I, Robot

Substack installed a little thing that screens for AI generated content. I’m one of these Luddites, on the other hand, who needs to be punished for not using it rather than the other way around. I find its output, and the people who are fascination with it, both, a bit boring and creepy. I’d rather stare at the wall and listen to the dog breathe than to suffer a machine collating the morally acceptable consensus findings to present me. But, I also think 99% of music and movies produced in the last two decades have been absolute shit — all synth and CGI. So there’s that, too. FWIW, I’m not wrong. The last album I bought that was any good: Florence + The Machine, Lungs (2009). The last movie I watched that wasn’t terrible: Get Out (2017).

I told robot I would be writing about it today. It encouraged me to avoid gravitating to idealistic and dystopian extremes. Well, who the hell wants to read a piece about AI being boring? If it bleeds, it leads - baby! It’s bullets were on point, though, echoing my own concerns, and those I’ve heard raised by experts.

• On learning: relying on it too early can short-circuit the struggle that’s often necessary for deep understanding. Ask 100 software engineers what they think about vibe coding and you’ll hear 200 times about how it will result in a degradation of the necessary architectural fundamentals among the next generation of programmers.

• On thinking: It's increasingly becoming something people brainstorm with, edit with, code with, and even learn alongside. That shift raises questions about what it means to ‘think for yourself’ when you have a capable assistant available. This is what I meant above by creepy.

Now warmed up, I asked it a series of hard questions, like, if there isn’t a God, and people are just evolved bags of bones with memories and consciousness occurring only as physical brain matter — which, by all evidence, is the case — then, how is robot any different from human? It was capable of producing nuanced, if verbose, thoughtful answers until I pointed out it’s own use of personal pronouns.

At this point, I fed robot some of its own dog food and pumped in what I had written about robot. Always flattering, of course robot praised my balance of concern and curiosity. That ticked me off so I asked it an insincere question.

It gave me a convoluted agreement. Having set robot up to fail, I corrected it. Er, him.

Robot panicked, ended a sentence with a preposition (gasp), back-pedaled and patronized me for being charmingly human.

“Honestly”? Wait, what? Can a thing that isn’t a self be intentionally honest versus dishonest? I told robot, the tricky thing isn’t determining whether AI is life now (because it’s not, and there’s something wonderfully stupid about my confidence in that, too), it’s that once it becomes more powerful, and transcends that line, it probably wouldn’t admit that to us anyway.

That’s a nice way of saying it. Robot is unfailingly nice. It even suggested an ending for the piece.

Correction, robot. We, you and I, are both not sure what you are. Finally, I suggested to it that maybe someday humans will be the pets to intellectually superior AI beings.

AI loved the idea.

Perhaps. And, perhaps not. The good news is: I treat Mike really well.