Exponential View

Azeem Azhar’s Weekly Wondermissive: Future, Tech & Society

This issue has been supported by our partner: Ocean Protocol.

Searching for the right dataset can be like looking for a needle in a haystack. With the right metadata, this search is optimized. Learn how metadata works on Ocean Protocol.

You are reading the free edition of the Exponential View newsletter

We aim to make Premium membership exceptional. Here’s what you’ll get next week if you subscribe today:

• Exclusive audio excerpt from my conversation with Shoshana Zuboff, the leading thinker on surveillance capitalism, in which we discuss the role of Apple in the surveillance economy;
• Access to my upcoming State of the Exponential live webinar briefing with Karen Stenner, whose research on authoritarianism is more important than ever;
• Unedited transcripts of my podcast conversations with Joanna Bryson, Cathie Wood, Mariana Mazzucato, Eric Topol and more.

Subscribe now

Dept of the near future

🗯️ Mary Meeker produces her 22nd Internet Trends report, more than 300-slides worth. Some highlights:

• It’s getting harder to reach new internet users,
• Global smartphone shipments fell by 4% in 2018, compared to 0% in 2017,
• 53% of the world’s internet users are located in the Asia-Pacific region,
• The average daily time consumers spend on mobile in minutes surpassed TV in 2019: 226 minutes for mobile versus 216 minutes for TV.

Report recap & highlights available here.

💱 Facebook’s new cryptocurrency, Libra, has attracted the likes of Visa, MasterCard, Farfetch, Kiva and Uber as partners. Each firm will invest $10m and some may run ‘nodes’ along the system that verifies transactions and maintains records of them, creating a brand-new payments network. There are excellent details of all of Libra’s partners here. Preston Byrne, an astute observer of the cryptospace points out that ‘these entities are each paying $10 million for the privilege of running a node aka unlimited access to transactional data. What this announcement actually means is that now Visa, MC, PayPal and Facebook are each getting unfettered access to users’ data.’ GlobalCoin will be a ‘stablecoin’ backed by a basket of fiat currencies, which has been lauded as ‘promising news in its design’ and ‘very clever’. (Separately, Facebook has found emails that suggest Zuckerberg might be connected to many of the firm’s problematic privacy practices.)

💨 Uber’s market dominance has been based in artificial narratives of ‘cutting-edge technology’, market overhaul and lying about working conditions and pay, argues transport analyst, Hubert Horan.

Uber’s investors understood that this market power allows Amazon and Facebook to economically squeeze workers, suppliers, and cities, to manipulate the data presented to consumers, to force users to cede control of their personal data, and to expand into other markets where they have no advantages other than platform ubiquity. Uber has exploited some of this artificial market power against its drivers. But even with billions in predatory subsidies, it could never achieve the dominance and immunity from competition needed to exploit market power on a scale similar to Amazon and Facebook.

📺 The root cause of radicalisation is not technology, but sociological and political circumstances, including searching for identity and belonging in the time of crisis.

👛 Daron Acemoglu: UBI is a simplistic solution to a complex problem. Better would be to provide a guaranteed income floor through negative taxes. We need to recognise that modern welfare is the result of a democratic ballot and not simply an exercise in ‘panem et circenses’ to pacfy a population. Worthy of a read.

Climate calamity: 414.71ppm | 3,993 days

Each week, we’re going to remind you of the CO2 levels in the atmosphere.

The latest measurement (as of June 13): 414.71 ppm; 12 months ago: 409ppm; 50 years ago: 326.66ppm; 250 years ago, est: 250ppm. Share this reminder with your community by forwarding this email or tweeting this.

I’ve introduced a rough countdown to reaching the 450ppm threshold, inspired by EV reader Gavin Stark. I started a 4,000-day countdown on June 10th, which means the counter now shows 3,993 days.

Here is a wonderful survey of how machine learning could be practically applied to tackling climate change. It is 97-pages long and I’ve not read all of it.

🇬🇧 Outgoing British prime minister, Theresa May, announced that the UK would be the first major economy to shoot for net zero carbon emissions, including aviation and transport, by 2050. My initial reaction was that 2050 remains far too late for an advanced economy to transition to net zero. But on reflection, I realise quite how important this could be. Here is a readable Q&A on the topic, but do note that Theresa May did not call for a global carbon tax.

It will require massive changes (and hence investment) to achieve them, and that creates opportunity. The deadline of 2050 will impact investment in non-renewable energy programmes in a couple of years. These investments are often made over multi-decadal timelines.

This makes other technologies politically and economically more appealing. Many underlying technologies that will see exponential improvements in their capabilities over the coming years. What seems intractable today (finding a mix of renewable energy and storage that competes with natural gas electricity generation) is on the cusp of being financially viable, especially if you net off fossil fuel subsidies. (See my discussions with Bill Gross on this point.)

What was missing was a compelling vision of what a net-zero world would look like. A net-zero world could, should, mean cleaner, more livable cities that have evolved to the needs of people rather than vehicles. More breathable air, more walking, cycling, and human interaction. It could mean more wilderness, more biodiversity, healthier oceans, more blue in our pale blue dot. It might also mean the presence of new values, those of stewardship, attention and care rather than those of extraction, enclosure and profit-at-all-costs.

