🔥 🔮 Apple & the megacorps; blockchain blocked; the humanities crisis; the pros of algorithms & ethics of data; esports, coal & robotic hands++ #177
|Aug 5, 2018|
Dept of near future
🍎 The era of mega companies. As Apple passes $1 trillion in market cap, “Apple and Google combined now provide the software for 99% of all smartphones. Facebook and Google take 59 cents of every dollar spent on online advertising in the United States... the so-called labo[u]r share of the economy has been declining in the United States and other rich countries since the 1990s, coinciding with the trend toward corporate concentration.” Bloomberg puts together a lovely photo essay on Apple at a trillion dollars. (See also: Farhad Manjoo on how tech platforms are being forced to overhaul their absolutist positions of non-interference and becoming arbiters of what is and isn’t malignant on their networks. It’ll be messy, he warns. Also, Facebook and Netflix are reaching their own limits to growth: the number of humans.)
⏳ Blockchains are stalling in corporate deployments. “The disconnect between the hype and the reality is significant -- I’ve never seen anything like it”, says one analyst. Another analyst reckons than 90% of blockchain experiments won’t become part of a corporates operations. (My comment: these technologies are powerful but immature. Real work is being done, but it is happening much lower down in the stack, at the base protocol level.)
👔 How important are non-monetary incentives in work? (Academic paper). The neoclassical model--which considers a worker’s tradeoff as between income from work and leisure time--has underpinned much economic policymaking and firm strategy. It doesn’t resonate with, I believe, the experience of many of my peers when considering their jobs. The authors agree: “A firm’s mission and the design of one’s job can create meaning and purpose for employees. As a result, firms will have reason to care about meaning of work. [E]conomists can usefully contribute to the debate about the implications of meaningful work.” I’m not going to lie, this is a dry economics paper; however, it frames the important issue of why people work in terms to which economic policymakers can relate. And as we think about labour force adjustments in the face of automation and gigging platforms, expanding the tools by which we analyse the role of work is important.
🎓 There is a crisis in the humanities, as the percentage of all US college degrees in those disciplines trends down below 5%. EV has long argued that repetitive focus on STEM subjects may mean we lose sight of what it is to rebuild the human systems that comprise society using modern technologies. Facebook’s inability to grapple with its role as a publisher or its civic impact is surely one example of these. Engineering professor, Vivek Wadhwa agrees: “we need our musicians and artists working hand in hand with our engineers. It isn’t either one or the other; we need both the humanities and engineering.”
🙈 The ethics of computer science. One leading researcher argues “the computer-science community should change its peer-review process to ensure that researchers disclose any possible negative societal consequences of their work in papers.”
Dept of artificial intelligence
Are we reaching data’s day of reckoning? Like previous technologies, data scientists need to grapple with the negative impacts of their art, “that shoddy work has an impact on people’s live.” Further, the authors argue “We need to incorporate ethics into all aspects of technical education and corporate culture; we need to give people the freedom to stop production if necessary, and to escalate concerns if they’re not addressed; we need to incorporate diversity and ethics into hiring decisions.” EXCELLENT READ
An alternative view: Want less biased decisions? Use algorithms, argues Alex Miller in HBR. A refreshing take on algorithmic bias “[because] the humans they are replacing are significantly more biased.” And “even if technology can’t fully solve the social ills of institutional bias and prejudicial discrimination, the evidence reviewed here suggests that, in practice, it can play a small but measurable part in improving the status quo.”
👍 A good example of better decisions through algorithms is this lovely visualisation of using a genetic algorithm to evolve floor layouts of a school based on optimising traffic flow and resource usage. Very helpful explanation of a simple example of generative design and how the selection of the fitness function impacts the outputs.
Karina Vold and Huw Price ask if we are to live with artificial intelligence, how will we get it right? Thought provoking read.
