BuzzFeed to Use ChatGPT Creator OpenAI to Help Create Some of Its Content | Hacker News Reader

BuzzFeed to Use ChatGPT Creator OpenAI to Help Create Some of Its Content

wsj.com

22 pointsjackallis15 comments

This is really the dangerous part of where we're going with this. Our culture will end up with a net increase of Buzzfeed type content and decrease of substantive content.

Substantive content is going to become the new vinyl. Sure, it will still exist. As a niche. You won't be able to find much of it.

It's like the indignity of Google replacing customer service with algorithms knowing that it's less effective, but not caring because it's cheaper and it scales. But now it's about to be applied to our entire society.

For every artist using generative AI to find a new way to be creative there will be 50 companies using it to give us a worse experience that saves them money, and with it becoming the new norm they don't have to worry about losing business from it.

I know someone who has an SEO business, and AI-generated is pretty much all he's doing. He has hundreds of websites with nothing but AI-generated content and all the SEO shit and they're all top of the pop at Google.

Ever since I know him I believe more and more that many websites I find are AI, because they all look exactly the same and have the same formatting for their content: articles are short enough to be read in a few minutes but not short; articles are split into 4 or 5 subheaded sections; at least one of the subheadings begins with "which", "how" or "why": last section is often a conclusion or summary; text reads like an idiot wrote it; text absolutely doesn't give you the information you were looking for.

OpenAI, DALL-E, Stable Diffusion and all the others are already being exploited to "improve" these websites. The pictures, the videos, the text, it's all fake.

> articles are split into 4 or 5 subheaded sections; at least one of the subheadings begins with "which", "how" or "why":

This is interesting. I always thought it was structured this way to get into the Google Answerbox. I hated this trend because this type of repetitive articles defeated the whole purpose of hypertext in the first place.

I wish places like MSN and Yahoo, etc would employ AI instead of college kids etc., because the present quality of those "articles" is abysmal and worse than AI generated content.

While this may be a valid sentiment, it doesn't change the fact that you're reading bullshit. Many of the articles created by the aforementioned system are of a medical nature, for example, and the information it spits out is often patently false.

The thing is AI generated content, much like illiterate college kids masquerading as journalists, are a symptom of a media that is fueled only by money and corporate interests.

Is he rich off of this? I was under the impression that there are already tons of sites doing this, and they’re watering down the product and revenue.

We already have enough Buzzfeed listicles. Now these will be produced for close to free instead of by humans. I'm hoping it diverts humans from working on this low societal value content and into producing new knowledge.

What do you think the odds are that AI creating an infinite amount of low-quality, low-value content will change the world in the way you hope?

There is only so much demand for this content, and all of it is already being filled. I will not suddenly want to consume this content just because it is cheaper to produce.

it will probably divert humans from writing mostly junk and occasional quality content to far fewer humans writing quality content while every other writer is out of a job.

maybe we need this divergence - high quality content should be compensated accordingly and will be paid well - and there will be competition from writers for this small pool of high paid jobs.

Just look at substack paid newsletters to get a glimpse of the future of writing.

free content however - will become ChatGPT's home base

If anything, this will likely improve the quality of their writing. I'm sorry, that's too mean. CNET recently ran into issues with their usage of GPT https://futurism.com/cnet-ai-plagiarism so I wonder how this will play out.

I was waiting for it eagerly and here we are. Dont think much will change for buzzfeed content-wise, it was always pretty indistinguishable from GPT.

Indeed, I thought they were going to admit to using GPT from the very beginning.

Yes, this does not include their generally high quality investigatory pieces.

ChatGPT is a boon to click-bait content generators like BuzzFeed. There's a ton of low-quality media content out there that's about to become fully automated. A lot of it is of course, like sports articles, and that content is probably going to improve with ChatGPT and its successors.

Reply on news.ycombinator.com