Interesting articles, April 2024

After decades of building tensions and proxy warfare, Israel and Iran exchanged direct fire with each other for the first time.

An Israeli airstrike killed the leader of Hamas’ three sons.
https://www.bbc.com/news/68783840

A Ukrainian kamikaze drone struck a building 1,200 kilometers inside of Russia. This is the deepest such strike conducted so far in the war.
https://youtu.be/rvhLrIWNCWg?si=ko4IecuKLR_tPtu-

A glimpse into the future of warfare: A Ukrainian quadcopter drone destroys a Russian ground vehicle drone.
https://www.yahoo.com/news/video-captures-rare-drone-drone-170705950.html

Russia’s use of “turtle tanks” shows how dire the threat posed by kamikaze drones has become.
https://youtu.be/s7p2-tMS4UE?si=O_Z1tP3tKkgXY86I

Ukraine just captured a Russian T-72 tank with the most ghetto drone jamming system tied to its roof.
https://www.twz.com/news-features/ukraine-situation-report-russian-anti-drone-electronic-warfare-tank-captured

Russia copied a damaged M1 Abrams tank the U.S. gave to Ukraine.
https://www.twz.com/land/first-confirmed-abrams-tank-variant-captured-by-russia-seen-with-inner-armor-exposed

Russia also captured one of the high tech German tanks donated to Ukraine.
https://youtu.be/LjXuTWfthRI?si=otGZUFRRqo6solmL

Compared to its losses of tanks and land weapons, Russia has lost few aircraft.
https://www.airandspaceforces.com/russian-air-force-has-only-lost-10-percent-of-fleet-in-ukraine-us-officials-say/

Soviet tanks were better than American tanks in the 1970s and 80s.
https://youtu.be/GOJLKOKCxWc?si=OUNQYG8ua0qx64tj

However, the Soviet 1970s-era T-62 is now obsolete on the modern battlefield. Russia is using them in Ukraine anyway due to shortages of better tanks like the T-72.
https://youtu.be/cJfvIOAs-2o?si=MKQ09SDfnwtCqVGi

The Germans had the best telescopic gun sights of WWII.
https://youtu.be/VFzNtqJ7tpY?si=Q_o_oD0ioCr656RE

Here’s an idea for a cheap terror weapon: Release helium balloons that have some steering ability and that carry small, guided bombs. Once the balloon gets above a target, it drops the bomb. It would cost more money to shoot one down than it would cost to build one.
https://www.twz.com/news-features/ukraines-explosive-laden-balloon-operations-against-russia

A massive cyberattack that would have had global consequences was narrowly averted.
https://www.schneier.com/blog/archives/2024/04/backdoor-in-xz-utils-that-almost-happened.html

DARPA held the first simulated dogfight between an F-16 piloted by a machine and one flown by a human.
https://www.darpa.mil/news-events/2024-04-17

Jerome Powell was right about the U.S. avoiding a recession in 2023. Does that mean he is really smart, or that he just got lucky this one time?
https://www.politico.com/news/2023/07/26/fed-hikes-rates-again-and-leaves-options-open-for-more-00108303

Good Lord, these predictions from 2022 were totally wrong:

‘House prices in the United States — which rose during the pandemic by the most since the 1970s — are falling too. Economists at Goldman Sachs expect a decline of around 5%-10% from the peak reached in June through to March 2024.

