20 August 2023

Welcome back to the Week That Was series highlighting things from the interwebs which are interesting, noteworthy and/or probably worth your time.

Articles📝, Tweet(s)📱, Videos🎥, Charts 📈 all fair game with or without attendant commentary.

“I am only interested in everything” ― Roger Deakin

🎨 Many Skies

Many Skies, Alex Hyner, Photography/Photoshop, 2023

₽🔫🎙️ Ruble Roulette

From the Ones & Tooze podcast:

Are Western sanctions on Russia working? The Russian central bank raised interest rates by a dramatic 3.5 percent in the past week after the ruble dropped sharply—clear signs that the Russian economy is faltering. But in many ways, Moscow has defied expectations. Adam and Cameron explain how.

In light of the recent geopolitical situation in West Africa, they also go into the economics of Niger

And on the subject of supply-chain jiggery-pokery, The Economist explores whether or not America’s China policy is working?

On the outside it looks like decoupling or de-risking is actually working. Business operations are relocating to other southeast Asian countries, India, and Mexico. But in crucial ways, the process is only skin deep. Take a closer look and the exodus from China is actually driving closer integration between the Chinese economy, and those of America’s friends.

❄️🏦🌍 Frostbite

Max Cuvellier who writes the 📝Africa: The Big Deal newsletter posted an informative graphic showing the trend change in equity fundraising on the continent - after the surprising VC heatwave of ‘22, correcting sharply in some in ‘23.

🤖🖥️ Dept. of Computer Cognition

Interesting paper has come out showing that bots now outperform humans in many “are you a robot” CAPTCHA tests. Distinguishing between real and artificial users will continue to be a real computing challenge.

Part of the abstract:

For nearly two decades, CAPTCHAs have been widely used as a means of protection against bots. Throughout the years, as their use grew, techniques to defeat or bypass CAPTCHAs have continued to improve. Meanwhile, CAPTCHAs have also evolved in terms of sophistication and diversity, becoming increasingly difficult to solve for both bots (machines) and humans. Given this long-standing and still-ongoing arms race, it is critical to investigate how long it takes legitimate users to solve modern CAPTCHAs, and how they are perceived by those users.

📚An Empirical Study & Evaluation of Modern CAPTCHAs

"Cars won't replace horses, horses with cars will."
-- Mati Roy pic.twitter.com/w5Xl8OQCrw

— Eliezer Yudkowsky (@ESYudkowsky) August 17, 2023

Decision-theorist Yud is the face of AI safety, having practically invented the field and is the quintessential AI doomer - as far out on the doomer spectrum as it’s possible to be (while being well-reasoned along the way).

On the other end you have builders like George Hotz (the guy who jailbroke iOS and reverse engineered the PS3 as a kid - and having since started self-driving company comma.ai) representing what has since been dubbed the e/acc (effective accelerationism) position.

The e/acc community believe some version of these tenets/goals:

• Increase the amount of energy we can harness as a species (climb the Kardashev gradient.) In the short term this almost certainly means nuclear fission/fusion.
• Increase human flourishing via pro-population growth policies and pro-economic growth policies (emergent altruism, placing economic power and agency in the hands of the meek).
• Create Artificial General Intelligence (allows supplementation of labor with capital) the single greatest force multiplier in human history.
• Develop interplanetary and interstellar transport so that humanity can spread beyond the Earth.

Anyway, blog favourite Dwarkesh Patel hosted these two for a debate recently. Probably worth a listen/watch if these topics interest you.

If you’d prefer visuals, here’s the current medium-res recording of their live debate.

💰📊 Buffett

Buffett’s Portfolio in Q2 2023

(Credit: LeverageShares)

👨‍🔬🧲 LK-99 Update

A fortnight after the 📚blockbuster claims by Sukbae Lee, Ji-Hoon Kim and Young-Wan Kwon regarding a “First Room-Temperature Ambient-Pressure Superconductor” - there’ve been numerous replication attempts in everything from indie labs to professional and academic setups as well.

All indications suggest all replication has failed and the pre-print paper’s claims seem dubious at this stage. Betting markets on Manifold and Polymarket are now pricing potential replication at under 10% chance.

Punters are reacting to the spate of papers which have been published in the last week or two which systematically dismantle many of the original paper’s claims as well outline the issues with attempts to replicate a superconducting version of the claimed modified lead-apatite structure.

One of the 📰more notable attempts came from a team at the Max Planck Institute for Solid State Research in Stuttgart, who synthesized pure crystals of LK-99. Their conclusions are that not only is the compound of copper, lead, phosphorus and oxygen not a superconductor - but it’s barely a conductor at all!

📰LK-99 isn’t a superconductor — how science sleuths solved the mystery

This is the rigorous scientific treatment of the sentiments shared by condensed matter physicist and University of Nottingham professor Phil Moriarty (in classically exasperated fashion!) in this video from a few days ago.

