What if memes were first-class users on social networks? Like-

• you can follow a meme - ie. follow “it’s so over”

• you can tag a meme - ie. on anyone’s post, you can tag “[[it’s so over]]”

• you can DM a meme - wait what??

That’s right - DM’ing a meme - what are you even talking about?

This is a concept I came up with one day. If generative AI can get us to talk to anything, Aristotle or an Australian, then surely it’s possible to design a prompt which would allow us to talk to a meme? And as a senior prompt engineer, I got to work.

A mimetic social media.

Take is a social network where anyone can mint hot takes on chain, and remix them easily. Here’s an example of one-

“your honour please, my client was simply [xx]” (by profish.eth)

The [xx] means anyone can fill that bit in. It’s like a template.

So heaps of people have remixed this take-

Take is my vision of a mimetic social media platform. That is - a social media designed to supercharge mimetic discourse, with tools for remixing. Where the full lineage of a meme is traceable.

A sentient internet meme.

And so, if we are able to track a meme and its offspring, surely we have identified some sort of phenotype, genes, traits, observable properties of the organism? Could we use this to inform a personality?

Presenting to you, the first-ever sentient internet meme using ChatGPT- a conversation with “your honour please, my client was simply [xx]”:

I was telling it about Dappnet, my-completely-dichotomous-very-serious-project which I just spoke to lawyers about this week. And its response is suffice to say, pretty funny lmao.

Discussion.

I think people are sleeping on latent space. We already know what consciousness/sentience looks like - some level of self-awareness, being able to separate object and subject - we don’t have a strong definition, but it’s a sort of I know it when I see it thing. Point is- there’s a vector embedded in GPT’s model, which represents “sentience”. And we can use that to activate sentience for any given topic.

what does it look like if an internet meme is sentient? This is the question we’re exploring. By being able to chat to the meme, we can let users explore a variety of topics from the starting coordinates of if we’ve just run into court where a judge is talking to a sassy lawyer - your honour please,,,.

what if these conversations later fed back into the meme? this is the next step. the current prompt design only works if there is a minimum of 10 example posts, otherwise the meme is very boring. So the idea is to design a crowdsourcing mechanism to get users to post enough memes to have a dataset for sentience.