In a “pessimistic” scenario, US prices could plunge as much as 20%, Dallas Fed economist Enrique Martinez-Garcia wrote in a blog post recently.’
https://www.cnn.com/2022/11/23/business/global-house-price-slump/index.html

Reality: Aside from the briefest of moments in mid-2023 when the average U.S. house price dropped by a tiny amount, prices have only increased since the CNN article was published.
https://www.spglobal.com/spdji/en/documents/indexnews/announcements/20240326-1471289/1471289_cshomeprice-release-0326.pdf

James Cameron released remastered 4K versions of Aliens, True Lies, and The Abyss. He used new computer technology to radically sharpen the images by removing the grains of the 35mm filmstock and tuning the colors. I predicted this would happen, but not until the 2030s:

‘Computers will also be able to automatically enhance and upscale old films by accurately colorizing them, removing defects like scratches, and sharpening or focusing footage (one technique will involve interpolating high-res still photos of long-dead actors onto the faces of those same actors in low-res moving footage). Computer enhancement will be so good that we’ll be able to watch films from the early 20th century with near-perfect image and audio clarity.’
https://www.joblo.com/james-cameron-4k-restoration-defense/

An algorithm transformed a bland software licensing agreement into a convincing, sad song. Do YOU have the musical ability to do this?
https://youtu.be/pGbodliLFVE?si=Uug-40eT5uO7ENN8

GPT-4 is as good at diagnosing eye problems as average human doctors.
https://medicalxpress.com/news/2024-04-ai-doctors-accurately-eye-problems.html

We need an expert consensus on what tests a machine must pass to be deemed a “general intelligence.” Right now, there is no agreement, so a computer could be declared to be an “AGI” if it passed one set of tests favored by one group of experts while failing other sets of tests favored by others.

Without a consensus on this issue, we’re headed for a multi-year, contentious period where different tech companies and labs will unilaterally declare they’ve built “the first AGI” while everyone else disagrees with them. This is part of why I say we’ve entered “The Era of Fake AI.” I think we’ll need the benefit of many years of hindsight to be able to say for sure when the first AGI was created.
https://apnews.com/article/agi-artificial-general-intelligence-existential-risk-meta-openai-deepmind-science-ff5662a056d3cf3c5889a73e929e5a34

Famed philosopher Daniel Dennet died. He recently said this about the future of AI:
‘AIs are likely to “evolve to get themselves reproduced. And the ones that reproduce the best will be the ones that are the cleverest manipulators of us human interlocutors. The boring ones we will cast aside, and the ones that hold our attention we will spread. All this will happen without any intention at all. It will be natural selection of software.”‘
https://www.bbc.com/future/article/20240422-philosopher-daniel-dennett-artificial-intelligence-consciousness-counterfeit-people

Boston Dynamics has retired its old model “Atlas” humanoid robot.
https://youtu.be/-9EM5_VFlt8?si=9oFPyFZZUK9hP_JC

Simultaneously, the company announced it had a made a new type of Atlas robot that was nimbler and more advanced.
https://youtu.be/29ECwExc-_M?si=6odXeLFnpXH7c-ng

The British retail company Ocado, already famous for the heavy automation of its warehouses, has roboticized even more work functions.
https://youtu.be/COBDSmx9QDw?si=oa1JaWuy9OkyrnCB

Humans are so optimized for a narrow set of living conditions. As with space, intelligent machines will beat us to colonizing underwater regions.

‘Key problems include low temperatures, high pressure and corrosion. The change in gases – such as an increase in helium – also breaks electrical equipment and makes people feel cold; the Sentinel habitat will need to be heated to 32 degrees to make it feel like 21. High humidity also creates the potential for a lot of bacteria build-up, with people at risk of getting skin and ear infections, and the pressure also means people’s taste buds stop working – so those of the Sentinel will be eating food loaded with spices.’
https://www.yahoo.com/tech/inside-300-long-project-live-190000639.html

Mark Zuckerberg is richer than Elon Musk for the first time in four years.
https://www.bloomberg.com/news/articles/2024-04-05/zuckerberg-is-world-s-third-richest-person-passing-his-arch-rival-elon-musk

Here’s a video that shows how a “continuously variable transmission” (CVT) works. They’ve become very common in cars, for better or worse.
https://youtu.be/xHWqlfDZnmQ?si=_vhLVEgsI4FZKOvR

A Chinese company has invented a diesel engine with a record-breaking 53.09% thermal efficiency. A typical, commercial diesel engine is only 46% efficient.
https://interestingengineering.com/innovation/china-diesel-engine-thermal-efficiency

‘The Space Shuttle launching from Cape Canaveral in Florida (28.5° north of the equator) is a 0.3% energy savings compared to the North Pole. If we move it to around the equator, such as the European Space Agency’s spaceport in French Guiana, we’d get about 0.4% savings. Maybe that doesn’t seem like a big deal, but every bit helps.

