Vinod Ganesan

Sports feel like the one place where outcomes are actually earned. Clear rules, shared field, visible effort and first is first. But the race doesn't start at the starting line. It starts years earlier, in who got access to coaching, nutrition, facilities, and time. This week's Almanac is on why sporting meritocracy is a partial truth and why partial truths are the most convincing kind. 🔗Link below!

@amrnth0 Amen!

Code can now be treated as a digital living creature - growing, learning, falling, healing, winning, losing on its own, fueled by tokens.

Impressed to discover (though obvious in hindsight) that agent harnesses like OpenClaw can self-rewrite - not just its prompts but also its code in the most barebones deployment (forget secure isolations for a moment). This is pretty cool and not possible with closed harnesses like CC. It opens up interesting interactions where designing a future agentic architecture is a conversation with its current form and not with an IDE or another harness.

The cost of making things has collapsed. You can generate something polished almost instantly. And yet, a lot of it feels interchangeable. The constraint isn’t execution anymore. It’s judgment.🎯 This week’s Almanac explores why taste, the ability to decide what to keep, cut, and refine is becoming the new moat. Link below🔗

@pHequals7 @bcherny Try with Cursor and salvage yourself :)

@bcherny is this a harness or a model issue?

this has been happening way too often with opus 4.6 these days where it confidently spits out an initial version of a fix and then mid way corrects itself saying this won't work too annoying - the model has been nerfed beyond recognition

@amrnth0 @grok @grok - do explain. Or @gork you explain

@vinodgansan @grok explain what @vinodgansan meant here

Everyone deserves the quality and elegance of Mac. So happy to see Neo doing well! Windows laptops have no excuse to be shitty now.

@iruletheworldmo What wasn’t working with your openclaw memory?

finally, a working fix for my openclaw memory

Bengaluru is full of surprises. Even SF can’t keep up.

Every idle moment is now a problem to be solved, a gap to be filled before some uncomfortable feeling slips in. But the idle mind was never broken. It was processing, connecting, wandering somewhere useful. 🧬 A new Almanac newsletter on what boredom actually was, what the neuroscience says it was doing for us, and what we've sort of lost by replacing it with an infinite scroll and a dopamine drip. Link Below!🧠

@naval Was?

Software was eaten by AI.

@pHequals7 Banger!

2026: how i learnt my masala dosa depends on the strait of hormuz

AI in warfare is often discussed as a future problem. In practice it has been arriving piece by piece. First as tools to analyze intelligence, then systems that recommend targets, and now software embedded directly in the pace of military operations. ⚔️ The recent Iran strikes and the Pentagon’s fallout with Anthropic revealed how central these systems have already become. This week’s Almanac looks at how AI moved from experiment to infrastructure on the battlefield. 🔗Link below!

Adding sources is fun in Almanac! No more starting from scratch. Try it out in beta ✨ try.almanac.so

Alan Kay, in his 1989's iconic "User Interface: A Personal View" makes this prescient statement, although in the old-age context of a PC. However, how so applicable today, more than ever!

@commandpaul @iScienceLuvr @thinkwithalma Thanks @commandpaul for the shoutout. Would love for you to try @iScienceLuvr

@iScienceLuvr Checkout @thinkwithalma , something my friend @vinodgansan is building :)

i think i should get back into writing threads/articles over here... i have a few ideas about different medical AI topics to work on, stay tuned...

Introducing Perplexity Computer. Computer unifies every current AI capability into one system. It can research, design, code, deploy, and manage any project end-to-end.

Singular. Steve.