Let's be Fwends #147:

How the Robots Will Overthrow Humankind

"The nearest."

~ Henry David Thoreau, when asked at table what dish he preferred

Hi, and welcome to edition 147 of Let's be Fwends. Today's issue is AI. Again. Of course. But there were a couple of interesting thoughts out there I wanted to share with you. First, there's Yuval Harari trying to set the record straight how it would look like if the machines really took over. There's also a prototype of how that might look like, and the question of scale came up again. Finally, some good news on plastics, and a really fun video.

Enjoy!

If AI Takes Over, it Will be Through Forms, not Machine Guns

Yes, AI-controlled machine guns are very, very scary. But that's probably not the path for machines to take over the planet.

In an interview with Noema magazine, Yuval Noah Harari argues that the Hollywood-scenario of heavily armed combat robots controlled by an Artificial Intelligence subverting its systems of governance to rebel against humans is not the danger we need to be prepared for.

What the cinematic image so far misses is that the AIs don’t need to start from scratch. They are inserted into our own systems, and they can take our systems over from within. I’ll give a parallel example. Think about lawyers. If you think about the best lawyer in the United States, this person is, in a way, an idiot savant. This person can be extremely knowledgeable and intelligent in a very, very narrow field like corporate tax law, but can’t bake a cookie and can’t produce shoes.  If you take this lawyer and drop him or her in the savannah, they are helpless — weaker than any elephant or any lion. But they are not in the savannah. They are inside the American legal and bureaucratic system. And inside that system, they are more powerful than all the lions in the world put together because that single lawyer knows how to press the levers of the information network and can leverage the immense power of our bureaucracies and our systems. This is the power that AI is gaining. It’s not that you take ChatGPT and throw it in the savannah and it builds an army. But if you throw it into the banking system, into the media system, it has immense power.

This is a kind of AI-warfare many marginalised groups are already experiencing.

AI-powered "analysis" of crime patterns have shown a strong bias against minority neighbourhoods. AI-"enhanced" health care is likely to exacerbate inequality based on gender, disabilities, ethnic background, religion or any other factor that can be used to marginalise people.

Often, when machine learning is inserted into any process, it multiplies human biases, misconceptions, and power structures. It might happen due to neglect, ignorance or carelessness by the people harnessing these processes. Maybe it also happens on purpose. Right now, we're not talking about it, because it's still humans on the levers, and the people affected by it have no voices.

But if you want to see how an AI might wage a war against humankind, just look at the way marginalised people are affected by contemporary AI-systems.

Meet Robo-Jesus

If you need a sign from the Lord Almighty that the war of machines against humans will be fought by machines embedding themselves into our social processes, then look no further than the church of Lucerne, Switzerland.

Turns out a computer can be a vessel of God. But what truly shows the faith of the people running the church was that they tested their chatbot on just 30 people before letting it loose.

Scale can't Fix All Problems

Last time, I briefly touched on the idea that "global scale things" oftentimes make things worse for us, and we maybe need more "human scale" things. Similar to slowing the fuck down, we'll maybe need to scale the fuck down as well.

In a startup-from-2014-thinking-dominated tech landscape (cough "founder mode" cough) where hustling is the only modus operandi conceivable, speed and scale are the two unquestionable pillars of everything. Building faster and growing faster promises to solve all your problems.

But is this culture of "now" running out of steam? It takes more and more resources to keep the ball rolling, and with diminishing results. As if Sisyphos's hill got steeper and steeper.

One field where the "scale, scale, scale" mantra is still dominant is all forms of Artificial Intelligence modelling. In their paper "Hype, Sustainability, and the Price of the Bigger-is-Better Paradigm in AI" Gaël Varoquaux, Alexandra Sasha Luccioni, Meredith Whittaker raise the question how this can still go on, what's next for AI research and if the real way forward lies in smaller models that avoid all the downsides of the "scale is the solution to everything".

Compostable Plastic

"Mechanically strong yet metabolizable supramolecular plastics by desalting upon phase separation" sounds cool, but a bit unfathomable. What it means is that scientists found a way to build plastic that dissolves in saltwater and can be eaten by sea creatures. Imagine if we had this material a couple of decades ago - the Great Pacific Garbage Patch would be the world's largest All-You-Can-Eat buffet.

10,946 Post-its

Every day for a year (well, nearly),  Daren Jannace made 30 drawings on Post-Its. Running them on 30 frames per second gives you a fun 6:58 videofull of good ideas, set to the background of random sound recordings he made on his phone.

That's it for this edition of Let's be Fwends. If robots really want to subvert humankind by penetrating our systems and processes, Austrian bureaucracy will save us. Nothing and no-one can beat it. Go ask Kafka. 👩‍⚖️