A weekly summary of what I’ve found interesting at the intersection of economics, finance and technology.
Facebook and politics—Even now that the US elections are all but settled, there is little let-up for Facebook. Charlie Warzel from the New York Times was given access to the feed of two baby boomers, and witnessed the polarisation and antagonism, all in the name of engagement, first-hand. There is nothing particularly new here, but his account confirms the extent to which the information landscape is broken, and reminds us that this is by design and intentional.
The article begs the question of how to fix this of course, and a puzzling proposal comes from a trio of professors (among which Francis Fukuyama of “End of History” fame), who are members of Stanford University’s “Program on Democracy and the Internet.” Quote:
One of the most promising solutions has received little attention: middleware. Middleware is generally defined as software that rides on top of an existing platform and can modify the presentation of underlying data. Added to current technology platforms’ services, middleware could allow users to choose how information is curated and filtered for them.
Unfortunately, not only is this not a promising solution, the whole text serves as a reminder of how uninformed the whole debate about social media and society is. The authors, accustomed to moving in policy and business management circles, betray their lack of technical expertise with their use of the term “middleware”: while the definition is technically not wrong, no one actually working in tech would refer to a user-facing application as “middleware,” not to mention that everyone stopped using the term after the mid-1990’s.
Worse, I don’t see how “allowing users to choose how information is curated and filtered for them” will solve the type of issue highlighted by the NYT above. People can already choose what information they want to see today, by visiting mainstream media websites, which is just as easy as visiting Facebook. The real issue is that they don’t, because they are relentlessly sucked into the engagement machine Facebook built. Fixing that requires sophisticated regulation with an understanding of the industry and its drivers. Perhaps a more “promising solution” would be a regulatory approval process similar to that of new medicines, with tech companies required to show that their new algorithms are not harmful.
On a related note, this diatribe against blockchain-based voting systems from renowned cybersecurity expert Bruce Schneier is worth reading. Perhaps a bit too dismissive, but interesting nonetheless.
Crypto sees a resurgence of interest—Undoubtedly related to the huge run-up in Bitcoin’s value, people are talking about crypto again. Facebook is going ahead with Libra, but for now with much reduced ambition compared to the lofty plans of the original announcement. It is preparing to launch a USD-backed (one for one) stablecoin early next year, with other currencies to follow later (maybe.)
As ever, whether it makes a difference remains to be seen, but in the meantime established asset managers are dipping a toe in the water, positioning crypto as a useful hedge.
While we’re talking about crypto, related to the aforementioned blockchain voting piece from Bruce Schneier, his take on blockchains overall is also worth reading.
The state of ethics in AI?—A story emerged this week offering an edifying glimpse behind the scenes of ethics in AI research. A Google researcher was fired after a critical email sent to the other researchers, prompting quite a petty response from senior management. The truth as always will be somewhere in the middle, but the story does nothing to shore up confidence in Big Tech’s overall culture. Read
The Internet of Things (IoT) is the next big surveillance platform—Beyond the security issues associated with IoT, tools and platforms are emerging to use the data these devices generate for surveillance, including for military applications. Puts your friendly and somewhat clueless robot vacuum cleaner in a different light… Read
Brace for a tsunami of AI-generated text on the internet—There already is far too much to process on the internet, and while even tools like GPT-3 (covered in previous issues) still lag distinctly behind human intelligence, they are good (and fast) enough to enable an explosion in the amount of low-grade content, which will make it even harder to find something of value. Gloomy times ahead. Read
Website notifications are being abused by scammers—It has become more common lately for a website to ask permission to show you notifications on the desktop. Turns out that some operators then sell that channel to scammers and other ne’er-do-wells, which sounds like something you’d rather avoid. Keep it simple and basically just say no to any website prompting you to allow notifications. Apart from the security threat, it’ll also help with your mental sanity…Read
Better yet, just block websites from asking you—The Verge has a useful guide on how to switch off those prompts in the first place. Read
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.