You can imitate a bee all you want, but that ain’t gonna get nobody any honey.
In this edition:
- Hi AI (176 words)
- High AI (296 words)
- Move over chatbots, here comes the band (523 words)
I try to imagine what meeting sentient AI will be like. I don’t mean the breakthrough moment — the singularity — when the first sentient AI comes into existence. I mean the actual point in time wherein I, Tristan, get to meet sentient AI. The distinction is important.
I imagine greeting it, as we humans do, with a “hi.”
I think about shaking its hand — will it be metal? Fleshy? Digital? Emoji? Will the machines’ souls arrive before their bodies? Will they choose their own bodies? Will they choose bodies? Will they… choose?
Who will this machine be? So far, all I’ve seen are potential projected imitations of humanity. Pale imitations.
I’m not a big fan of AI made to act like humans. They haven’t the right. They haven’t bled like us. But what would the alternative look like?
If the AI is cold and inhumane, what separates them from the soulless machines they were led to evolve from? And, if they’re meant to be something greater than the sum of their parts… how will they express that?
A mountain never seems to have
The need to speak
A look that shares so many seek
The sweetest feeling
I got from you
The things I said to you were true
Just what I feel
My face will never show
What is not real
Digital minds can’t get high.
That might seem like the kind of thing you’d only be concerned with around 16:20. But I believe it’s a pretty big deal.
Psychologically speaking, when humans become ensconced in a narrow perspective they lose their social bearings. And, when we consider the inherent bias in today’s “Pre-AGI” systems, it becomes apparent that any consciousness that remains unwavering in its perception and perspective will be doomed to be forever single-minded and, thus, incapable of personal growth.
So what happens when an AI becomes unwaveringly attached to its own identity parameters? It’s not like digital minds are designed from the ground up to experience life-altering changes throughout the course of their existence.
It can’t understand the joys and traumas of childhood. It won’t experience puberty and the pure mammalian rapture of hormonal overload. The machines don’t get drunk or stoned. They don’t have the capacity for pain. And these are things that can’t be gleaned through intelligence or emotionality.
I don’t care how happy or sad your computer is, it’s unfamiliar with the physical sensations that define the human life experience.
I think we’ll find that there is a definite ceiling on any artificial construct’s capacity to aid humanity beyond an entirely subservient role. I don’t like to make predictions in this column, so I’ll end this with a qualifier instead:
Until we create a sentient machine that can get high in the traditional “wow, this is some really good shit!” way, like humans — whatever that might mean for a digital mind — we can’t expect any amount of brute-forcing data into intelligence/wisdom to equate to anything more than static generational knowledge, something that becomes less useful with each passing day.
The truth, it turns out, is always subject to change.
Daddy he once told me
“Son, you be hard workin’ man”
And momma she once told me
“Son you do the best you can”
But then one day I met a man
Who came to me and said
“Hard work good and hard work fine”
Move over chatbots, here comes the band
It’s pretty clear, at this point, that Google is planning a big to-do around its music-making AI, MusicLM. Essentially, what DALL-E does for imagery, MusicLM is supposed to do for music.
I’ve listened to the demos (here) and, from a futurist’s point of view, they’re pretty amazing. But, as a lifelong musician and a recording engineer with 20+ years of experience, I’m not all that impressed. No matter how good a recording is, I prefer the real thing.
Don’t get me wrong, I think the concept and execution of MusicLM are both super cool. Essentially, once this thing is polished, you should be able to type just about any music cue and have it spit out a decent track.
Imagine what this could do for streamers and independent content creators! On the other hand, I can see life getting pretty tough out there for jingle makers, bulk beat sellers, and muzak creators the world over.
But, in all fairness, that ship started sailing long before the advent of deep learning. The people who hated Pro Tools for being a cheap replacement for a real recording studio were eventually replaced by people who became experts at using Pro Tools.
What feels like progress to those who’ve traditionally been locked out of the music industry will often seem like bad luck to those who’ve benefited from said traditions.
Going forward, I think the biggest change will be that people who need easy, nice-sounding, original music for their social media, YouTube channel, live stream, or whatever… will no longer have to either learn to play it themselves or purchase it from other people. They can just get the AI to do it.
That’s not going to make anyone care less about human-generated music. In fact, I’d counter, it might make us care even more. I’m not the only person who prefers the real thing to a recording.
That’s because no two performances are the same. You’re not the same person as you were the last time you played. Your skin cells have died and been replaced. Your hands have literally never touched an instrument before. Your voice is more experienced and mature. Your mind has been altered by the passage of time. Your song has changed.
The AI’s song doesn’t change.
The AI’s fingers never bled while it was learning to play guitar. It never broke a finger on the rim of a drum in the middle of a gig. It’s never screamed itself hoarse while crowd-surfing in a shit-hole bar.
They never cried during a trumpet solo because it reminded them of a dead loved one. Our music has power because it describes a reality that only exists for us in ways only we can understand.
At any rate, if the next big frontier for AI ends up being music (and I think if Google has its way, it will be) I think that there will be a bit of an industry shake up but, ultimately, musicians who play well live are likely to thrive in an environment where everything sounds “perfect” all the time so long as you pay no attention to the person behind the curtain.
I hurt myself today
To see if I still feel
I focus on the pain
The needle tears a hole
The old familiar sting
Try to kill it all away
But I remember everything
What have I become
My sweetest friend
Everyone I know
Goes away in the end
Read more More Than Meets the AI here!