[This version has embedded images]
A long time ago, some primitive apes got addicted to rocks.
The earliest stone tools were crude bastards, made by smashing large river pebbles together and calling it a day.
Stone choppers like the one above took the prehistoric neighborhood by storm almost 3 million years ago. However dull the tools themselves may have been, this was the cutting-edge technology for literally more than a million years, a timescale I have no capacity of comprehending. Not until around 1.7 million years ago (again, no idea what this means) that someone got the bright idea of chipping away both sides of a rock. You can see what the (tedious) process looks like.
The end result is the unassuming tear-drop shaped hand axe, by far the longest used tool in human history. There are no accessories here with the hand axe, its name comes from the fact that you use it by holding it directly with your hands.
On top of being tedious and painful to make, you can imagine that it's not terribly comfortable to hold while using. Hand axes also have to be somewhat bulky because of the necessity of combining the sharp useful end with the blunt holding end. But what if --- stay with me for a second --- instead of holding the thing directly with our pathetic squishy hands, we held something that "handled" the tool for us? It took humans about another million years to discover hafting, with the earliest examples from around 500,000 years ago but the technique didn't really find its stride until the microlith era of stone tools around 35,000 years ago.
Then humans found metal.
"Technological advance is an inherently iterative process. One does not simply take sand from the beach and produce a Dataprobe. We use crude tools to fashion better tools, and then our better tools to fashion more precise tools, and so on. Each minor refinement is a step in the process, and all of the steps must be taken."
-- Chairman Sheng-ji Yang, "Looking God in the Eye"
The historian Bret Devereaux has an excellent and highly-recommended series on the history of iron. The popular depiction of iron being a rare commodity (typified within medieval and fantasy genre) obscures some of the reality. As a material, iron is extremely abundant --- the fourth most common element in the Earth's crust, making up 5% of its mass. The hurdle with iron wasn't finding it but rather getting it out of the ground and into a useable form. It required a lot of dead trees and broken shins. One of the illustrations Devereaux cited is from 1556, and shows how workers wore shin protection as they crushed the ore into useable chunks.
Think about how many mangled limbs had to accumulate before medieval OSHA cared enough about this hazard. After the ore is dug out of the ground, the next hurdle was figuring out how to reach the high temperatures needed for processing. Because of how finicky iron is about absorbing too much carbon, the only feasible avenue was charcoal, which is made from wood, which is cut from many many trees. As Devereaux notes:
To put that in some perspective, a Roman legion (roughly 5,000 men) in the Late Republic might have carried into battle around 44,000kg (c. 48.5 tons) of iron -- not counting pots, fittings, picks, shovels and other tools we know they used. That iron equipment in turn might represent the mining of around 541,200kg (c. 600 tons) of ore, smelted with 642,400kg (c. 710 tons) of charcoal, made from 4,620,000kg (c. 5,100 tons) of wood. Cutting the wood and making the charcoal alone, from our figures above, might represent something like (I am assuming our charcoal-burners are working in teams) 80,000 man-days of labor. For one legion.
To understate it, much has changed since. A stainless steel spoon today is a trivially manufactured artifact. But just the material from that spoon would have represented thousands of times its weight in stone and tree, all excavated by hand. I think about what this spoon, held in the palm of my hand, would have previously cost in terms of human toil and crushed limbs.
This post is about AI.
I feel like I'm holding a hand axe right now, while everyone around me is revving up their chainsaws. I feel like I'm a peasant awestruck at the intricacies of a steel spoon, unaware of its bargain bin progeny.
It's difficult, and exhausting, to keep up with the pace of AI developments. I also question my ability to make any sort of concrete or realistic predictions in this field, so I'll try to keep it semi-grounded in the present.
What already seems evident is that, even if we assume a complete halt to any further developments, content creation is already utterly trivialized. Do you want a picture of a cat riding a unicycle while smoking a hookah? Here's 50. Do you want those same drawings but done as if Picasso was tripping out on LSD? Done. Do you want the script from an 80-episode television series involving these psychedelic Picasso unicycle cats as they work to solve a murder mystery on a cruise ship in a black hole? And you want each cat voiced by a different rap artist from Kanye West to DMX? Why not also make it a choose-your-own adventure series controlled by each viewer? Sure, whatever, done. Some of these require a little work to stitch together, but you can have it all.
Part of where my feelings are settling are a bizarre mix of trepidation, ennui, fatigue, and...excitement? I'm not the only one to ever experience mild frustration that a given movie, TV show, book, video game, etc. wasn't exactly just right, and if only the creators changed this one thing that would've been so much better.
I encounter this feeling constantly with video games and for that same reason I tend to gravitate towards extensively modifying big-budget video games to my liking with mods. For a period of time, I definitely sunk in more hours finding, installing, and configuring Skyrim mods than actually playing the game itself. This was only possible because other people were insane enough to pop the hood open and get their hands dirty. If I wanted cold weather survival elements added to Skyrim, I was lucky enough that someone else had the gumption to analyze the game files, draft up pseudo-scripts, and collect custom-made assets into a coherent package that actually worked.
