Comment by CosmicShadow | Hacker News Reader

Comment by CosmicShadow | Hacker News Reader

Parent Full thread

This is the annoying thing I've found about AI writers past and present.

Maybe almost a year ago I was playing around with Rytr.me (and their competitors) trying to generate factual content I could use for SEO purposes on hundreds or thousands of band specific pages. Many top sites already have this, a bio of the band, list of albums, awards, some random Q&A's, etc.

This should be an easy job for an AI to do, yet in many cases it was plausible, yet not always accurate. They would miss key facts, point out songs, albums, events that weren't really that important (despite the info being there). Sometimes the data would be off, and I'd have to go look it up, and if I'm going to do that, I might as well just do it all by hand, because I can't trust anything generated. It is a pain enough to do for bands I think I know, but for bands I couldn't care less about, blech.

In theory it would have been good enough, but I didn't want to associate my site with that kind of crappy or incorrect data. Being dated info as well, it felt wrong publishing a page about an artist that died like a year ago and not mentioning it.

I've since circled back around and played with ChatGPT for this, but it's still just off enough that I don't want to use it.

The thing about playing with these tools is that you get really good at recognizing content written by an AI and you really start questioning a lot of things you read online for misinformation. It's kind of gross and scary to see these simple answers appearing at the top of Google for any question and realizing they could be completely wrong, but people don't know.

Just the other day we cut our dog's nail and hit the quick and he started bleeding like crazy, so we look up what to do and find a content blog with what appears to be a few paragraphs with a clear solution, but it also had like 7 other articles showing up with similar queries that all had different answers. I scrolled to the bottom of the Wordpress site and it had 7000+ pages, I go to the last one, they were all added to the site on the same day. Almost every query I typed into Google lead me to this site as one of the first results. It's fucked.

If you don't mind my asking... Is there any merit to the pages you're trying to create?

It sounds as if it feels automatable because it's not really something any human needs to read. Is there anything on that page worth searching for?

I'm so sorry that this sounds judgmental. I don't mean it that way. I'm just curious in that it sounds as if the web really is complete, and we're now just repeatedly rearranging it.

That's kind of the open question with all of these AI projects. We deny their creativity because all they can do is rearrange. But it sounds as if we humans are the ones running out of creativity.

Does a duplication of information need to exist? Definitely not and I don't want to do it (especially if I have to do it shittily with bad info via AI), however the only way to show up in a google search result for what I want is to include this information to trick Google into thinking I offer unique and relevant information so I can help rank up my offering and play the game like every other site fighting to be found online.

I am self-employed and run my own online business that requires traffic and sales. Apart from wasting a ton of my money into ads, I can work on SEO, which only costs time and some effort. I've done this and found that after 6-9 months many of my desired search queries put me in the first 10 results (and without AI or spamming crappy content).

In order to go further and attempt a longtail strategy I offer something customizable and relevant to the fan of every single band and artist out there, so Googling "Band Name [my offering]" would ideally lead to my result being one of the first and having it precustomized to that band. This is how every result is ranked for the biggest players that take up page 1.

Do I want to make the web a shittier place? Definitely not, and unlike many other folks I'm very hesitant to just throw thousands of crap articles that can't provide any value out there to try and get a few ad clicks. I don't have ads on my site. I want to help fans of an artist find the offering they are looking for. I often get amazing feedback and love letters from customers who are so thankful they've found what I sell, but the problem is how to get found, especially when competing for search relevance against some of the biggest websites on the internet. If I can do this instead of giving money to Facebook, I'd be much happier and more profitable.

> Is there any merit to the pages you're trying to create?

No, but maybe they can be monetised for a few extra dollars in ad revenue.

(Do I sound cynical? Yes, I'm tired of the recycled garbage that fills so much of today's web.)

That's my real concern with ChatGPT. That it will be used to fill up ad-infested sites with crap generic content for even less money than paying pennies to a new college grad to spew out 5,000 words a day.

There's lots of new quality content being created all the time. That said, it's reasonable to ask whether there's a need for another music database, another movie database, etc.

Reply on news.ycombinator.com