My favourite reaction to all this has been this meme whose provenance, for the life of me, I can’t remember.

This tweet’s trendline is probably also true of the Google Trends search term for “LK-99”

It was good while it lasted pic.twitter.com/iNBJCJUrS1

— Nirmalya Kajuri (@Kaju_Nut) August 12, 2023

A fun two weeks to track but the most important take-home is that science has functioned the way it’s meant to - which is heartening.

Alas, long-time reader Oskar Groblewski encouraged exercising caution a few weeks ago - and I’m sure he watches yet another bold claim live and die within a fortnight news-cycle with some bemusement. Amusement even 😄. I just hope purveyors of big headlines will also be quick to issue follow-ups throwing cold-water on previous headier exaltations.

👶🌍 Fertility

📰The Economist’s Mike Bird:

My briefing this week: None of the world’s 15 largest economies now have fertility rates over 2.1. The ageing and shrinking of the world will pose a threat to the most disruptive innovation and entrepreneurship, disproportionately pioneered by the young.

One wonders who of the nations with positive TFR (Total Fertility Rate) can leverage their demographic dividend in coming decades or if the rich nations with negative TFR will find technology-based ways to reverse the implications of falling populations (via robotics & AI - per our e/acc friends).

🖼️🎮 Clone-a Lisa

Another eclectic game by Matt Round where you get 60 seconds to try and counterfeit a famous painting.

Here’s Matt on how it works:

I made this game; in case anyone’s wondering how the scoring works: It draws your version over the original with ‘difference’ compositing, then downscales the result (using ‘medium’ quality to avoid most pixels getting skipped) and adds up the R, G & B values for every pixel in that small image. This total is then compared with the total difference for the initial state (which represents 0%).

Darker areas have the biggest initial difference, and the whole pic is darker than the initial canvas, so get a lot of paint down and pay most attention to the dark bits, you’ll soon get over 70%. 85% is v good, and I think 90% is probably possible with enough skill and speed.

🖼️

Here’s Steph Ango’s attempt at the Mona Lisa which scored 85%

🖌️Clone-a Lisa

🌎📈 Largest Economies

Founder & CEO of NextML Henric Larsson often posts on Reddit’s dataisbeautiful subreddit and recently (re)shared an infographic showing two decades of changes in the world’s largest country GDPs - both in today’s USD as well as 2015 bucks.

⚗️🧮 9 Lines

All of physics in 9 lines

Per 📝Motion Mountain:

The 9 lines contain all present knowledge about nature, including all textbook physics and all observations ever made.

The simplicity of the 9 lines and their vast domain of validity are intensely fascinating. No known observation and no known measurement contradicts these 9 lines, not even in the last significant digit. Discover a single contradicting observation and you will become famous.

The 9 lines determine all equations of physics, as the references in the pdf show.

Lines 1, 2 and 3, together, uniquely determine special and general relativity, cosmology, the Hilbert Lagrangian and Einstein’s field equations. (This is shown here.) Line 5, with line 1 and 4, uniquely implies thermodynamics. (Shown in https://arxiv.org/abs/2307.09914) Lines 1, 2, 4, and 6 to 9, together, uniquely determine quantum field theory and the Lagrangian of the standard model of particle physics, extended with massive Dirac neutrinos and PMNS mixing. Written out, the Lagrangian is somewhat involved, as Jim Shifflett showed.

The 9 lines show how little math is required to describe nature with high precision.

The 9 lines describe everything that moves - and thus everything that happens in nature. The 9 lines given above improve the humorous 9 lines found here.

The 9 lines resulted from the work of countless scientists and engineers over 400 years.

Galileo started around the year 1600; line 1 resulted after over 150 years. Line 5 arose from 1824 to 1929, line 6 arose around 1860, lines 2 and 4 around 1900, line 3 from 1915 to 2002, and lines 7, 8 and 9 from 1936 to 1973.

The 9 lines imply the lack of trans-Planckian effects, as detailed here.

No such effect has been observed and no such effect will ever be observed.

The 9 lines contain all natural sciences.

They contain physics, chemistry, material science, biology, medicine, geology, astronomy and all engineering disciplines. Again: any single observation contradicting the 9 lines will create a sensation.

🎨 Friday Night

A Perfect Friday Night, Leopoldo D’Angelo, 3D Art, 2023

💬 Deep Cuts

“There’s always someone blocking my clear view of an artwork, and that someone turns out to be myself” - Stephen Fry

🐰🐕 One more thing

3,5 minute thriller. I wonder who you’ll be rooting for…

This was genuinely crazy to watch pic.twitter.com/pmcmHU3sbP

— What the media hides. (@narrative_hole) August 14, 2023

📧 Get this periodically in your mailbox

Thanks for reading. Tune in again and please share with your network.