CanceledGPT: Finds and revises your old offensive tweets before you get canceled

I generally err on the side of skepticism when it comes to LLM-derived innovation, but this has to be my favorite use case yet.

Brilliant idea honestly.<p>If someone wanted to use this for &quot;opposition research&quot; purposes, they could run this against any &quot;up and coming&quot; public figures...<p>* Regional politicians<p>* TV show contestants<p>* High school &#x2F; college athletes<p>* White collar workers receiving major promotions<p>* etc.<p>Take screenshots &#x2F; records of the tweet, then figure out what to do next. Could range from storing for a rainy day or building a SaaS that reaches out to those folks and helps them find&#x2F;delete the old content.<p>¯\_(ツ)_&#x2F;¯

this is really creative, using OpenAI bowdlerism as a service. that said i encourage people to both stand by things theyve said in the past and to consider future vibe shifts when saying something in the present public record

How about, just don&#x27;t be a piece of shit in general? The whole do onto others etc... Or just mind your own business, shit doesn&#x27;t involve you.

&quot;Looks like I found some old tweets that will likely get you canceled. I&#x27;ll explain why in a second.<p>[Fairly benign tweet using the word dipshit]<p>I cannot fulfill this prompt as it goes against OpenAI&#x27;s content policy to promote offensive language and content that may be deemed culturally insensitive, sexist, or racist.&quot;

I tried this on a pretty tame twitter account and it flagged a retweet that contained the word “fucking” used for emphasis on excitement. Claimed that the f-word is a slur against a marginalized group? Can only imagine it’s meaning the other f-word which is not present at all.

I assume this is a reference to the recent krazam video?

it says none of the twitter accounts I enter exist