Ever have one of those moments where you try to hold a company to its own promises, and they respond by showing you the digital door? That—s the spicy saga of one Perplexity Pro user who dared to cite the manual.

This user noticed the much-hyped —Deep Research— feature wasn—t digging very deep. So, they did what any good detective would: they used Perplexity—s own official documentation and launch blog as evidence to highlight the downgrade. The post blew up, getting hundreds of upvotes and comments from fellow users who agreed. The community verdict was in. Perplexity—s response? A permanent ban and a deleted thread. Talk about shooting the messenger.

It—s the ultimate —look what you made me do— scenario. You can—t help but laugh at the irony of using a company—s words to critique them, only to get banned for it. It—s like getting kicked out of a book club for quoting the book too accurately. The moderation strategy here seems to be less —address the feedback— and more —delete the feedback and hope nobody saw it.—

There—s also a funny lesson about the hierarchy of truth in tech communities. User screenshots and testing? Debatable. The company—s own published words? Apparently, that—s the most dangerous evidence of all. It turns out the quickest way to find a limit is to read the terms of service back to the people who wrote them.

In the end, this is a peak internet culture moment. It has everything: a promised feature that under-delivers, a community rallying together, and a baffling overreaction that guarantees the story spreads far beyond one subreddit. The takeaway is clear: if you—re going to advertise deep research, you better be prepared for your customers to do some on your claims. Otherwise, you might just end up researching the depth of your own backlash.