…Launching from the top of Mount Everest would give you a 0.2% savings in energy per kg.’
http://ifsa.my/articles/space-launch-equator-vs-mountains

During the last Ice Age, the planet wasn’t just colder, it was drier. Because so much water was locked up in the enlarged ice caps and glaciers, the atmosphere was drier and it rained less in the parts of the world closer to the equator. The deserts were larger than they are now and the rain forests were smaller. The equatorial regions were more clement to human life, but that wasn’t saying much.
https://en.wikipedia.org/wiki/Last_Glacial_Maximum

Whenever there is a climate disaster, the media always announces it with trumpets. However, whenever a disaster gets fixed or when there are extraordinarily clement climactic conditions, it gets buried on the back page.
https://www.redding.com/story/news/local/2024/03/28/lake-shasta-could-fill-by-end-of-april-finally-after-drought-years/73118853007/

‘Watch your garden glow with new genetically modified bioluminescent petunias’
https://www.npr.org/2024/04/08/1242346659/genetically-modified-bioluminescent-petunias-make-their-own-light

1% of people have “extreme aphantasia,” meaning they can’t visualize ANYTHING in their minds.
6% of people have lesser degrees of aphantasia.
3% of people have “hyperphantasia,” meaning they can see mental images that are so vivid they can’t tell them apart from real images they’re seeing in front of them.
https://www.bbc.com/news/health-68675976

Anti aging therapies are entering mainstream medicine.
https://www.science.org/content/blog-post/rejuvenating-blood-cell-population

French Canadians have high rates of genetic defects due to inbreeding in the distant past.
https://blog.23andme.com/articles/french-canadian-health

The first “conversation” between humans and a whale happened, though we don’t know what exactly was said.
https://www.seti.org/press-release/whale-seti-groundbreaking-encounter-humpback-whales-reveals-potential-non-human-intelligence

‘Amphibians use scream inaudible to humans for self-defense against predators’
How much experience do humans miss because we can’t hear or see outside of narrow bands of sound and light?
https://phys.org/news/2024-04-amphibians-inaudible-humans-defense-predators.html

‘“The textbooks say nitrogen fixation only occurs in bacteria and archaea,” says ocean ecologist Jonathan Zehr at the University of California, Santa Cruz, a co-author of the study. This species of algae is the “first nitrogen-fixing eukaryote”, he adds, referring to the group of organisms that includes plants and animals.’
https://www.nature.com/articles/d41586-024-01046-z

‘Experts suggest the discovery of rare variants in the BSN and APBA1 genes are some of the first obesity-related genes identified for which the increased risk is not observed until adulthood.’
https://www.walesonline.co.uk/news/health/scientists-find-genes-can-raise-28934600

Brain scans that map the structure and activity of a brain can predict whether it belongs to a biological male or female with 99.7% accuracy.
https://arxiv.org/abs/2309.07096

Men are likelier to be gay if they have an older brother.
https://www.npr.org/sections/health-shots/2024/04/15/1243861703/sexuality-birth-order-gay-siblings

He Jiankui, the Chinese scientist who created the first genetically engineered humans, has finished his prison sentence and is again working on genetics research.
https://www.dailymail.co.uk/sciencetech/article-13259289/Disgraced-scientist-genetically-editing-babies.html

What my broken down car taught me about the future

When I was in college, my mother bought me a new, cheap car for my 21st birthday. It lasted me for 19 years and 209,000 miles–my companion through two or three chapters of my life–before finally dying of a seized engine last month. Finding a replacement in a hurry plunged me headlong into the world of cars, and a side effect of all the research and car inspections I did before buying a new one was an understanding of how future technology will revolutionize cars and the industries related to them.

