Hmmm… I wonder why? Could it be anything to do with the fears of the super-rich about the consequences of rising inequality? These vehicles will essentially be mobile gated communities.
- The internet is getting less free Election interference and government surveillance on social media are hurting internet freedoms.
- Human Rights and the Grenfell Tower Inquiry Expert critique by Aoife Nolan.
The view from here
From the 4th floor of the University Library where I’m happily occupied reading a collection of essays by the great historian of computing Michael Sean Mahoney.
King’s College chapel in the centre of the frame.
Linkblog
- A Simple Combinatorial Model of World Economic History I bet that W. Brian Arthur will not be surprised.
- Queen’s dresser tells all Including what she said to Daniel Craig.
- The best books on Charles de Gaulle and the French Resistance Chosen by Jonathan Fenby, who was once my Editor on the Observer.
- Technology and rock-climbing Or, the importance of rubber. Fascinating, even to those of us who suffer from vertigo.
Right diagnosis, wrong remedy
Robert Reich is right about one thing:
A major characteristic of the internet goes by the fancy term “disintermediation”. Put simply, it means sellers are linked directly to customers with no need for middlemen.
Amazon eliminates the need for retailers. Online investing eliminates the need for stock brokers. Travel agents and real estate brokers are obsolete. At a keystroke, consumers get all the information they need.
But democracy can’t be disintermediated. We’re not just buyers and sellers. We’re citizens who need to know what’s happening around us in order to exercise our right to self-government, and responsibility for it.
If a president and his enablers are peddling vicious and dangerous lies, we need reliable intermediaries that help us see them.
The problem is we have a president who will say anything to preserve his power, and two giant entities that spread his lies uncritically, like global-sized bullhorns.
We can’t do anything about Trump until election day or until he’s convicted of an impeachable offense. But we can and should take action against the power of these two super-enablers. If they’re unwilling to protect the public against powerful lies, they shouldn’t have as much power to spread them.
And his solution? Use antitrust law to break up Facebook and Twitter.
That’s not going to solve the problem. And even if it did, Trump would be into his fifth term before break-up was accomplished.
Linkblog
- Sir Tim Berners-Lee’s Inrupt startup receives £5m investment
- Rosemary Hill on Auberon Waugh Lovely LRB piece.
- Remembering Edward Shils Nice memoir by his friend Joseph Epstein.
- Ten of the most important papers published in Nature Includes monoclonal antibodies, the structure of DNA, the first exoplanet and the discovery of the Antarctic Ozone hole.
Linkblog
- How Trump Reshaped the Presidency in Over 11,000 Tweets Funny what you can do with 140 characters.
- Russia just brought in a law to try to disconnect its internet from the rest of the world ‘Balkanisation’ gathers pace.
- MIT president acknowledges women, minorities on campus feel belittled, excluded Boston Globe report. The surprising thing is that anyone who knows MIT would be surprised.
- Newton vs the machine: solving the chaotic three-body problem using deep neural networks Since Newton’s time the problem of solving the equations of motion for three bodies under their own gravitational force has remained practically unsolved, and even partial solutions require massive brute-force computing. But four researchers at the University of Edinburgh have trained a neural network which provides accurate solutions at a fixed computational cost and up to 100 million times faster than a state-of-the-art conventional solver. Interesting because it suggests a novel way of using machine learning.
The Liberal failure
From Dave Winer:
Just thinking out loud here. I am sure there’s a new journalism out there, that it’s not the journalism that gets so much acclaim, the reinvention of Woodward and Bernstein, the two Washington Post innovators who brought down Nixon. We should be way ahead of that by now. We need to be, because the forces opposing democracy, the equivalent of 1974’s plumbers, are moving much faster. We’re erecting Maginot Lines now, getting ready to fight the Battle of 2016, ignoring that the enemy already controls our capital. They’ve been innovating. We haven’t seen the results of their most recent innovations, yet.
Yep.
How “Don’t Be Evil” panned out
My Observer review of Rana Foroohar’s new book about the tech giants and their implications for our world.
“Don’t be evil” was the mantra of the co-founders of Google, Sergey Brin and Larry Page, the graduate students who, in the late 1990s, had invented a groundbreaking way of searching the web. At the time, one of the things the duo believed to be evil was advertising. There’s no reason to doubt their initial sincerity on this matter, but when the slogan was included in the prospectus for their company’s flotation in 2004 one began to wonder what they were smoking. Were they really naive enough to believe that one could run a public company on a policy of ethical purity?
The problem was that purity requires a business model to support it and in 2000 the venture capitalists who had invested in Google pointed out to the boys that they didn’t have one. So they invented a model that involved harvesting users’ data to enable targeted advertising. And in the four years between that capitulation to reality and the flotation, Google’s revenues increased by nearly 3,590%. That kind of money talks.
Sign up for Bookmarks: discover new books in our weekly email
Read moreRana Foroohar has adopted the Google mantra as the title for her masterful critique of the tech giants that now dominate our world…
What if AI could write like Hemingway?
This morning’s Observer column:
Last February, OpenAI, an artificial intelligence research group based in San Francisco, announced that it has been training an AI language model called GPT-2, and that it now “generates coherent paragraphs of text, achieves state-of-the-art performance on many language-modelling benchmarks, and performs rudimentary reading comprehension, machine translation, question answering, and summarisation – all without task-specific training”.
If true, this would be a big deal…