What ChatGPT is good for here
Explaining the trend and the song, picking which friend takes the hook and which takes the verse, writing a prompt for a video model, and writing the caption and hashtags for your post. It is also handy for checking which photo of you is the clearest.
When you ask it for a prompt, tell it that 58th Floor is the Anderson .Paak and Cordae COLORS performance in a teal studio with one hanging mic. Otherwise it may describe a high-rise apartment.
What it does not do
It does not take two photos and return the finished clip: the teal studio, the hanging mic, the four shots, one person finishing the hook and the other taking the verse, both lip-synced to the original song. That needs a video model guided by a motion reference, which is what the template sites run.
It also cannot add the original recording to a video for you. Template sites that keep the song mux it onto the finished file.
The two-photo route
On 58th Floor AI: add a photo of each person (or pet), pick Standard or Pro and the quality, check the price on the button and render. In a few minutes you download a 15-second vertical MP4 with the original sound, or a silent one.
Then come back to ChatGPT for the caption if you like.