5 min read

🔮 Massive AI models; social media extremism; carbon dioxide removal; resilient cities, chastity cages & renting people ++ #305

🔮 Massive AI models; social media extremism; carbon dioxide removal; resilient cities, chastity cages & renting people ++ #305
Written by Azeem Azhar

Hi, I’m Azeem Azhar. I convene Exponential View to help us understand how our societies and political economy will change under the force of rapidly accelerating technologies.

How should we think about vaccine passports? I discuss with Exponential View members in our Friday discussion. It’s not too late to join the conversation.

If you missed it, this week’s Chart of the Week looks at Japan’s energy crisis, the flow of venture capital into Africa, Tesla’s worth, and more.

Dept of the near future

Google recently trained a new language model with 1.6 trillion parameters. The more parameters, the more sophisticated a model can appear. OpenAI’s GPT-3, about which I’ve written,  had 175bn parameters and good craft some technically impressive outputs. Now that firms have figured out that increasing the complexity of these models drops jaws, there is obviously some competition going on. Microsoft (which pulled out a 17bn model last year that was briefly the world’s biggest) is also in the running. So long as bragging rights matter, and given powers-of-ten, this means that we shouldn’t be surprised to see a 10 trillion parameter model within a year, assuming the teams can find a sufficiently large training corpus. Of course, none of these obviates the limitations of such an approach. Indeed, the “Stochastic Parrots” paper by Gebru, Bender, et al outlines many of the risks of such large models, in particular, the way they will rigorously encode whatever problematic material is fed into them and they become rather complicated to analyse, debug and eliminate the source of such issues. (Elsewhere, I found this overview of deep learning in longevity research interesting.)

🎙We’re known for a long time that social networks with recommendation algorithms usually incentivise people towards more extreme positions. Charlie Warzel and Stuart demonstrated in a well-researched piece that those radicalisation dynamics were at play leading up to the attack on the US Capitol.  They argued that “Facebook’s algorithms have coaxed many Americans into sharing more extreme views on the platform — rewarding them with likes and shares for posts on subjects like election fraud conspiracies, Covid-19 denialism and anti-vaccination rhetoric.”

👀 Joan Donovan’s critique is an excellent read: “As we, as a society, consider next steps, we should keep in mind that emphasizing scale has a trade-off with safety” Mark Zuckerberg finally pulled the plug on Donald Trump’s account when other execs finally pointed out that  Facebook’s “business as usual” approach isn’t working.

📱With more pressure coming down on the major social media platforms, smaller alternatives are gaining followers and support. MeWe is one such platform that has seen an explosion of new users due to a crackdown on extremist groups on the likes of Facebook and Twitter. It puts the leadership of MeWe in a difficult position. The platform was founded as a privacy-focused alternative to Facebook and now has the tough task of moderating 15 million users.

🕵️ State back use of computational propaganda is rampant. The Oxford Internet Institute identified 62 countries in which state agencies themselves were using social media to shape public opinion. Of these, state actors in 48 countries were found to be working with private computational propaganda companies. Computational propaganda provides incredible bang-for-your-buck, the researchers tracked that despite its growth, total private contract value was below $60m over 13 years. (A single B-2 bomber mission runs to about $3m, excluding the cost of the munitions, supporting tanker or airborne command costs.)

So we’re all jumping off WhatsApp onto Signal, are we? Cracking profile from October 2021 on Moxie Marlinspike, the person behind it.

🔋 Dept of decarbonisation: 414.87 ppm | 3,433 days

Each week, I’m going to remind you of the CO2 levels in the atmosphere and the number of days until we reach the 450ppm threshold.

The latest measurement (as of January 13, 2021): 415.02 ppm; January 2020: 412.37 ppm; 25 years ago: 360 ppm; 250 years ago, est: 250 ppm. Share this reminder with your community by forwarding this email or tweeting this.


🌎 Carbon dioxide removal (CDR) is critical for offsetting emissions. Holly Jean Buck and Roger Deane Aines go deeper than most on the issue in their new book CDR Primer (you can read it online for free). As new technologies to remove carbon come to market, the simple role of CDR can’t be ignored.

🌞 The cost of solar power has dropped dramatically over the last 40 years. Advances in technology have helped but so has the role of governments.

🇨🇳 An interesting analysis of the three, sometimes mutually-conflicting, pillars of Chinese environmentalism: stability, science and trade. The piece goes into the coercive and non-coercive tactics Xi’s government is likely to keep using as the country works toward its decarbonisation goals.

Short morsels to appear smart during the next impeachment

📱TikTok’s exponential rise continues. UK TikTok users doubled the amount of time they spent on the app last year, and in doing so TikTok overtook Facebook in time users spend on the platform.

🐎 Akbar the Great, India’s ultimate Renaissance ruler, knew a thing or two about political pluralism. Maybe it’s time to revisit his largely forgotten socio-political views.

🦠Exponential technologies enabled the rapid development of vaccines for Covid-19. When the urgency of the pandemic subsides, the lessons learned could help find answers for other health priorities.

👀 Sometimes you have to scratch your head and ask why. Hackers were able to remotely take control of internet-connected chastity cages and demand ransom to unlock them.

China’s hold on rare earth metals.

🌆 While the exodus from cities might be temporary, the effects are no less profound. San Francisco might be the most interesting example of these trends.

🔋 Mohammed bin Salman, the young and ambitious crown prince of Saudi Arabia, unveiled The Line, a 170km-long city built in a straight line near the Red Sea. Will this be a milestone or Saudi Arabia’s own Brasilia?

💁‍♂️Rent a person who does nothing. A man in Tokyo rents himself out to other people to “do nothing”. He is in great demand.

Endnote

The podcast season restarts after the holiday break in ten days. If you haven’t already tuned in to the Christmas specials, let me recommend them to you.

Have a great week!

Azeem


What you’re up to – notes from EV readers

Doc Searls published “social media shell games.” Good read.

Calum Chace published “Into the Roaring Twenties” in Forbes.

Sangeet Paul Choudary published a 90-page deep-dive analysis on key themes in the platform economy for 2021.

Tom Raftery started a new podcast called Climate 21. The podcast is named after SAP’s Climate 21 programme – an initiative to help its customers (which include most of the world’s biggest emitters) calculate, report, and reduce their climate emissions throughout the 21st century.

Claudia Chwalisz published “The pandemic has pushed citizen panels online” in the journal Nature.

Christian Toksvig published four OTT predictions for 2021.

Meera Clark discusses why we work, and how answering this question can help us achieve professional transformation in the new normal.

John Thornhill published “Conspiracy theorists destroy a rational society: resist them” in the FT.

To share your news and projects with other readers, email joseph@exponentialview.co

Comments

Sign in or become a Exponential View member to join the conversation.