🔮☀️ The AI economy; achieving optimism; blurring machine & human; future business models; peak car; bumper summer special++ #180
It is a bumper summer issue this week as it is a public holiday on Monday for many readers. Enjoy at your leisure.
Dept of the near future
✋ “The boundaries between body and machine, between mind and world, between standard, augmented and virtual realities, and between human and post-human…[are blurring]” says philosopher, Andy Clark. It will create “world marked more by possibility, fluidity, change and negotiability than by outdated images of fixed natures and capacities.” THOUGHT PROVOKING. (One of my favourite examples of how humans adapt through technology is in Vaclav Smil’s Energy & Civilization where he describes how the human colon is relatively smaller than those of other primates, primarily because our early technologies allowed us to switch to higher-quality, energy-dense foods.)
📈 Alibaba as a model of the future of business: “Alibaba’s special innovation, we realized, was that we were truly building an ecosystem: a community of organisms (businesses and consumers of many types) interacting with one another and the environment (the online platform and the larger offline physical elements). Our strategic imperative was to make sure that the platform provided all the resources, or access to the resources, that an online business would need to succeed, and hence supported the evolution of the ecosystem."
💯 Hedge funder, Steve Cohen, joins the chorus on the power of data network effects and argues AI models will run the world, and that such effects represent "a new, more powerful, business model". (See also, this fascinating framework for AI-first startups, moving from product-market fit to model-market fit and “minimum algorithmic performance” (MAP) instead of minimum viable product. Also see my longer discussion in Dept of the AI Economy, below.)
💈 It isn’t technology that is disrupting our jobs, argues Louis Hyman, rather “[t]he insecure nature of work is a result of decisions by corporations and policymakers.”
💨 We’ve reached peak car and it heralds the end of an industry say Stefan Nicola and Elisabeth Behrmann. A good survey of attitudes to car ownership (down, even in Germany) and the growth of alternative mobility models (think fleet networks) and how incumbent car firms are trying to pivot into them.
🎯 The decimation of industry weakens process knowledge, while “[o]ur obsession with the digital world has pushed our expectation of the technological future in the direction of cyberpunk dystopia”, we need to recapture it in order to have a pragmatic optimism. Brilliant, long essay from Dan Wang which I consider a MUST READ. (h/t Claudia Chwalisz)
Dept of the AI economy
I’ve discussed superstar companies many times before in Exponential View. Now the FT presents more some superb data on the scale and depth of the so-called superstar companies, mostly Internet giants. Amongst other arguments:
the economic weight of a small number of highly profitable and innovative “superstar” companies has increased, workers’ slice of the pie has fallen in their industries
and
increasing risk “that innovation and investment may ultimately fall after an initial spike as industries become highly concentrated and big companies rest on their laurels.
There are clearly policy considerations and not just the risk of excessive market power (and for various reasons most of the activities of these superstar firms fall outside the American framing of antitrust), but also emergent social and political risk from a barbelling economy.
Many of these superstar firms are AI at their heart, taking advantage of the data network effects Steve Cohen refers to above. And they are streets ahead of old economy companies, which were the stars of the past decades.
To help understand how far ahead the Internet superstars who have understood that “models are eating the world”, have a read of the detailed CB Insights overview of the main tendrils of Google’s strategy of turning itself into an AI company.
Elsewhere
Bill Gates reviews one of the books that has most impacted my thinking in the past year, Capitalism with Capital by EV subscribers, Jonathan Haskel and Stian Westlake.
New research suggests the machine translation can boost international trade. Data from eBay suggested that a moderate improvement in the quality of machine translation”–a 7% increase in human acceptance rate–... generated an export increase of 17.5% in quantity and 13.1% in revenue.”
Dept of artificial intelligence
Mariarosaria Taddeo & Luciano Floridi: making a AI a force for good. Excellent framework for understanding how to design the ethics & value systems for an AI future.
This is amazing. The algorithmic gaze: art which depicts the world the way algorithms do.
Nice survey of AI design typologies by Gideon Rosenblatt.
😵 Estimating the gender split amongst AI researchers. (only 12% women)
Many years ago, I used Amazon’s Mechanical Turk extensively to generate training data for classification systems. Turkers have become the de facto standard for producing gold standard data. Now researchers have discovered that bots, rather than humans, might increasingly be doing the work.
👀 DeepMind’s team successfully applies a novel DL architecture in assessments of retinal diseases, training the model on only 14,884 scans.See also from DeepMind, AI autonomously controls Google's data centre cooling.