Better designs

My old car was a Chevrolet Cobalt. Over the years, I’d learned a lot about it from working on it in my driveway, so it was sensible for me to consider buying a new one, but the model was discontinued in 2010. That led me to consider its successor, the Cruze, which I assumed would share many design elements with the Cobalt. 

The engine bay of a Chevrolet Cruze

Unfortunately, I discovered the Cruze has an average-at-best reputation among compact cars thanks to problems with its engine and some of the components directly attached to it. The use of lower-quality components was the main culprit, and there was also a case to be made that some aspects of the engine design itself were not as well thought-out as they should have been. 

I bet GM’s engineers didn’t know about these problems, or at least didn’t know they would turn out to be so pronounced, until after a million Cruzes had been sold and at least two years had passed so the problems could be exposed through real-world driving conditions. I also doubt the problems would have arisen at all had those engineers had access to the kinds of advanced computer simulations we’ll have in the future. 

Using hyper accurate, 1:1 simulations of materials and physical laws, car designers could test out unfathomably large numbers of potential car designs and experiment with different components and combinations of components until optima were found given parameters like maximum cost and minimum performance. Each simulated car could be “driven” for a million miles under conditions identical to those in the real world, thus revealing any design or material deficiencies before any vehicle was actually built. (These kinds of simulations already exist, but are so expensive to create that they’re only used to model things like nuclear weapons and stealth bombers.)

Thanks to this, cars in the future will be better and more reliable than they are today, and there won’t be such things as specific car models like the Cruze that have bad reputations for unforeseen problems. All vehicles will be optimized and all car companies will use the same tools for designing their products (which I also imagine would lead to many convergences). 

More diligent maintenance

With the Chevy Cruze out of the equation, I considered another compact car, the Nissan Versa. My research quickly led me to discover that Nissan cars have become infamous among owners and mechanics for transmission failures. This is because most Nissans have “continuously variable transmissions” (CVTs) instead of traditional 6-speed automatic transmissions or 5-speed manual transmissions. 

CVTs are cheaper to manufacture than the traditional transmissions and improve the fuel efficiency of the cars they are integrated into. However, CVTs require more maintenance because they get hotter during operation and produce more metal particle debris due to more metal-on-metal contact between moving parts. Replacing the transmission fluid and filter largely solves the problem and should be done every 30,000 miles in a Nissan car with a CVT. 

Old transmission fluid draining out of a car

To put this into perspective, a 2013 Toyota Corolla with a 5-speed automatic transmission only needs the same transmission service every 100,000 miles. Most car owners still expect that kind of maintenance interval in all new vehicles, and this mismatch between expectation and reality explains most of the Nissan Versa’s bad reputation. It doesn’t help that Nissan itself has downplayed the higher maintenance requirements of its CVT vehicles, or that the kinds of cash-strapped people who buy Versas tend to know little about cars or how to take care of them. 

More broadly speaking, improper maintenance is something that car mechanics constantly complain about (even if it generates a huge amount of business for them). Most cars die prematurely due to owners ignoring obvious problems and not properly maintaining them. Some “bad” cars like the Versa aren’t actually bad, they just need more maintenance than others to stay functional. However, learning about this through research and then staying mindful of your particular vehicle’s maintenance requirements is too much for most human car owners thanks to a lack of time, energy, and sometimes intelligence. 

Intelligent machines won’t have those same limitations. Future cars will have better self-diagnostic capabilities, and will be maintained by robots that will never skip preventative care. And since machines will work for free unlike today’s human mechanics, the costs of this will be much lower. Even poor people will have enough money to change the transmission fluid in their Nissan Versas. 

