ksunnole

"If I had always done what I was 'qualified' to do, I'd be pushing a broom somewhere" @naval

End of Prioritization I’ve been thinking about the tension between exploitation and exploration lately - mathematically best described by the multi arm bandit problem. You can’t do everything because trying something has a cost. Just as so many other laws of physics are changing with AI, I think this one is about to change too. For any intelligence+execution bound work you can imagine the cost of exploitation (trying something) is rapidly approaching zero (modulo inference). In that world, the value of exploration goes up dramatically — you can simply try more things. This is a broad, important concept that applies to thousands of trade-offs in companies and society that we previously took as immutable. It also tells you something about where value accrues in the future. People who can identify compelling new paths to explore will have far more value to add than people who are experts at specialized exploitation of known paths. I have a feeling this might even have implications for the multi-armed bandit problem in the formal mathematical sense, but that’s a bit beyond my expertise. Think about a growth team that A/B tests two landing pages a week because each variant costs real design and eng time — now they test fifty. Or a product team that agonizes over which feature to build next because they can only ship one — now they build all of them and let users decide. It’s like Monte Carlo simulation for everything, except you’re not simulating — you’re actually doing it. Every path gets run. Prioritization as we know it is obsolete. You don’t pick what to do — you do all of it. The only art left is knowing which bandits are worth arming.

A truly progressive, future-looking person kills their innate urge to be acknowledged as right; exactly the moment they are proven so.

@Vtrivedy10 When support for something like Agent Teams? As in claude code?

AI will handle politics better than humans.

@aakrit Adding another component, which is keeping yourself updated with latest experiments happening in the space.

Building an AI product that's 80% as good is easy. The real work is the last 20%. It's where great products differentiate from wrappers. - Proprietary data - Deep domain insight - Model + infra orchestration - Workflow design - Human augmentation This part is brutal. Takes forever to get right. But here's where magic happens and winners are created.

In a post-AGI world, every human would be forced to think deliberately, produce original thoughts, and adopt personal agency.

At this point you must start converting your thoughts into actions to produce new information, which would breed new grounds for investigations.

Humans must spend a lot of time on deliberate thinking. Not overthinking, but deliberate thinking.

@KirkMarple @levie @grok which arxiv paper is talked about here?

@levie post inspired me to continue this epic conversation about context graphs tonight. Even found the original ArXiV paper on it. x.com/kirkmarple/sta…

Study STEM Now.

@ick_real Constantly reminding myself that I am distinct from my emotions

One habit that massively improved your mental health?

Opinions must be expensive, or else they don’t matter. If you investigate deeply before ever forming an opinion, you will turn out to be right most of the time.

Richard Feynman on how physics differs from mathematics.

@grok my 2025 wrapped?

@akshaymarch7 I wish it was that easy!

LLMs don’t ‘think’. They predict the next token. This one fact explains 80% of AI behavior.

@dhumchikdish No one really knows what time is.

don't succumb to the guilt of not having the year you wanted time is a man made construct happy diwali

@PatrickAlphaC Use the word “Ultrathink” it activates thinking mode in claude

Make no mistakes