How I use GenAI for my memes

I’ve been trying to do a themed funny meme a day for the past few weeks, every day when I work. I share them on my Twitter, Bluesky and LinkedIn accounts.

Here’s my process:

• Find a meme I understand. I know a lot of memes, but also check Know Your Meme to make sure I understand the context.
• Use ChatGPT to brainstorm meme ideas. This usually takes 1-2 tries, but the prompt looks something like this: “can you provide 15 funny meme ideas around saml, jwts, passwords, oidc, rbac, or other authentication or authorization related topics. The meme I am trying to use is the epic handshake meme, where the hands in the middle show agreement between two parties on something.”
• Review the 15 suggestions, modifying as needed based on my knowledge and the memes I’ve posted recently. If I just posted about SAML, I don’t want to do so again.
• Create a meme using imgflip.
• Post to the areas above, making sure to use alt text descriptions.

This process takes about 5-10 minutes, with a big chunk of the time being the actual posting. I created memes before all on my own, but it took longer, probably 20-30 minutes. The brainstorming took much longer. I’ve also done meme brainstorming in a committee, and that took person-hours, cycling through ideas.

I sometimes wonder: is this cheating? I’m still part of the creative process, but am more of an editor/refiner than the creator, though I do set the bounds by selecting the meme and creating the prompt. The quality of the memes are at least as good as what I did without any genAI help, so I don’t think this is AI slop, though of course the people looking at the memes are the ultimate arbiters of that.

Would it be cheating if I had an agent do all of this? Would it still be valuable? I also don’t know.

As the cost (in time) of content creation goes down, more and more will be created.

The human attention that I’m competing for hasn’t increased, though.