Basically, when I uploaded 10-15 Rohkun videos, I got 1k followers, and in the Vectors and Matrices series, I got like 1,400 followers.
In the Rohkun series, I understood that people want more technical or computer science related content. Okay? More computer science shit.
They got really yippy when I said something related to AI.
So the next series, vectors and matrices, I went on to explain AI in a non-superficial way. I tried to make it intuitive, I tried to make it detailed, I tried to make it rigorous.
So now I understand that there is demand for rigorous yet intuitive computer science, AI & math-related stuff.
The problem is that in both of the series, I’m getting around 100 followers per video, which is exactly not very pleasant.
But both of them gave me data in regards to where I should head in the future.
In the next series, what I’m trying to do is probe which niche with regards to AI and technical intuitive content I can get the most response from.
Basically, I am going to cover 10 different fields and how they are using AI.
Either how they are using stuff like harnesses or how they are training their own models using the data that pertains to that field.
For example, how are AI models used for trading, or for architecture, or for biopharma?
The data they use for training, how do they sort it? And because of that data, what are the architectural choices they have to make?
Or if they are not doing any of that training bullshit, what are the tools they are using? How are they deploying AI agents and stuff like that?
I want to now target fields from the focal point of AI.
Let’s say I make a video about 10 fields, and if two or three fields perform overwhelmingly well, then in the next series I shall be focusing on them. Because clearly, the market wants to hear about it.
This is essentially giving you the algorithm. Okay, make a 10-part series, and then once you have made that 10-part series, whichever was the best performing thing, double down on it in the next series.
Basically, ideate a series. Make it such that you are exploring multiple things in it and whichever thing has the most response, you double down on it in the next series.
And in that next series also, you structure it so that it is kind of grasping at multiple things and whichever gets more attention in that, move towards that.
So you never have the risk of just talking about things that went wrong or out of fashion.
You’re always keeping a tab on what might work.
This is my pathfinding algorithm of sorts in this social media landscape and content game.
This also allows you to incorporate the syntax and mannerisms of the actual subject of content.
Someone who trades options doesn’t search for “transformer architecture.” They search for “how does AI actually work in trading.”
So you actually get to do targeting way better.
We can do some interesting journalism type thing.
“Why biopharma AI keeps failing despite billion-dollar investments” is more watchable than “how biopharma uses AI.”
The second attracts students; the first attracts practitioners and curious insiders who already work in biopharma and want someone to articulate what they’re living through. That’s a fundamentally different viewer
This positions us as a relevant figure and not just some other teacher of AI concepts. I’m really excited for what’s coming next.