Super excited about all the innovation in chips for AI. Detailed profile of Softbank-backed Mythic which produces highly-power efficient inference chips to live on the edge of networks. (About 100x more power efficient than a GPU.)
What went wrong with IBM Watson? It was never meant to really work.
Dept of democracy, governance & jurisprudence
We know that cyberspace and the firms in them are blurring the traditional boundaries of national sovereignty, and none more so than in cyber warfare.
Two examples:
This stellar story goes behind the scenes of the Netpetya cyber attack, most likely an assault by Russia’s Sandworm hacker group on Ukrainian infrastructure, which rapidly spread and debilitated large firms like Merck and Maersk, and inflicting an estimated $10bn of damage. This is all about negative spillovers.
And Microsoft took “steps against broadening threats to democracy” when it took control of several fake domains run by Russia’s Fancy Bear outfit designed to attack right-wing politicians in the US. A couple of interesting points: here we have a private corporation assuming very state-like responsibilities. (This is fascinating from a corporate governance perspective, after all, companies have in recent decades been in thrall to Milton Friedman’s paean, and their executives are heavily incentivised by the short-term rise in the stock price not the long-term rather amorphous benefit of protecting democracy.)
Elsewhere:
Russian trolls are sowing vaccine disinformation across twitter.
Understanding Facebook content moderation. Super reportage by Joseph Cox on how the firm is trying to “fix the mess [its] worldview and business model has created, without changing the worldview of business model itself.” Can this circle be squared?
The veteran journalist, Frederic Filloux pulls no punches: Facebook’s flawed DNA makes it unable to fight false information: “Facebook’s DNA is based on the unchallenged power of an exceptional but morally flawed — or at least dangerously immature — leader who sees the world as a gigantic monetization playground.”
Who needs democracy when you have data? Excellent, detailed report looking at how Chinese city, regional and national governments are taking advantage of data and advances in AI to maintain political control. (Interestingly, Chinese teams dominated the 2018 ActivityNet Large-Scale Activity Recognition Challenge. This challenge focuses on the semantic visual understanding of videos, in other words understanding what is happening within video frames, localising actions & describing them etc, essential capabilities for widescale urban surveillance.)
Large-scale incidents of violence in India have been triggered by the circulation of fake messages on WhatsApp. Ravi Shankar Prasad, a senior Indian government minister, is arguing for WhatsApp to take more steps to comply with local laws.
On the subject of law, here is another argument for granting legal rights to rivers, one way of transforming how we see our relationship with the natural world. (We’ve already given rights to other non-persons, in particular, corporations, which essentially grants them new classes of freedoms to act and protect themselves. AI systems and distributed systems may soon also be entities which may need to have some form of legal rights & obligations attached to them. To be discussed further!)
And, finally, a revealing read about the ways poverty is privatised, increasingly with the help of technology.
Short morsels to appear smart at dinner parties
A luxuriant profile of how TripAdvisor, used by 1 in 16 humans every month, has changed travel (mostly for the better, in my view). Great details of its struggles with scale and trust.
Chinese bike startups are reversing out of many Western markets.
Alcohol probably isn't as bad for you as a new Lancet study recently suggested. Critical tear down by David Spiegelhalter.
🏗️ Swiss startup wants to store excess energy using concrete blocks and cranes.
Global off-grid renewable energy capacity tripled between 2008 and 2017.
🥗 Nearly 10% of the Canadian population is vegan or vegetarian. See also: if everyone in the US ate beans instead of beef, the country would meet its greenhouse-gas emissions goals by 2020.
♀️♂️Ubang, a Nigerian village where men and women have different languages.
A female who died around 90,000 years ago was half Neanderthal and half Denisovan.
Weaning electric cars off cobalt.
End note
Dan Wang’s essay on optimism is my one recommended must-read in this week’s bumper issue of Exponential View. He articulates a very similar point to my core thesis: that in order to make sensible decisions we need to understand the technologies and how they can be assembled into meaningful systems, and the importance of “process knowledge” is being able to imagine and realise positive futures. Try to find some time to take it in.
I'm restarting the Exponential View podcast in a few weeks. Most of the season has been recorded, but it isn't too late for you to introduce me to potential guests. Just drop me mail. Incidentally, I love hearing from you. I know, due to volume of email, it sometimes takes a few weeks for me to reply, but the twitter channel (@azeem) is also a good way to connect with me.
🏖 Sending this from a beach in Spain. How is that for commitment to the cause?
Asta la vista,
Azeem
P.S. Share EV with your network: email at least two colleagues, tweet about this issue or post it on Linkedin.