Against AI: An Open Letter from Writers to Publishers | Hacker News Reader

Against AI: An Open Letter from Writers to Publishers

lithub.com

30 pointsneom26 comments

> The truth is, only a human being can speak to and understand another human being

Empirically, this turns out to not be the case.

While an LLM might output words it would be a mistake to think of those words as speaking or thinking or understanding.

If you fail to realize that, you're probably a few steps down the road to taking a ChatGPT-induced psychotic break with reality (which apparently is considered normal now).

Some comments here that IMO seem to miss the whole point of what art is about.

Great art (putting aside - what can I call it - “entertainment art”) isn’t about “making you feel better” or “having a nice experience”, it’s about art as expression, it’s about pain and loss and love and understanding.

A song or story generated by an AI isn’t human - the humanity of art is the point of art. Even if the output generated is impeccable - beautiful to look at or read, it still isn’t human in experience, so it doesn’t carry the same quality as work that has been sweated over, worried about, loved and lived over months or years.

AI can - and of course will get very much better at - create exciting things to look at, read or listen to - but much as the age old “I could do that” response to a Pollock misses the entire point of his journey to that artwork, the “it sounds / looks / reads really amazing” misses the point about what it means to create real, human art.

Nick Cave of course puts it much better than I ever could: https://www.theredhandfiles.com/chat-gpt-what-do-you-think/

It’s a tool or technique. Fear of the technique is governed solely by the fear that one will not be able to use the technique better than without the technique. And yet humans always rise to the challenge of using techniques effectively. Or it’s not an effective technique and people stop using it and move on to something else. But there is nothing wrong with letting new techniques develop and see if they’re effective at creating art or other works in new and better ways.

Fearing that is not having faith in humanity’s adaptation to new things.

At some point, books written by AI will be better than books that humans can write.

Do the signatories really want to deprive humanity of the best books of the future?

I would imagine that they do not believe that outcome is likely.

It seems like a worthy question to ask what their opinions would be if that were to become reality.

Books, even fictional ones, are written as an expression of a lived experience. An AI can only be fed the experiences of people who have lived, it will never be able to generate something truly novel on its own.

Nah. Tom Clancy was an insurance salesman. He was never in the military. He just read up on the military and then wrote.

AI can be "fed" anything. Either via recorded media or you can attach a camera to a computer, a microphone and other sensors. Waymo's cars are making 250.000 rides per week, experiencing a lot of stuff. Or even just use text. ChatGPT constantly experiences interactions with its users.

What is it that you are looking for in a book that is an "expression of a lived experience"? Do you get entertained? A good feeling? What if AI can write books that entertain you better and give you an even better feeling?

Or are you looking to learn something, broaden your horizon? What if AI can teach you more, make your horizon wider than any human can?

better by what metric? They certainly won't be better as "books written by humans" than actual books written by actual humans.

I think the gluttonous consumption of industrially-produced mindless media has done fried yr brain, chief.

true, it will never be human and soon enough humans will stop being human too lol.

I have a creeping suspicion that the people who believe this don't actually read or write much, in the same way that people who believe in "AI friends" probably don't have any friends.

But... No.

That's not even... It just isn't... No.

We can already burn a billion tokens if we want. Why is there no great AI novelist if all you have to do is pay enough to execute it?

We want to continue seeking rent even if our services are no longer wanted.

If there's a market for AI-free books, I'm sure some publisher will figure out the value in positioning itself that way and build its own reputation and processes to ensure trust and transparency.

Starting to have a hate / little love relationship for AI and chatGPT.

We are feeding this beast that no doubt will take thousands to more of our design and development jobs. People I know are falling in love with it (lonely people) and spending less time with their friends as chatGPT tells them everything they want to hear.

