I don't want anything your AI generates

I'm not sure your premise even makes any sense here, because it doesn't take an artist much more resources to produce art then it took them to just exist for the same amount of time. They're still just eating, sleeping, making basic usage of the computer, using heating and light, and so on either way. Whereas someone using dall-e is doing all of that plus relying on the immense training costs of the artificial intelligence. That basic usage of the computer in order to use the machine learning model might be shorter than the basic use of the computer to use procreate or something, but they'll still be using the computer for about the same amount of time anyway, because the time not spent not making art will just be shifted over to other things. So it doesn't seem to me like having machine learning models do something for you instead of learning a skill and doing it yourself will really decrease emissions or energy usage noticeably at all.

Furthermore, even if there is some decrease in emissions using pre-trained machine learning models over using your own skills and labor, the energy costs of training a powerful machine learning model like you're thinking of are way higher than I think you are imagining. The energy and carbon emission cost of training even a 213M parameter transformer for 3.5 days is 626 times the cost of an average human existing for an entire year according to [this study](https://arxiv.org/abs/1906.02243). Does using a pre-trained machine learning model take that much emission out of people's lives? Or a day's worth out of 228,490 lives, perhaps? I doubt it.

But we aren't even using such a small transformers anymore either — they actually aren't that useful. We're using massive models lile GPT-4, and pushing as hard as we can to scale models even further in a cargo cult faith that making them bigger will fundamentally qualitatively shift their capabilities at some point.

So what does the emissions picture look like for GPT-4? The study above found that emissions costs scale linearly with number of parameters and tuning steps as well as training time, so we can make a back of the napkin estimate that GPT-4 is 8,592,480 times more expensive to train than the transformer used in the study, since it is rumored to have 1.76 trillion parameters versus the 213 million of the model in the study, and GPT-3 was said to take 3640 days to train (despite using insane amounts of simultaneous compute to scale the compute up in conjunction with the scale of the model) versus 3.5 days. This in turn means it is 5,378,892,480 times more expensive to train a GPT-4 than it is for a human to live for one year. And again, to reiterate, no matter what work the humans are doing, they're going to be living for around the same amount of time and using roughly the same amount of carbon emissions as long as they're not like taking cross country or transatlantic flights or something. So it's more expensive to train gpt4 then it is for almost 6 billion people to live for a year. I don't think it's taking a year's worth of emissions off of 6 billion people's lives by being slightly more convenient than having to type some things in or draw some art yourself. And there are only 8 billion people on the planet, so I don't think there's enough people to spread smaller gains out across to justify the training of this model (you'd have to take a days worth of emissions off of 1,963,295,755,200 people to offset that training cost!), especially since in my opinion the decrease in emissions of using machine learning models would necessarily be absolutely miniscule.

> So it's okay to learn, but only if you do it very slowly?

No, it's a fundamentally different process with different results. An artist learns from previous artists to express things they themselves want to express. An AI digests art to become a viable(ish) tool for people who want to express themselves, as long as that expression resides somewhere in the weighted averages of the art the model has digested. Two fundamentally different things, apples and oranges, and, also not without it's own set of limitations. Despite the rhetoric around this stuff that anyone can create anything, that's just not true: you can create anything that you can find a model that's suitable to create it, that was itself trained on a LOT of similar material to what you want to create. Effectively automated ultra-fine scrap-booking.

Honestly if creativity is your thing, even if you find creating difficult for whatever accessibility reason you feel like pretending you care about, you will find AI more frustrating than anything, and the bounds of your creativity are the model itself, and the safeguards whatever provider has decided are important to put in place. You've just exchanged one set of limitations you probably can't control for another set you definitely can't control.

> I surely don't need to point you to the existence of forgers - you know a human can study the art of Rembrandt and paint plausibly Rembrandt style paintings

Yes and those are worthless once found, just like AI art. And again, you've sidestepped the scale: Adobe Firefly can bash out 3 images in roughly 2 minutes of solid resolution. No human can even dream of getting close to creating Rembrandt forgeries at that rate.

> We have been doing that since the dawn of history - what makes this any different from cars obsoleting the horse drawn carriage?

Because cars costed a fortune when new and were toys for the wealthy, before Henry Ford came along some three decades later to fix that. And then, the former farriers had time to retrain for new work. Not to mention, carriage builders were still employed through many decades with the rise of cars, because originally buying a "car" meant you got a chassis, suspension, engine and the essentials, which you would then take to a coach builder to have a "skin" if you will build around it. Hence the term "coachwork."

> Computers have been automating people's jobs for decades - should we ban programming writ large?

Is this the "debate" you were saying you were open to? Hyperbolic statements with zero substance? I can see why few want to have it with you.

> Where, exactly, do you feel the line ought to be drawn?

Consent. Tons of people's creative output was used to build machines to replace them, without their consent and more often than not, explicitly against their wishes, under the guise of a "research project" and not a monetized tech product. Once again a tech company bungles into the public square, exploits it for money, then makes us live with the consequences. I frankly think that question aught to be reversed: what makes OpenAI entitled to all that content, for a purpose it was never meant for, with zero permission or consent on the part of it's creators?

I'm not opposed to ML as a concept. It has it's uses, even for generating images/writing/what have you. But these models as they exist now are poisoned with reams of unethically sourced material. If any of these orgs gave even the slightest shit about ethics, they'd dump them and re-train them with only material from those who consented to have it used that way. Simple as.

It's evolution and our responsibility to make it good.

Doesn't matter though if we like it or not because it's happening.

The only thing preventing this is a total economy collapse so crazy that our society doesn't continue chip production.

Loud and public objections are the opposite of not caring.

[dead]