I also appreciate the estorism of open-source oeuvres made entirely by coding hobbyists, like the suburban apocalypse simulator Cataclysm: Dark Days Ahead. Cataclysm is a jury-rigged amalgamation, cobbled together over the years by dozens of drive-by developers. Some aspects of the game are painfully undercooked, such as the lack of any real ending, while others are pathologically overdeveloped, such as the ridiculously intricate vehicle physics system which manages to accurately simulate drag resistance in a game where no one will ever the difference. The only reason there's any progress made on these projects is because there are enough enthusiasts roaming around with actual coding talent, but they'll only chase after their own whims and then move on. Anyone else with ideas either has to convince one of these sensei to take up their cause, or drudge through hours of coding tutorials on YouTube to ever stand a chance. Lots of fields stay fallow then.
Outside of play and in the realm of work, much of my time is chasing after tedium. A few tasks manage to reliably trigger my procrastination reflex with the main one being legal research and writing. Let's say I'm trying to have incriminating statements or evidence suppressed. If the scenario is even slightly interesting, I am not likely to find a case precedent within my jurisdiction that is perfectly on point. Instead, I dump a few search terms into a legal database and then spend hours with dozens of tabs open, dutifully reviewing each hoping I can find enough adjacent precedent to triangulate into an answer into my own case. Judicial opinions are almost never written in a uniform manner, so I often find myself realizing a given case is worthless only after already wasting several minutes reviewing it. After all that research, I have to synthesize it into something legally accurate without boring the overworked judge to death.
It's all tedious boring work. It's also a perfect use-case scenario for chatGPT because it would be trivial for me to just ask it to quickly find and summarize whatever is analogous to what I'm looking for, then write something custom-tailored. The day that Westlaw incorporates chatGPT is the day that Thomson Reuters will become a pseudo-branch of the Treasury Department, for its ability to just print money from the legal profession. To be clear, my concern here is not job loss. I imagine that with greater productivity comes greater expectations, especially with AI helpers at our side.
I wonder, why bother with any of it now?
On the consumption side, whatever game I choose to play now will only get way better in a few months as I'm able to trivially customize it to my mind's whim. Same with whatever television show, or movie, or book. Or existence.
On the production side there's so much more I want to write but I also wonder, why bother writing anything if it's just going to be swallowed up whole and incorporated into the labyrinthian halls of a Borges infinite library. Realistically the only effect this post will ultimately leave upon the world is a faint whisper of an errant memory. The rest will either be carved up into individual tokens or buried under a figurative mountain of indecipherable pages. I see the entire corpus of mankind's creative output as a tiny ship, a gnat really, about to swallowed by a towering ocean wave. Part of me just wants to sit and wait for the flood.
I wrote this entire post without chatGPT, to prove something I guess. It took hours. I had to look up some new concepts, read enough to understand them, revisit old essays I read, and review them to refresh my memory. After all that, I had to use my dumb fingers to tap buttons on my dumb keyboard, over and over again.
I'm the idiot holding the hand axe. I'm the imbecile mangling my shins with rock debris. Why bother?
Jump in the discussion.
No email address required.
Notes -
Hmm. How to explain how I feel.
"What would it be like?"
A being born after the singularity might ask-
"To have been born on ancient earth during the rise of machine intelligence? To integrate capability after capability into their being that had spent years beyond their reach? Would it be thrilling or frightening? Being so limited and watching those limits fall away?"
And so they go on to simulate the whole experience, just to taste it.
So what are you waiting for? You can already ask GPT how to mod Skyrim.
And by the time you've finished learning how the engine works, Text to Mesh will be ready for you to expedite the models.
You don't know whether you're the one in the sim or the one in the original origin story. And either way, we're not here for the ending. We're here for the taste of it.
More options
Context Copy link
This is equivalent to any other fatalistic argument, is it not? Why do anything when you will be forgotten, since you will never surpass the greats, since there will likely be countless betters in the future, when your civilisation and culture will succumb to obscurity, when the Earth will end, when the Universe will end... render unto yourself what is yours. No need for concern about the passage or time unless you can figure out how to stop it, or much else that is outside of your control. A few months is yet to pass, etc.
More options
Context Copy link
"A person had to work hard for it, but a minute of perfection was worth the effort. A moment was the most you could ever expect from perfection."
More options
Context Copy link
My fear of AI is that corporations and politicians will one day realize that blindly trusting the outputs is (in the long run, on the whole) better and more expedient than ever checking the outputs. Why check the work when we know that trusting it without wasting time will make us more money? This is what our competitors are doing, after all. So a political organization will be creating the perfect propaganda, the marketing corporation will create the perfect marketing, and the AIntertainers will be creating the perfect entertainment, all without any significant human input. When AI employees are finished at work, they will go home and be captured by the most addictive entertainment, food, AI parasocial friends, porn and so on. In this sense, AI will be creating a consumerism death spiral, with the median human unable to resist the opium of AI, while going to work to expedite AI’s power.
