Hacker News Reader

We are already drowning in noise, there's too much blogspam and other BS. Now that algorithms can produce orders of magnitude more BS in a fraction of the time, the question of “is this article even worth reading” only gets harder. I'm already suffering from “link fatigue” as it is.

The answer to that is, obviously, another AI reading the AI created content. Should make training of new content AIs so much easier!

I don't see how Medium or even the concept of shared content has a future.

I am sick of chatGPT in its current form but a future AI assistant with even the slightest bit of personalization will put Medium out of business.

"assistant, show me something interesting" I would think is the future prompt. The slightest customization based on browsing history would put the assistant far beyond anything someone else can show me on Medium or any other platform.

I just can't see articles scored by humans as worthwhile has any future. "Interesting" is not objective or statistical, it is entirely personal.

Most of these currently popular content sites are going to go the way of Myspace. They are just tinkering with myspace music or whatever as if it matters.

> I suppose if the text is good and adds value and is factually correct then it does not matter if it's a human or an AI assistant who wrote it.

The real issue is the bad content. It’s perfectly acceptable for a human to publish a bad article/blog post but when language models do it it’s more or less spam.

Why is it less spammy just because a human wrote it? Plenty of humans produce spammy content.

The difference is the scale it allows. The amount of writing a human can generate is relatively limited, even if it is bad writing. GPT and its equivalents allow you to turn electricity into spam with minimal human intervention. That allows unprecedented levels of spam of sufficient quality to make it hard to detect without being examined more closely.

Spam is spam whether made by a human or an AI.

Good, valuable content is good whether made by a human or an AI.

Bad or uninteresting content is spam if it’s being pumped out by an AI, but it’s a learning experience/personal expression if it’s made by a human.

The difference is intent.

It is not, and most of the internet is already trash. We're speaking a lot about scale - but there are currently billions of humans and most of what they put online is already trash. It's not going to materially change - it is currently almost impossible to manually (on your own, at least) figure out what is good, and it will be impossible in the future as well.

I've definitely used AI generated text to help me out with suggestions. It's pretty nice when you have a bit of writers block. All the important stuff like facts, numbers, and code examples should be written by yourself in my opinion.

And how do you trust a reviewer? They might be an AI, too.