TL;DR
We have all seen what ChatGPT can do (it’s not new to anybody).
So many articles are being written over and over using ChatGPT.
...
For further actions, you may consider blocking this person and/or reporting abuse
Well, but we still don't talk about the philosophical catch behind all this: why read a blog post generated by a bot when you have access to the same bot, and you can ask the same bot yourself more relevant questions?
Would you rather read a book about your issue written by your colleague sitting next to you, or would you rather ask them personally?
By the time we can effectively generate blog posts, books and emails, they would be all obsolete. Zero value.
The moment I see an article written by a blog, or I suspect it, I rather close it.
How did you arrive at this figure?
You’re being very polite in your question @robole
When I read this phrase I stopped immediately to look in the comments to check if someone found that strange.
This affirmation - without any source - is a bit offensive. This community has years of existence, it’s here much before than ChatGPT, and it’s made by all kind of people.
Also, let’s suppose this is true, then that’s actually an issue, in IMHO, that would need to be solved.
(I support all our fantastic writers that grow and share their talent in this community 🙇🏻♂️)
“The calculator is a tool for someone who already knows basic arithmetic,” - Suydam
In 1980s, math teachers protested against using calculators. Today, Large Language Models (LLMs) are like super advanced calculators and much more.
If someone who understands their topic uses LLMs to shape their message, I think it's fine.
But when used by someone clueless about the subject, I see problems. LLMs will change the way we write, whether we like it or not. They'll become so smart that it'll be hard to tell their work from human writing.
BTW: 1/3 paragraphs above is written by GPT 4. Can you guess?
This is interesting! Training the agent on your own documentation. How well does it work in your experience?
I'm going to save this tutorial and implement this into a test project. This is awesome!
Nice one! Did you test it out? Is it good enough that you would use it for your blog posts or more as an aid/ tool?
Great article Eric. I wonder when Dev.to is going to start punishing pieces of content which have more than X% that comes from GPT. Anybody with this insight, I'd love to hear your thoughts!
The problem is the huge amount of false positives most of the detections produce.
I was wondering if we can connect generating posts to a self hosted WordPress blog.