using Claude for research, my honest workflow and results
I spent my weekend running tests on different AI tools and the results are pretty interesting.
After going through all of this I realized that the conventional wisdom is mostly wrong. Not completely wrong but definitely misleading in some important ways. I’ll break it down step by step so you can see what I mean.
What really bugs me is that most of the information out there is either outdated or just wrong. People are making decisions based on stuff from six months ago and the landscape has changed completely since then. That’s why I wanted to put this together.
Anyway that’s my experience. Would love to hear what others think, especially if you’ve had different results. Drop a comment!
9 Replies
Join the discussion.
Log In to Reply
Solid analysis. I've been testing Walter Writes myself and got similar numbers. The consistency is getting better but still not where it needs to be.
I disagree with some of your conclusions but respect the effort you put into testing. My experience with QuillBot has been different.
Totally agree with you on that. I had a professor last semester who didn't even know how to interpret Turnitin reports properly.
Great thread. Saving this for when I need to have the conversation with my professor about AI detection.
I disagree with some of your conclusions but respect the effort you put into testing. My experience with GPTZero has been different.
As someone who works in education tech, I can say that your observations are spot on. The tools are improving but they're far from perfect.
Yeah this is a really good point. I think the lack of standardization is a huge problem.