More Meme-ing with GPT4

This ad campaign is kind of cringy:

But GPT4 does a pretty good job reading straight from the image and interpreting the terms.

The image appears to be a public health campaign attempting to utilize Gen Z slang to communicate the dangers associated with street pills, presumably containing substances like fentanyl, which are known for being highly addictive and potentially lethal.

Each panel in the image takes a different approach:

– The top left panel uses “No cap, those pills are sus,” which translates to a serious statement that the pills in question are suspicious or untrustworthy. “No cap” is used to emphasize honesty, and “sus” is a shorthand for “suspicious.”

– The top middle panel’s “We’re not frontin’, this is meant to stand out” suggests that the message is sincere and is intended to be attention-grabbing. “Frontin’” means to put on a pretense or to falsely represent oneself.

– The top right panel plays on the popularity of Rihanna’s Fenty beauty brand with “The only Fenty we like is the beauty brand @rihanna,” signaling that the drug fentanyl is dangerous, unlike the admired beauty brand.

– The bottom left panel with “Keepin’ it 100, street pills kill” aims to be straightforward about the lethal risks of street drugs. “Keepin’ it 100” means to be completely honest or to keep things real.

– The middle bottom panel states, “Bruh, did you know overdoses are up 20% in CLT?” The term “Bruh” is equivalent to “dude,” and “CLT” is likely an abbreviation for a specific location, like Charlotte.

– Lastly, the bottom right panel uses “You + Street Pills = We Don’t Ship,” employing the term “ship,” which typically refers to endorsing a romantic pairing, to say that the combination of people and street pills is not supported or approved.

The campaign uses the hashtag to underscore its central message. The strategy is to frame a serious and educational message within the vernacular of younger generations, presumably to increase engagement and awareness about the issue of dangerous street drugs.

I would say this is a pretty solid answer. The fact that it can get the ocr correct is nice, but I also liked how it understood it was 6 panels and gave the answers about each panel. Really curious what sort of embedding space they are using for the images.


Posted

in

by

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *