21–30 of 275 posts
The fear reminds me of the early days of hip hop, where songs remixed from the past were blasted as unoriginal. I think we've all mostly agreed now that you can build new content that honors the old whole being completely fresh.
1. Before 1970
2. By someone who has consumed AI-generate content
3. With a ~2000-era spellchecker
4. By someone using ~2020-era neural speech-to-text software
5. With a ~2020-era spellchecker
6. By someone with an AI in the traditional editor role (reads it, gives detailed feedback)
7. By a human and an AI working together where the AI generates everything initially but the human fixes things and nothing goes out without human endorsement.
I'd probably draw the line at 7, but you could also argue for 6 or even 5.
The history of the term "handmade" and discussions about what to allow on Etsy come to mind: https://whileshenaps.com/2013/10/etsy-redefines-handmade-aut...
[edit] Actually, just to expand on this a bit, this is effectively an argument for establishing a mark of authenticity that all literate civilizations have always striven to place on their intellectual output. There has never been a single rule uniting these efforts, but civilizations which placed more emphasis on safeguarding and defending the precision of the written word have tended to be rewarded with greater longevity. There's no reason that trend shouldn't accelerate when faced with the threat of inundation by meaningless language models. Just like gold coinage, we're looking at a period of debasement and inflation.
You could actually argue that language models themselves are an expression of anti-semitism, in the sense that they're an attempt to undermine the sacredness of the written word, to destroy or wash out the way that the meaning of words ennobles humanity, and to eradicate the special relationship that the law of language and the language of law create between God and Man. I only say it seems anti-semitic because that particular concept, as a high/sacred value, seems unique to Judaism (from my perspective, I can't think of another culture that considers it an inviolable precept) and so this attempted abolishment of the human hand in the written word seems particularly targeted at those who consider the word sacred; maybe this is yet to be threshed out. Maybe Bari Weiss will write about it once some nazis have ChatGPT come up with a totally bunk but plausible corruption of the Talmud. But love of the written word is something that should rightly be a general human value, because we'll live or die with it, Jews and everyone else, whether we want to or not. All civilizations fall when their coin is debased, and our coin today is information.
[edit2] also, I'm drunk, and I love y'all. I hope this stimulates debate, not hate.
There is so much stuffing for a simple idea that I'm not sure if this piece deserves its own title, but I'll give it the benefit of the doubt. One thing that I wonder though is how we will draw the line. If I'm writing a piece and do a Google search, and in that way invoke BERT under the hood, is anything that I write afterwards "AI-tainted"? What about the grammar checker? Or the spot removal tool in photoshop or gi…
We could get this whole discussion back to some semblance of sanity if we stopped calling any form of remotely complicated automation "AI". The term might as well be meaningless now.
Nothing about any of all these "AIs" is intelligent in the sense of the layman's understanding of artificial intelligence, let alone intelligence of biological and philosophical schools of thought.
> Published content will be later used to train subsequent models, and being able to distinguish AI from human input may be very valuable going forward I find this to be a particularly interesting problem in this whole debacle. Could we end up having AI quality trend downwards due to AI ingesting its own old outputs and reinforcing bad habits? I think it's a particular risk for text generation. I've already run into…
As we move forward, suspect we will see an increase in curation services and AI models will do more with less. You can bootstrap a productive adult human on an almost infintismal slice of the training sets we are using for the current gen of AI, can't imagine future approaches are going to need such an unbounded input to get better results - but might be wrong!
If content is curated for its quality, whether or not it's AI generated (or assisted) doesn't matter.
Why we use text? Half of it is about getting something from someone else BUT even more importantly we write text to change something in the worlds using our words.
The problem with AI pretending being a human only exist when that AI doesn't get anything from us and it's only good for extracting information from.
It's utterly futile to discuss with AI generated content here on HN but it's amazing experience on ChatGPT precisely because when we write each other something here on HN, we expect that our words will create some impact somewhere. Some option will change or we will befriend someone.
I have 0 problems with having interacting with AI which is an individual machine in Lisbon, and it is learning and evolving as the life happens. On the other hand, I hate the idea that I'm talking with a machine in Palo Alto and the only output of my conversation is some statistics that VC will gaze over and optimize for his own gain.
Just think about it how meaning is transferred from human to human: we can compress meaning in few markings and extract it on the other side only because as individuals we have experienced life and some markings saying "bored" is enough to transfer very complex situation from person to person. In the current state of the AI, being bored doesn't mean anything to LLMs but if individual AI machines lived human-like lives, I think it will start having meaning for them too.
IMHO, the problem with AI generated content currently is exactly the same with SPAM or other non-genuine content and has nothing to do with it's origins being biological or electronic.
We people will put eyes on a ball and call it our friend, we don't have problem with it.
> do you really want to cry watching a movie that was 100% produced by robots? The fear reminds me of the early days of hip hop, where songs remixed from the past were blasted as unoriginal. I think we've all mostly agreed now that you can build new content that honors the old whole being completely fresh.
EDIT: Okay, now I'm shocked to see your comment grayed out all the way at the bottom. Interesting times we live in.
Meanwhile you never get any results from 4chan or KiwiFarms (sites with much more organic content), unless you go out of your way to specifically ask for it.
"AI-free" is pretty clear. "Organic" is much more subjective. The same applies to Organic food, if I'm not mistaken.
> Published content will be later used to train subsequent models, and being able to distinguish AI from human input may be very valuable going forward I find this to be a particularly interesting problem in this whole debacle. Could we end up having AI quality trend downwards due to AI ingesting its own old outputs and reinforcing bad habits? I think it's a particular risk for text generation. I've already run into…
Step 1. Release a tool that generates nonsense code across a thousand repositories, and allow anybody to publish crap to GitHub.
Step 2. Copilot trains on those nonsensical repositories because it can’t distinguish them from the real thing.
Step 3. Copilot begins to generate similar crap.