What Makes Some Technology Good?

My friend Malaika recently posed the question: "What sets apart technological experiences that worsen the human experience (VR/AR, fast fashion, tiktoks, bad spotify algorithm) from tech/mediums that makes it much better, makes us more present and freeing our time for what we want to do (google maps, uber, granola, nts.live)?"

My best guess is that the distinguishing factor is whether it enables a new capability (saving time w/ uber, not getting lost with maps, note-taking + multitasking, music and shows that we'd never hear about), or trades-off against another one that is already pretty good (real life, slow fashion, reading, music recs.)

The free market exploration of the tech tree does not privilege a capabilities-enhancing view by default. It does much better than chance — the market, after all, is the best preference aggregator that we've been able to come up with. But often, particularly at the edges, for the most profitable companies, we see what we want and what we want to want come apart.

One useful little abstraction that I heard about recently in the AI Safety textbook splits normative ethics into three buckets, ones that prioritize:

Pleasure

Preferences

Objective Goods / Capabilities.

Funnily enough, macro-econ has a homomorphic mapping too — gross national happiness, gross domestic product, and the holistic "global development index."

Often times these are aligned — fulfilling your basic needs satisfies all three — but does come apart at the edges, particularly for technology.

As an AI blog, we ultimately do want to think about AIs that can fundamentally enable a new capability (curing cancer, predicting the weather, automating some mundane work) while not replacing an existing capability (some jobs, social relationships).

That's the key challenge: the market will find things that satisfy 2) preferences, by default — and not even long and considered ones, just wherever we put our money. And so we'll have to shape technological development such that it ultimately ends up enhancing our lives, rather than replacing what good existed there with a profitable but otherwise hollow alternative.

I guess this makes me a humanist, at core. I believe that the free markets are a powerful tool to find things that enhance capabilities, but must not let run unfettered, lest it discover an attention algorithm that hacks our brains, or a technology that destroys the world — but in the meantime we'll have great companies.