Gentler driving

Facebook Marketplace was my primary source for my used car search. In a huge fraction of the ads, the owners wrote their cars had “Salvaged titles” or “Rebuilt titles.” That means the car sustained so much damage that its insurer declared it “totaled,” meaning the cost of fixing it exceeded the resale value of the car in its state. Instead of being scrapped, many cars like this are bought at very low prices by mechanics who fix them themselves and resell them for a profit. Those profits tend to be small because having a Salvaged or Rebuilt title is a scarlet letter in the open market because buyers know such a vehicle was badly damaged at some point, and can’t be sure of the full extent of the problem or of how fully it was remedied. I ignored all the cars without clean titles. 

Why do cars end up with Salvaged or Rebuilt titles? Mostly because they were in serious accidents, floods, or caught on fire. Autonomous vehicles will, once fully developed, drive much more safely than humans and get into far fewer accidents. Eventually, they probably won’t even have steering wheels or pedals, making car thefts and ruinous joyrides impossible. 

As I discussed in my blog Hurricane Harvey and Asimov’s Laws of Robotics, autonomous cars could also avoid floods by keeping watch of their surroundings and driving to higher ground if they were at risk of being submerged. Better monitoring systems would also reduce instances of car fires since the cars would be able to shut down their systems if they sensed they were overheating, or to immediately call the local fire department if they caught on fire. 

More careful driving and avoidance of other hazards will sharply lower the odds of a car having to worry about getting a Salvaged or Rebuilt title. Gentler driving that stayed mindful of the car’s engineering limits and avoided exceeding them would also lengthen vehicle lifespans since components would take longer to wear out. 

Conclusion

In the future, vehicles will drive safer and will last much longer than they do today. They will be designed better and will incorporate more advanced materials like future alloys. Moreover, once battery technology reaches a certain threshold, the vehicle fleet will transform to almost 100% electric in a few decades, and electric vehicles are inherently more robust than gas and diesel vehicles we’re used to because they have fewer parts and systems. 

On a longer timeframe, autonomous driving technology will achieve the same performance as good human drivers, and the average vehicle will become self-driving. Machines will drive much more safely and gently than humans, making it much rarer for cars to be damaged in accidents or by driving behavior that overstresses their components. 

Future technology will also benefit car maintenance. The vehicles themselves will have better inbuilt self-diagnostic capabilities, so they’ll be able to recognize when something is wrong with them and to alert their owners. The proliferation of robot workers of all kinds will also lower the costs of maintaining cars, meaning it will not be so common for owners to skip maintenance due to lack of money. The robot butler who hangs around at your house could work on your car in your driveway for free, or your car could drive itself to a repair shop where machines would service it for low cost. 

Under all these conditions, the average car’s lifespan will be over 500,000 miles in the future (today, it’s about 200,000 miles), being stranded because your car broke down will be much rarer, and personal vehicle transportation will be within the means of poorer people than today. Ultimately, cars might only get totaled due to unavoidable freak accidents, like trees suddenly snapping in the wind and smashing down on one of them, or to deliberate vandalism by humans. Likewise, after humans discover the technologies for medical immortality, we’ll only die from accidents, murder and suicide.  

These technology trends will also upend the used car industry. With machines carefully doing and logging all the daily driving and maintenance, secondhand buyers won’t have to worry that the vehicles they’re looking at have secret problems. With highly accurate data on each car’s condition, haggling would disappear and pricing would reflect the honest value of a used vehicle. 

People in the used car industry who make a living off of information asymmetries (the worst example is a car auctioneer who only lets potential buyers examine a car for a few minutes before deciding whether to buy it) would lose their jobs. In fact, AI and autonomous vehicles would let car manufacturers, fleet owners like rental car companies, and private owners sell their vehicles directly to end users without having to go through any middlemen at all. AIs that work for free would replace human dealers and would talk directly with customers who wanted to buy cars. A personal inspection and test drive could be easily arranged by sending the autonomous car they were interested in to the buyer’s home, no visit to the car lot needed.