As the models get better, I should be doing bigger tasks and not the same task but with more ease, you know? When first LLMs came around, I was using it to write stories. Then I was using it to build a SaaS. Then I was using it to make videos, and now that models have gotten better. Like they have in the previous phases, it would be highly stupid to try to continue building software and videos using it. It has leveled up, so the arena has leveled up. The bottom rungs have been occupied by others. So I shall rise up. I shall move beyond petty things like animations or software. or just writing simple scripts using AI. AI is reaching godhood.And I cannot disrespect God by making him write trivial things like this.
These days, I am using AI to play around with the transformer architecture itself. This would not have been possible two months ago, or three months ago, but now it is. So imagine how stupid it would be to still try to do what I was doing three months before using AI. I was making manim animations to explain transformers. What’s incredible waste of opportunity and efforts, right? That’s why as LLMs get better, I should keep getting more audacious. The better the AI gets, the more audacious and more impactful projects I should undertake.
So it’s not like you stay stuck at a project, and you try to harness AI, even as it keeps getting better to try to do the same task. No, rise up with AI, make things that are now possible that were not before. That’s how we move in this chain.
So next month when a even better model comes along, maybe instead of just tinkering with transformers you might make something that is completely different from scratch. That’s how we roll.
We are going from a bonfire to a gasoline store to now a fucking power plant. Why the hell are you still trying to boil potatoes using it? We can do so much more.
Keep running experiments that will shock culture to its core. That will impact a lot of people’s dreams and nightmares. Don’t play the games of olden days. Do what models make possible today. Whatever arena was unlocked, it is yours. Manifest destiny. This is the new frontier. And the gift of this life is that each month you will keep unlocking new arenas as AI keeps getting better.
Don’t fall for silly trivial things like automation and reducing your work by ninety percent and other things like that. That’s really, really pathetic. That’s lowly. That’s subhuman. You are meant for great things greater than that. Go do what is now possible that was not before. Go shock the world using that.
The objective should not be to become increasingly efficient at yesterday’s ambitions. It should be to acquire ambitions that were impossible yesterday.
And without knowing, you have been doing that your whole life. If motherfuckers are good at solving five problems out of twenty, you become good at solving six. If everyone becomes good at solving six, you become good at solving ten. You rise above. To be valuable is to be someone who can do things an average person cannot with the tools of the era. And growing ambition is the only way you can do it. In past you could just let the corporate hierarchy give this to you that hey as you get promoted each promotion will bring enough bullshit that you are always doing something that you couldn’t yesterday. But you are alone on the journey in past. Now you have AI with you. Don’t fall for it. Keep rising at the same pace as AI is doing.
| Era | Your arena |
|---|---|
| Early LLMs | Generating stories and language |
| AI coding | Building entire software products |
| AI + creative tools | Producing animations and videos |
| AI agents today | Conducting experiments on neural architectures |
| Next frontier | Discovering and engineering new computational systems |
Another thing that I want to talk about is impacting the culture. Everybody says, hey, this song is so famous, it has such an impact on culture, blah blah blah. No, motherfucker. It is culturally impactful. That’s why it is fucking famous. It has an interesting experiment underpinning it. It might be a experiment like drake where he is like hey I gotta make something that gets played a lot at the clubs and the radio or it can be like kendrick where he is trying to experiment with social issues such way that people just have to listen to it. A product can exist, and it can be such that no one knows about it. A famous product can only be such that it is culturally impactful.
Calling a famous song culturally impactful is like calling a virus contagious after it has infected a million people.
Motherfucker, it didn’t become contagious because it spread. It spread because it was contagious.