AswinPratap

Is it just me or is Claude eating up 50% of weekly limits for even the most basic of tasks? #conspiracy

WORLD MODELS are all the rage, as the AI community tries to pivot from the perceived shortcomings of large language models to AIs that use internal “world” models of the environment in which they act. Such AIs, instead of predicting the next token, will predict the next set of states of the environment+agent, conditioned on some action that the agent may or may not take. World modeling AIs promise much—but this is not a new concept by any means. Psychologists, cognitive scientists and now computational neuroscientists have known for a while (the history goes back 150 years) that our brains must be modeling the world and using these models to hypothesize the external causes of sensory inputs. These hypotheses are our perceptions. In my first post in a series exploring world models for the WHERE MACHINES THINK Substack, I discuss the neuroscientific rationale, and put the current interest in historical context. wheremachinesthink.substack.com/p/the-case-for…

@shuklashobhit x.com/i/status/20198… one possible explanation

What 🤔

If there's one true egalitarian space, it's Bengaluru's Udupi restaurants. Why do we not have such places elsewhere? Clean, tasty, no nonsense affordable food for all. @peakbengaluru

Can sympathize with him on one thing. Keep visiting the customer and make a passionate pitch, even when you know that there is a nil chance of conversion.

#TheHedgehogStrategy The Greek poet Archilochus wrote: “The fox knows many things, but the hedgehog knows one big thing.” When attacked, a hedgehog does exactly one thing - it curls into a tight, spiky ball. No tricks. No improvisation. Just consistency. The fox, clever and adaptable, tries everything. And still loses. Jim Collins later popularized this idea in business: long-term winners aren’t those doing many things well, but those doing one thing exceptionally well. Watching the current AI gold rush, it feels like most companies are choosing to be foxes. Everyone is chasing the newest model, the latest agent framework, the buzziest feature. Strategy keeps changing. Focus keeps shifting. Effort is spread thin. You can’t blame them as well — every day brings a new flood of ideas, innovations, and hype in the AI world. But the teams making real progress in Enterprise AI are doing the opposite. They’ve made a conscious decision to narrow their scope and go deep. They are focusing on the Last Mile execution. At @EvamLabs, we’ve chosen to be a hedgehog. Our focus isn’t on models or hype cycles - it’s helping our customers on the Last Mile execution. That moment where AI actually meets real workflows, real data constraints, and real humans. That’s our “spiky ball”: E.D.G.E. execution. The market can throw new models, new narratives, or new competitors at us. But if you keep executing the last mile better than anyone else, you don’t need to chase every trend. In a noisy market, clarity and not complexity is the real advantage. 🦔 @michellemzhou @sachinjose @caesarpratap @SauravBharadwj @shuklashobhit

Brewing a perfect V60 suddenly makes the world a little less chaotic and slightly more fixable

What's the point of having so many parks if they are closed even during the day over weekends? @BCCCofficial @DKShivakumar

@sundarpichai Truly a @rajinikanth movie style turnover. Unbelievable!

iykyk

@deepigoyal What does a hypothesis even mean?🤣 It’s fascinating how the word has become the adult version of ‘I just made this up but hear me out’ - except now with extra confidence and zero data.

Just came up with a more visual explanation of how gravity could interact with human aging across the gravitational spectrum. (Again, this is part of the hypothesis, not proven yet).

We have been having mini-celebrations over the past months whenever a major customer agrees to do even a trial run with us. Not all of them are wins, but what the heck, we should keep finding reasons to celebrate. Big customers take time. But when they convert, they can propel a company into an entirely different orbit. The Databricks-Microsoft partnership is a perfect example of how transformative deals rarely happen overnight. That collaboration faced multiple setbacks at different organizational levels, despite having support from senior leadership including people at the top. But when it did happen, it catapulted Databricks into a whole new level. There are countless other examples like this: Salesforce - Apple, Spotify - Facebook, and many more. All of them died multiple times before becoming a success. Over the past months (and it’s early days for me on GTM at this scale), I have picked few nuggets which have been effective in winning customers: #1 Deliver real impact: Your product has to create clear, unmistakable value. If the impact isn’t meaningful, interest fades quickly - no matter how innovative your tech is. #2 Patience is a virtue: Enterprise sales and GTM demand long-term commitment. You often have to persuade a broad range of stakeholders, each with unique priorities and concerns. #3 Expect setbacks: Deals stall. Champions change roles. Priorities shift. The teams that win learn, adapt, and come back stronger. #4 Over engage: Credibility and trust-building across various customer touch-points and functions increase your chances of winning and sustaining partnerships. #5 Big wins take big effort: When such deals finally close, they unlock new channels, integrations, and possibilities that justify the effort, often resetting benchmarks in the industry. And above all not giving up on customers is critical. @shuklashobhit @michellemzhou @sachinjose @sameer_pendse @caesarpratap @manojumapathy1

Humans in the Loop is a powerful reminder that AI isn’t just code, it’s people. An essential viewing about a first-gen Adivasi data annotator. The movie was a powerful reminder of the hidden human workforce shaping our AI future - with all its biases, empathy and possibility.

@shuklashobhit Questioning the status quo (from past, present ) unapologetically - it’s the first step. Pride with perspective > pride without purpose

What does it take for a nation to take pride in its past and be optimistic about the future? (1) Understand the past beyond what is in the history books. (2) Believe the current/future generations have it in them to change reality - for the better.

@nasqret If discovery in mathematics is as much about choosing what questions matter as it is about solving them, how can a system trained only on existing questions ever transcend the boundaries of its data?

I encourage you to read this article, in which we describe the current situation and the directions in which, in our view, mathematics is heading. Many thanks to Ken Ono for including me in this extraordinary project. I look forward to a wide-ranging discussion and will be sharing more about activities related to this article. nature.com/articles/s4156…

#lifeupdates Purchased #spotify premium.

@UiSavior Ha ha. But I still don't get the notifications when they matter.

"Design is easy" This is how Slack decides when to send notifications.

At least it is authentic and authenticity compounds.