Surveilled #57 — Exposure notifications, Amazon’s labour relations, and circumventing internet censorship
Surveilled
Issue 57
A weekly summary of what I've found interesting at the intersection of economics, finance and technology.
Apologies for the irregular posting schedule lately, building a new set of routines and an operating rhythm after moving countries in the midst of a pandemic is proving to be challenging. We’re getting there though, so I’m hopeful we’ll be back to our regular schedule soon.
Six Links
Google and Apple Move To Provide End-To-End Solution for Exposure Notifications — Android and iOS will provide exposure notifications directly to users, whereas this was initially meant to be done by the health authorities' apps and infrastructure. Many national and local governments are lagging behind however, hence the change of plans. More generally, evidence seems to be emerging that exposure notifications can help manage the pandemic’s spread, provided that it is accompanied by consistent messaging from the government, and in a comprehensive framework that also emphasises testing. In that sense, the technology’s adoption seems to follow the familiar path: first overpromise, then a disappointed backlash, and finally a mature implementation that becomes ubiquitous. Maybe we’re at the end of the second stage. Read (Wired $)
Amazon Accidentally Reveals Its Stance on Organised Labour — The tech giant posted a couple of jobs for analysts to track “labour organising threats.” Applicants could be expected to help with court cases to secure injunctions against organisers etc., and the postings lumps organised labour together with hate groups and terrorists. This shouldn’t be surprising anymore by now, but it’s a good reminder that this period’s arch-capitalists are not much different to those from a century ago. Read (Wired $)
How Belarussians ‘Hotwired’ Their Internet — Following the standard dictator’s playbook, the authorities in Belarus shut off the internet to prevent protesters from organising. Only this time, people knew what to do, and switched to encrypted chat app Telegram and VPN’s en masse. Incidentally, and inevitably, these random internet shutdowns wreaked havoc on the country’s sizeable tech sector. Read (Gizmodo)
Hong Kong’s Surveillance Environment Is Deteriorating Rapidly — Speaking of disrupted internet, authorities in Hong Kong are stepping up their electronic surveillance, monitoring activists’ social media accounts and trying to get into their phones by none too subtle means. This serves as another good reminder to review your security hygiene, e.g. your phone passcode, how to disable biometric security in an emergency etc., even if you’re not exposed like the protesters in Hong Kong and Belarus. Read (NYT $)
A Practical Adversarial Technique To Hinder Facial Recognition — There is also good news. Researchers at the University of Chicago came up with a practical adversarial technique to fool facial recognition systems that depend on photos posted to social media. The software, called Fawkes, makes pixel-level changes to photos that are invisible to the naked eye but confuse the algorithms. This kind of “guerilla warfare” against AI is an important tool to check its rise, so any positive developments are worthy of applause. Read (University of Chicago)
Maybe Palantir Is Not as Fearsome as We Thought — The data analytics slash surveillance company filed for IPO recently, providing a detailed look at its business through the mandatory S1 filing. It turns out its economics don’t look stellar, and its technology may not be all that advanced either. This tends to match with what I’ve seen at previous employers who were clients. Newsletter The Diff has an in-depth look at the S1 filing. Read (The Diff)
That's it for this week's edition. As always, thanks for reading and please forward this to anyone who you think might be interested, it would be much appreciated.