If you’ve been peeking in on my Thoughts From the Throne, then you know I’ve been having some fun with ChatGPT. Or at its expense, take your pick. I’ve also been playing around with a lot of the other AI tools out there in everything from image creation to music design.
So I’m getting more experience with these tools all the time and, sitting on my throne today, I decided to take some time away from them and throw out some neither-here-nor-there, off-the-cuff, Presidential-level-tweet thoughts about it and our future as creators.
Cause there’s a lot of fear out there, and uncertainty, and hype-or-not-hype, ranging from “it will end us as a species and extinguish the light of consciousness from the universe” to “it will guide us into a utopia of endless abundance and creativity.”
So, big stakes. And that’s a pretty wide, confusing range. So, what are the best takeaways as it stands right now, and what do I think is likely for our future?
Where We’re At
First, I think even the biggest doomsayers would admit that ChatGPT 4 isn’t going to go Skynet. GPT 5 won’t either. They’re worried about something else, AGI, that could be a year away or a century away. No one really knows. They don’t even really know if it’s possible.
So, the extreme on that side is about what could happen in a somewhat distant future. No repercussions with these tools as they are. Ditto for the opposite, “AI will bring utopia” side.
That limits this for now to a disagreement between those who think it’s moving too fast for us to adapt, and those who think that it offers too much for us to slow it down, with any attempt to do so leading to us losing to another society without as many scruples as ours (ahem, CCP China).
These are still pretty serious stakes, and as a writer, I know many people who are very worried about losing their jobs or even the point of what they’re passionate about. Just look at the WGA strike, which began as a fight for more residuals and has become as much or more about AI.
Writers and other creators fear that AI will let the executives of the world effectively take their role from them, and they’re a stand-in for others who fear effectively the same thing.
Some intellectual heavy-hitters back them up here from a societal view more generally. People like Yuval Noah Harari, who wrote recently in The Economist that AI has “hacked the operating system of society” because by manipulating language, it can manipulate our thoughts.
“In a political battle for minds and hearts, intimacy is the most efficient weapon, and AI has just gained the ability to mass-produce intimate relationships with millions of people.”
He and a number of others, including Elon Musk, signed an open letter calling for a cessation of AI advancement for at least six months so that society can catch up. While this is mostly symbolic (probably), it calls attention to the concern.
And yet, I know firsthand how people are diving into these tools and putting them to work, and while Sam Altman can go in front of Congress and claim he’s as concerned about ending human life as much as he wants, there’s no sign he’s slowing down on research.
So, are we headed for a misinformation poo-pool of Bladerunner proportions, or a…what? I’m not even sure what the positive vision is here.
What Does it Really Mean?
Well, scratch that. I kind of do. To get there, I’ll go through my own personal experiences with this stuff.
I’ll start with ChatGPT, both because it’s the most popular of these tools and because it’s probably the one with the widest range of use cases. It’s a lot of fun to use. I’ve found it can be a great brainstorming tool and a great way to get a pre-vomit draft (a gag draft?) out in a second.
It’s also really fun to mess with it. I at least make myself laugh doing it.
But it has its limitations, and I think some of them are never going to go away, no matter what number you put after the T.
Now, I can already hear the people thinking, “JuSt WaIt uNTil GpT 25!” And yes, there are many areas where later iterations will improve on what exists now. But there are some limitations I don’t think it will ever overcome.
First, it has no POV. It’s not writing with any real voice or outlook, because it doesn’t have one. Despite the sensationalist articles that have come out, there’s no reason to think that it’s conscious. It has no perspective unless you give it one, and even then, it can only imitate.
Yes, it can write something that seems original, but however technically adept it is at doing that, it has no way to impart its own spin on something. It’s just remixing what you tell it to. And however advanced it gets, that’s all it’s likely to do.
I’m not alone in this outlook, either. Here’s Jaron Lanier giving an interview about AI. You can read the transcript here. He’s one of the pioneers in VR and is working at Microsoft on AI.
Basically, he puts forward the idea that we make a mistake treating AI like it’s a human, one that limits us. Really, these AIs are remixers.
And that’s what I’ve found. These tools are best thought of as samplers.
In music, we’ve gotten used to artists sampling other artists and using their stuff all the time. But when rappers and others started doing that in the 80s, it was seen as vulgar, even theft. Well, like so much else on the internet, what started with music has now gone into every other media.
All an AI does is take humanity’s collective output from the internet and remix it for you. Want a painting of a toilet in the style of Salvadore Dali? Midjourney will do that for you. Want an HR statement written as a Shakespearean sonnet? ChatGPT is happy to oblige.
And because that’s what AI is, I have a different view than what seems to be the prevalent one right now. I think that rather than not empowering artists, AI is going to make them able to create more work of a higher caliber for a cheaper price.
In fact, if I was an executive or suit who was used to having artists depend on me for a paycheck, I’d be worried.
What the Future Might Be
In a world where creating media is so cheap, the best work is what’s likely to stand out, not necessarily the work with the most money behind it.
Using the WGA strike as an example, I think the WGA has this backward. It’s not writers who should be worried about Netflix making stuff without them, it’s Netflix that should be worried about writers being able to create without their money.
To give the devil its due, though, this will also make creating BS as easy as firing off a tweet from the toilet is now.
But that’s an article for another day. I’m finishing up in here. Let me know what you think in the comments. Am I crazy, or am I onto something?
Neil