Let us be grateful for what was present in the announcement: an acknowledgement of the scale of the climate crisis. Considering the hapless state of the British government and its governing party, that is something to be happy with.

🕳️ A senior member of the Commodity Futures Trading Commission, a US financial regulator, said that the financial risks from climate change were comparable to those of the 2008 financial crisis. My gut says it’s worse than the global financial crisis but the intervention of the CFTC might encourage more investment or financial products that address climate change.

Dept of artificial intelligence

💯 I had a really fun conversation with Jürgen Schmidhuber, one of the pioneers of deep neural nets that power applications that billions of us use daily. Have a listen.

Balanced argument by Bruce Schneier and James Waldo that AI and open societies can thrive. ‘Surveillance is not justified by the needs of machine learning, and real progress in AI doesn’t need it.’

Speech2Face was trained on 100,000 educational videos of people talking, learning to create associations between vocal cues and characteristics such as age, gender and ethnicity. It accurately reconstructs photorealistic faces and the three attributes based on voice alone.

😛 Video: Chinese school kids check-in using facial recognition.

‘Sensors and new machine learning models are what allow humans to extend the interpretive capacity of our brains’ argues Gideon Rosenblatt.

A machine vision system was worse at identifying household items from lower-income countries: algorithms were 15-20% better at identifying items from the US than from Burkina Faso or Somalia.

An analysis of RADAR, an automation tool that allows six journalists to produce 8,000 localised news articles per month. It relies on manual tuning and looks a lot like a bunch of nested ‘IF-THEN-ELSE’ statements right now.

😛 For an example of how Transformer, a more sophisticated AI text generation system than RADAR, works, have a play here.

Bank of America’s chatbot has interacted with more than 7 million customers, completing more than 50m requests.

A network of Russian spies uses fake Linkedin accounts and GANs to build a virtual rapport and connect to the important Washington figures.

Short morsels to appear smart at dinner parties

Pupils at an elite primary school in Hanghzhou wear brainwave-detecting headbands that allow teachers to track concentration levels in real-time.

💢 Walmart’s Jet.com was touted as a masterstroke to digitise a sluggish incumbent. But the strategy didn’t seem to work.

Amazon replaces Google and Apple as the world’s most valuable brand.

👶🤳 Snapchat’s toddler and gender-swap filters appeared to have doubled app downloads.

🛰️ A GPS outage would cost the world economy about $1bn a day. (GPS has generated $1.4trillion in benefits for the private sector. Not bad for a state-funded innovation!)

😠 Saudi Arabia uses military-grade technology to track around 1,000 women who flee the country every year.

A Princeton team found a way of controlling Majorana quasi-particles. (The Majorana is a particle which is its own anti-particle. Microsoft’s quantum computing efforts rely on manipulating Majoranas.)

A German flying taxi startup announces plans to hire ‘hundreds of software engineers’ at a new London base.

🦠 Evidence that some gut microbes can eat our medication.

A further study highlights the mortality risks of eating red meat.

🧠 Nice survey of animal cognition.

End note

This week I was helping out at Cogx, a conference on artificial intelligence.

It is a fantastic and stimulating opportunity to meet people who understand how to create a better world. And there are a lot of interesting things to be learned there.
My aim here was to get to know the world and to give you a clear perspective on what’s possible in AI and how we can use it to create new and better ways of working together. And in the process I’m really encouraged by a lot of people to take part in the discussions.

Everything in italics above was automatically generated by Transformer, a dialog model based on OpenAI’s GPT. It has produced some vapid but human text. It doesn’t look like it would be out of place in a corporate presentation. You should also have a play with it here.

If auto-generated text really takes off, I look forward to the next innovation: automatic reading of this algorithmic pabulum and more time on the tennis court!

Cheers,
Azeem

P.S. Scroll down to see what your fellow readers are up to

This issue has been supported by our partner: Ocean Protocol.

Have questions about Ocean? Join their Telegram chat and participate in their monthly AMA. Here is what was covered in June

What you are up to—notes from EV readers:

Joanna Bryson and De Kai on Azeem’s panel discussing Google as the world power and the failed Ethics Council.

Elise Thomas on surveillance technology on the streets of Australia.

Michael Keating’s startup Scoot has been acquired by Bird.

Walter Pasquarelli shares the world’s first Government AI Readiness Index.

John Taysom’s Privitar raises Series B.

Saul Klein’s LocalGlobe just closed two funds totaling $295m.

Salman Farmanfarmaian’s has written an introduction to the personal server paradigm.

Daniel Hulme’s TEDx Talk answers the question: what would a completely new social, economic and political system look like?

Michael Szul released a book Building Chatbots in TypeScript with the Microsoft Bot Framework.

Amit Patel invites you to the 6th edition of A Conversation about Design in London on Friday June 21st, exploring the Brand Strategy in a Digital Age. Get 50% off with discount code ‘exponentialview’.

Codrina Cretu invites you to an event exploring New urban visions: Bringing digital social innovation into City Hall on June 25. Bursary scheme for under-represented groups and countries available.

To share your project and news with the EV community, email marija@exponentialview.co