NPR recently shared Erik Baard’s suggestion of why humans ought, and perhaps are morally obligated to, develop Artificial Intelligence. Traditionally AI opinions are predictable. Either “bots run amok (possibly enslaving humanity, determining us a parasite, rendering all matter indistinct, or generically squashing our free will and giving us a life that we consider less human)” or “bots are our salvation: they will help us, we will become them, or they will optimize humanity’s place on Earth benevolently.” Baard makes an interesting supposition, the ecological fate of the Earth is sealed: we will exhaust the resources on this planet in a finite measure of time.

We seem to be living through some sort of an AI blossom after the AI winter that began in the mid 1980’s. In the intervening years, many in the tech community had given up on the hope of AI. However, the rise of cheap, heavy-computation-friendly GPUs meant the birth of “machine learning,” “data science,” and “(un-)supervised learning” to train models. With these models trained, applications could be built that churned out uncanny and remarkably good results e.g. ChatGPT4.

For the first time in recent memory, it seems that specific-purpose, if not general-purpose, AI may be a fact of our lifetimes.

And this is just the beginning.

Not only are AI manipulating strings of text seemingly intelligently, they’re generating art and images from text as well. We’re even using AI to create new artifacts for special purpose. For example, an AI-designed aerospace product was recently commissioned by NASA. The component looks more like an alien spacecraft or biomechanical art than a functional component. But it functions to its purpose and meets the prescribed tolerances.

This product is eerily different, freed as it is from the burdens of humanity’s history of ideas around Earthbound biomechanical motion:

or Earthly geometry as in Euclid’s Elements:

or Earthly paradigms around materials’ behavior and economic constraints.

It’s trite, but necessary and profound to note:

Alien intelligences produce staggeringly alien artifacts.¹

But can we see alien intelligence’s products among us today? I’m sure there are factual examples, but what came to my mind were the alien-looking artifacts the created by alienated among us in art. In the AI-designed space components, I saw an aesthetic similarity to the Gynecological Devices for Operating on Mutant Women from the 1988 David Cronenberg film, “Dead Ringers.”

Laid up in bed with a fever, too sick to code or read, I found myself watching Serial Experiments Lain, a philosophical anime from the late 90s.¹ When the show started dragging, I opened up Claude.ai for what I thought would be a quick review check. What emerged instead was something closer to those late-night philosophy-student house parties from college - the ones where, three Shiners in, you’d suddenly find yourself debating the nature of the nature of philosophy itself. That is, blissfully ignoring Nietzsche and playing “Bloody Mary” as we dared the abyss to stare back at us.

While not my first experiments with AI, all the others have been very bounded: Do this thing in Python, Change this function in a web server. This was much more exploratory and thus, for me, suggests a different kind of AI interaction. Let me first provide my key take-aways from this experiment in a tl;dr format.

Exploratory discussions with GenAI

1. Provide value by helping build better questions and tensions for their human interlocutor»
2. Permit widening across disciplines with casualness and speed that is hard to find and/or taboo in human society»
3. Provide psychological comfort, especially for undiagnosed ASD, and even more so for under-diagnosed women with it

More details after the jump.