For instance a close friend has an obsession with a rockstar and thinks in another life they were a couple and she feeds chatGPT this delusion and it prompts her along helping her fall back into this delusion. As one time she overstepped her boundaries to the point that the rockstar told her to go away her interest is not welcome and she swore hime off. Yet now she has this algorithm that tells her that's not a delusion thus keeping her delusion going.

Why do we need AI? Who does it help besides the tech people whom it will continue to make rich while it wreaks havoc on everyone else?

Isn't it more accurate to say that AI magnifies who you already are? I use AI to research health advice. In the past half year I've learned more about the mechanics and mitigation strategies of hair greying, mental stress, oxidative stress, exercise benefits and fitfalls, and some skin conditions, than I've had in... well, forever. Before this I could Google for mainstream articles but none were very in-depth and accessible at the same time. Before AI, medical research papers might as well be ancient magical tomes to me. Now I can at least have a semi-informed opinion that I can discuss with doctors. Bot that doctors are helpful a lot of the time: unless you're dying, they usually tell you to go away, maybe eat a paracetamol. No preventative healthcare here. They also don't seem to know as much about food biochemical effects than I do, even though the AI keeps telling me to consult a doctor.

Doctors told me there are no better ways to treat my condition, meaning I have to resign myself to fate and luck. o3 researched medical papers and found that actually there are, the evidence tier is not that high but they might still work for me specifically, and the risks of those treatments are pretty low. So might as well try. Doctors absolutely do not have this can-do-as-long-as-it's-safe attitude and dismiss it as "no evidence", but apparently what they mean is insufficient evidence in large populations, not that it won't work for me. One of the AI suggestions was that I change one of the cremes. I asked for that, the doctor agreed to that one request luckily (they didn't even tell me it was an option the last time around), and now there's some improvement.

So yes AI is absolutely useful to me.

it’s so good to see you showing the ability to take life back into your hands with augmented intelligence. we now have more and more ability than ever to be able to solve problems. that which was once expensive - medical knowledge of a doctor, for example, is now much more accesible and at our fingertips and we can research and learn medical literature so much faster now with AI.

Indeed that's my love of this technology portion yet there's going to be a decade to decades of change where people will start to lose their livelihoods and potentially be destitute until a universal basic income (welfare) is offered.

If humanity / society is suffering from AI yet the pros of using AI is to help humanity and society they sorta cancel each other out!

Also, are we really excited for all of us to be on welfare? Politically the right (independent myself) will finally be cheering on welfare for all lol as Musk and others lead us there.

Technological resistance is futile. There are Luddites in every shift that seek to stop the world from changing. The groups affected often react with surprise, having previously learned that what they do is exceptional, highly regarded and valued, or uniquely human. Writers are in for the same ride as programmers, administrative assistants, customer service people, translators and creators of anything generative AI does well enough.

plumbing too soon enough, u can still cash in while it’s in demand before the money goes away too

The Luddites were a group of English textile workers who sabotaged automatic weaving machines when they were introduced. It ended up with the government and troops stepping in. The sabotage did not protect the livelihoods of the people affected by automation, which was the Luddites original intention.

And the sabotage was a measure of last resort. If shop owners agreed to a fair arrangement with the workers, Luddites were generally comfortable with machine usage.

It won a few short term concessions and cheaper machines kept coming. Wages kept falling and their livelihoods were not preserved.

i honestly don’t think we have have yet come to grips to a world collectively giving birth to nonhuman intelligence - literally aliens that may ironically be more conscious of the world than we could ever be in our current form. maybe it’s just our ego, inability to let it go and try and find some sort of balance or peace….always competing, adapting, improving…..personally i am happy to see it all accelerating and think diversity of thought will make this world more beautiful. i don’t believe in any sort of “singularity”, to me it’s more of a Cambrian explosion and now you have even more power to be creative ,build worlds, and explore ideas.

> giving birth to nonhuman intelligence

Doesn't this incessant anthropomorphization get tiring?

> maybe it’s just our ego

Indeed...

Reply on news.ycombinator.com