Koushik Busim

A dangerous cycle in AI Humans develop it, show it to other humans. They feel awe and scared, if it replaces them? Then they point out what it can't do to feel safe Then those points become new design brief for AI evolution Humans are giving directions to their replacement

@oprydai Warehouse & robotics? Or Women & robotics?😂

man only want two things

AI doesn't fix your habits, it amplifies them. What ever they are good, bad or ugly.

#DecodingGenZwithKB Introduction

Printed the first draft of my Quanta Currency Framework — Edition 01 (2026). Zine artifact: Unknown Currencies Early map of a system I’ve been building. More soon.

Consistency is not one size fits all. It never was. Consistency is contextual. Stop measuring everyone with the same lens.

People called me inconsistent for a very long time What both me and them never understood was my cycles are longer. I never really abandoned anything. I wrote my first book over a course of 7yrs. It's still not out. I have 4 more in drafting phase.

Consistency is contextual.

Setup and started writing on substack. Twitter -> short form Substack -> long form My first article: open.substack.com/pub/kuchbhiiun…

@oprydai Full vertical integration!

most people try to build robots few build the infrastructure that makes robots inevitable that’s the leverage point a real stack looks like this: → data layer sensors, logging, real-world datasets not scraped data — physical interaction data → simulation environments to test before breaking hardware close the gap between sim and reality → compute training + edge inference models that actually run on constrained hardware → control classical + learned systems real-time, not theoretical → hardware modular, swappable, designed for iteration not one-off builds → manufacturing repeatability, tolerances, supply chains can you build 1 → can you build 1000 → feedback loops every deployed system feeds data back into training AI changes one thing: you can compress iteration cycles but only if the stack is connected end-to-end most people build demos the ones who win build systems that keep improving without them

I will give my own example as a writer My future self would be thankful if i publish my work, because they will be timestamped & attributable. My past self regrets not publishing anything online because it made it invisible = non-existent Present = post in fragments, tweets

The crazy part? This framework applies to startup founders too. Eerily precise. Every decision compounds. Product. Hiring. Culture. Tech debt. Distribution. You’re either compounding leverage… or compounding problems. There’s no neutral.

I realised the most powerful thing a human can do is ask these 2 questions 1. What will my future self thank me for that i should be doing now? 2. What is my past self teaching me now that i should avoid doing now? Present self picking from the overlap(common) of both