My positive anticipation of AI is that it will allow us to focus on interesting moral questions and questions of happiness maximization. I like this possibility, because it means my favorite hobby will just get better over decades.
According to Nick Land, this has already happened.
We're just going through the motions.
More options
Context Copy link
More options
Context Copy link
I mostly gave up on 'narrative' games due to AAA rotting, and have been wasting time playing games no one ever meaningfully 'wins at', against other people. It's pointless as any game, but at least I always going to end up being surprised and/or outraged.
Meanwhile, in ordinary games, once you figure out a few tricks or the mechanics, it's pretty much over. The only thing that makes them exciting is if mistakes are severely punished, but that's not the norm in games.
More options
Context Copy link
I thought for a minute that the first half of your post was AI generated. But AI's still can't write stylistically with that degree of verisimilitude.
More options
Context Copy link
This feels premature. Your post is still more interesting than anything I've read that I know to be generated by GPT. Just look at the paragraph above this. In about a hundred words, you spontaneously generated and peppered in about a dozen evocative analogies and metaphors that GPT is terrible at. And that's hardly the only place GPT struggles.
You may be useless eventually, but for now LLMs are a labor saving tool. The better comparison for now is that you need to pick up and learn to use the chainsaw. It doesn't know where to point itself, yet.
We may in fact be entering a golden age of hobbyists creating art/mods/open source software projects. Many times in my life I've thought "Wow, I have a really cool idea for X", but to actually implement X would require me spending 100+ hours mastering some boring technical skills like obscure markup languages or a certain style of pixel art. With AI image generation and LLMs? You'll be able to marshal your special (for now!) human creativity and intuition and let this fancy chainsaw cut through all the boring wood for you.
Absolutely, that's how I see it. I don't think that came across as much as I wanted, but I am legitimately excited about the developments. ChatGPT is not good enough to generate interesting prose on its own, but I already use it as a pseudo-editor and thesaurus and summarizer as an occasional supplement to my writing. There's definitely an element where it feels like "cheating" but I'm not a Luddite and that's not going to be enough to stop me from using the chainsaw.
Part of the ennui here is just impatience. The tools out there already have obvious capacity but awful UI, and enthusiasts have to stitch together a workflow across several command line prompts to get something useable. It's not so different from the Louis CK airplane wifi segment, I want it now and there's a reluctance with trying to chop down a tree with a hand axe when the chainsaw is in the mail.
More options
Context Copy link
I'm reminded of Kasparov's commentary a while back on Human/AI pairings for chess games: He suggested that the ideal end state would be what he called a centaur: The human makes the broader strategic decisions, and then AI then executes the plan tactically.
More options
Context Copy link
Yeah, it occurred to me recently that AI could possibly blow open-source software and freeware wide open. You might not have the time and effort to maintain some dependency thing, but the AI can solve your problem quickly.
Check out this example where a guy generates a frontend for a social media site connected to a working database with just voice commands, in less than five minutes.
More options
Context Copy link
More options
Context Copy link
More options
Context Copy link
For the same reason people have done anything since time immemorial: because the process itself is enjoyable. Why do I bother to cook food, when professional chefs can turn out far better food than I ever will? Why do I bother to play the guitar, when I will never even come close to the level of Eric Clapton (for example)? I do it because I don't actually care if someone else can do it better than me, the point is to enjoy doing it myself.
Certainly it's true that the economics of content creation will change. There will no doubt be large swathes of artists who will never turn a buck again (though as has been pointed out, it may just mean that the artist's tools become Stable Diffusion instead of pen and paper). But then again, being an artist has never been a strong money-making proposition. We have the "starving artist" trope for a reason, because it's by far the most common outcome. So even the economic angle is going to be much as it always has been.
IDK, I guess I just don't understand the nihilistic argument here. Presumably you agree that the journey matters as well as the destination. After all we all wind up just as dead no matter what we do, but life is still (generally) considered worth seeing through to the end. So if you agree that the journey matters (as I think you would), then why would the journey not matter in the case of creating art or other content?
Depends why you're creating it, particularly in the case of "other content". I'm sure Eliezer's letter in TIME wasn't written because he wanted to spend time writing a letter. It's because he wanted the letter to exist.
Now, I mean, I doubt he used GPT-4 to write it, given that he'd come off as a giant hypocrite if he'd done so without advertising it in the letter as "it's bigger than you think, panic now!". But if you shelve that, there's a very real question of "why would you ever write a thinkpiece when you could get a robot to write a more compelling one?" (Of course, then the AI company has the question of "why would you ever sell this when you can use it to directly take power via mind-controlling the government and/or voters?", at least if they don't wind up mind-controlled by their own AI along the way.)
More options
Context Copy link
I don't know about you, but I cook because the alternatives are vastly, vastly more expensive.
More options
Context Copy link
More options
Context Copy link