the gradient-descented normie
Do you not want to marry someone who doesn’t know what 6-7 is?
mimetic desire
Like my most unhinged post I ever made, it all started because of a vimothee chalamet video.
He used the term mimetic desire (the idea that humans desire things not because they intrinsically desire them but they see others desiring them) and something kinda clicked for me the second I saw the term which led to me frantically writing this very blogpost.
The idea that exists in my mind may not map 1:1 with the actual concept as popularely conceived of, but this is mine.
To illustrate what I mean, we can use the very simple example of being “pro vs anti AI” on a y-axis and “politically left vs politically right” on an x-axis. (It’s obvious that these cluster somewhat.)
The actual graph might look something like this

Now my point is that because social media algorithms, and how human beings communicate, if you exist outside these 2 clusters, you will day-by-day shift beliefs (since you are what you eat when it comes to media) and kind of gradient descent into one of these clusters.
Now expand this into like, a couple thousand dimensions, that’s what I’m getting at. It’s really hard to intuitively explain but it is something I noticed.
This is how we get normies and TPOT and incels and all the other things.
uniqueness
When I go looking in my soul and think about who I want to be, who I think are the coolest people, I conclude that, in a lot of ways, derivation and not being unique is one of the greatest sins. I don’t think I am entirely alone in this.
Think about the most interesting people, the most ambitious, the most fascinating.
They didn’t have a typical gradient descent did they? They probably don’t exist outside of their very narrow network of mentors and peers.
And this is where the concept of “strong mentorship” comes in again, the one thing that people have repeatedly explained to me that is like the most important thing when selecting a program or a PhD etc etc... The best researchers eating dinner together, The young hegelians!
So if you accept the thesis of “you are what you eat” and that social media algos cause polarization by reinforcing/radicalizing existing beliefs. Then to a certain point you have to accept this thesis as well.
A friend recently told me that he has a new heuristic for how cool someone is: “how long does it take you to explain who you are, the longer the better”. I feel this way as well.
Ofcourse I cannot you know, *explain* why I think being unique is an aspiration for me, it feels visceral and almost axiomatic. I could think of a couple reasons why it could be good for society, I guess?: Like people in these clusters would be able to conceive of unique ideas that haven’t been tried before, while the cluster probably generates a lot of sames-y ideas.
If you care about being unique that’s up to you, but I think I do.
So idk, follow my advice maybe: go and be the suburban wine-mom that somehow uses LessWrong. Or go and be the Kantian who knows a lot about Peter Singer.
Or be anything but college student #29573 that sends instagram reels to his friends and gets drunk on the weekends and doesn’t have a single unique thought.
Do you not want to marry someone who doesn’t know what 6-7 is? Do you not want friends who *feel* like high-entropy individuals? A high perplexity wife? Am I alone in this!?


The "High entropy individuals" made me laugh. My take on authenticity is that some people cant help themselves, they have to be authentic, and some people try very hard to hide themselves out of fear of rejection - but that's the biggest fallacy of all. Re: societal value: In my experience original thoughts are what makes an individual/company. Read zero to one by Peter Theil. I think authenticity is what people call "research taste" on Lesswrong. Favorite movie on the subject is “A star is born".
Hey, great read as always. The gradient descent analogy for how our beliefs shift is super insightful, especially when thinking about the AI debate. It really makes you wonder how one can cultivat genuine, independent thought outside these algorithmic pressures. You've hit on something fundamental here.