Tesla intends to build its own chips to power its autonomous vehicles. In a sense unsurprising as the specialist hardware can result in much better performance. Several firms like Google, Apple, Alibaba, Sensetime (a privately held image recognition firm in China) are developing their processors, and are being joined by a range of new entrants (like Graphcore, Cerebras, Mythic), all of whom recognise the specialist and heterogeneous needs of intelligent applications. As a reminder, in the past three years, the amount of compute used on ‘blockbuster AI’ (such as AlphaGo Zero) has increased 300,000 fold. I regularly talk to people on the demand- and supply-side of compute who believe the amount of compute we will want/deliver in the next few years will be a million-fold, even a billion-fold what we currently do today. Machine intelligence is cycle hungry. (As an example, in the past few months I’ve spoken to different entrepreneurs who are seeking to use classical, non-quantum, approaches to give a power-consumption/compute performance improvement of 100x to 1000x of traditional GPUs which power most ML applications today.)
Uber shuts down its self-driving trucks initiative.
The 640,000 people crossing between Shenzhen and Hong Kong will now be automatically checked by facial recognition systems.
IBM’s Watson regularly recommended erroneous cancer-treatments.
Dept of China & policy
💯 Bradford DeLong: Why the 20th century was not a Chinese century. A stunning piece of economic history exploring how China lost its technological and economic lead in the centuries running up to the 20th. (Long read - est. 20 minutes.)
Beijing and Shanghai are likely to be the next tech hubs of the future. According to CB Insights, Beijing’s tech investment levels have hit $72bn since 2012, about half that of the Bay Area.
The Party and China’s AI giants: “Even though leading Chinese technology companies are starting to become international in their presence, workforce, and activities, the Chinese Communist Party (CCP) has reasserted its authority over them”. The majority of large internet firms have party secretaries who represent the CCP interests.
🔇 Google is planning to launch a censored version of its search engine in China that will blacklist websites and search terms about human rights, democracy, religion, and peaceful protest. Project Dragonfly has been underway since spring of last year
The Pentagon has created a “Do not Buy list” of software from Russian and Chinese origin.
Likewise, the UK government has published proposals for a new “national security screening regime” for foreign investments in British firms which might have “military or dual-use” implications. The UK looks to strengthen its current rules and reckons as many as 50 transactions a year will result in regulatory remedies. We’re going to see much more of this around the world. From an investor perspective, it might point to diminishing markets or a need for firms to establish novel forms of multiple domiciles. But it also suggests that crypto-based systems (software which enables trust to emerge out of an agreed protocol) will have a role to play. Those systems can, in theory, be more transparent, readily auditable and more resistant to attack than traditional software.
Short morsels to appear smart at dinner parties
Peloton, a connected training bike system, is valued at $4.15bn in its latest funding round. Quite a phenomenon.
🎮 The top decile of eSport’s fans spends $100 per month on Twitch. (Great data.)
Facebook launches tools to manage the amount of time you spend on its products.
☎️ End of the line: the number of voice calls dropped for the first time ever in the UK last year.
End of the seam: We’re getting very close to peak coal.
✋✍️ ✌️ Using a mind-controlled robotic third arm to improve multi-tasking. (It turns out that even when you lose the third arm, your rapid task-switching skills improved.)
Interactive story on progress being made with robotic grippers, including a bed-making robot.
😨 Climate anxiety is real, and we don't have tools to deal with it.
I've long argued for the importance of the understanding of the humanities within tech firms, large and small. Technology is building the systems through which we will live our lives. Our very access to the resources we need will be mediated by a new class of technical systems. Wouldn't it be helpful if the people building those systems had, at least, dipped their toe in the vast and deep set of disciplines that have thrown a light on the human condition and the communities within which we reside?
I'd love to put together a reading list, a primer for technologists (or most of us) of key texts from the humanities and social sciences. What are the best reads on ethics? fairness? the political economy? the nature of power and structural advantage? the evolution of our understanding of wants and needs? the importance of identity & tribal affiliations... and more. Drop me a mail or a tweet with your suggestions, and I'll see if we can build something to share.
I’m on holiday for the next couple of weeks, and three of my friends have agreed to take care of Exponential View. They are Linda Liukas, a digital educator and award-winning author, and Nicolas Colin and Laetitia Vitaud, who are thinking very hard about the nature of work & new social contract. Please stay tuned for those superb editions of EV.
Have a great summer,