Comment by matheusmoreira | Hacker News Reader

Comment by matheusmoreira | Hacker News Reader

Parent Full thread

Please define "substantially". Exactly how many words must come from my hands and brain in order for a work to qualify as human?

Imagine that I write an article manually and keep it to myself. Then I prompt my favorite AI model and it somehow generates the exact same article, against all odds, despite having never seen it before. Then I publish the AI's article, which is equal to mine. Is the article rubbish?

> I can send a prompt myself, don't waste my time republished AI amswers.

I can do it too. I consider it rude to reply to people with AI answers during conversation. Not because I hate AI, but because it made the social interaction pointless.

That doesn't somehow invalidate the value of an AI generated article though. Not everyone thinks to prompt it with that exact idea. Not everyone has an AI they can prompt. Not everyone wants to waste their limited tokens on the idea.

I realize that reading Claudish can be aggravating but that doesn't somehow eliminate the ideas contained in the article. Those can still be very valuable, which is why I advocate for quality filters.

That's a little metaphysical for my taste. If its the same bytes of text, provenance is destroyed anyways, so who knows if the tree fell in the forest.

"Substantially" to me would mean not your ideas and not your writing. I'm sure there are borderline cases that are hard to decide but a rule of thumb for me would be if it takes more time to read than it took the "author" to write.

The issue is metaphysical. If AI can produce the exact same output a human would, does it matter who made it?

It's possible for the ideas to come from the author while the writing is outsourced to someone else. Ghostwriting is an established practice.

Effort involved in writing sounds a lot like the "sweat of of the brow" doctrine, which is rejected even by copyright monopolists.

It's about effort into thinking. If no effort went into either thinking or writing, then I'll plead guilty to sweat of the brow doctrine. Again, just send me the prompt and I'll feed it to AI myself.

Reading is about learning interesting new ideas, enjoying a narrative. The suffering of the author is beside the point.

I can sympathize with the notion that AI writing currently sucks, but not that it's automatically invalid or deserving of mockery. Maybe one day we'll have far better models. Maybe they'll do a good job without inserting any load-bearing vocabulary.

Asking for the prompts is fine, but might not be sufficient. The process iterative, and may incorporate data that won't be published. For example, I recently published a Sol/Fable benchmark article. It would be trivial for anyone to prompt a new version of the article if they had the raw benchmark data, but this data will never be published.

It's not about how good or bad the AI writing is. It's about the human interaction, if someone outsourced their thinking to AI then they are bullshitting me and I will rate their thoughts appropriately at the human level.

You can value human authorship without automatically ridiculing anyone using AI.

If someone's already burned the trees down to generate it once, why would you want to burn more trees, just so you can do it yourself? Do you hate the Earth or something?

Reply on news.ycombinator.com