Category: LinkedIn

What do Paris Hilton and Large Language Models Have in Common?

Paris Hilton was famous for being famous, which set up a self-referential loop where visibility generated visibility, and since the metrics said she mattered, she did.

I’m starting to think AI is productive for being productive.

In The Wall Street Journal this week: 40% of executives say AI saves them more than eight hours a week, while two-thirds of their employees say it saves them two hours or less. How do the same tools in the same companies create such opposite realities?

Workday found the mechanism: time “saved” by AI gets eaten by correcting errors and reworking AI-generated content. Because executives are generating, their subordinates are fixing, and work isn’t moving faster as much as it is sliding downhill faster. Workslop, they call it.

Meanwhile, 56% of CEOs say AI has delivered no financial benefit, though the dashboards look great and everyone’s producing more of, well, something?

IDK if this is very different from philanthropy/social change, where individual interventions can look impressive but leave the underlying system untouched. More of a feeling/vibe of progress decoupled from the fact of it.

So long as we measure outputs, not outcomes, count what individuals produce, not what organisations accomplish, the gap between those is where collective capacity will go to die.

The thing is, Paris Hilton was in on the joke and built real businesses behind the performance.

---

Originally written for LinkedIn on 23 January 2026. View original →

The thing about the three-body problem is that it exists in chaotic equilibrium, and no closed-form solution ever exists. The trajectories of each body become exquisitely sensitive to initial conditions and impossible to predict over long timescales.

This is exactly the predicament of philanthropy and social change.

Call it impact, and you can measure it. Call it capacity building, and you can do a three-year grant. Call it systems change, and you get a systems map and a theory of change. None of these are wrong. But they are incomplete.

What I love about Tanya Kak’s piece is that it asks us to see differently—to see adaptation not as a catalogue of climate fixes but as the social infrastructure of bonds, knowledge, and institutions that allow communities to bend without breaking.

And once you see differently, you can’t fund or govern in the same way. It is a one way door. Walk through it.

---

Originally written for LinkedIn on 19 January 2026. View original →

Vitalik Buterin’s “Let a Thousand Societies Bloom” is animated by a real discomfort with how thin and distant many of our institutions have become. His answer is exit: pop-up cities, network states, tribes and zones, each positioned as freedom — choice as pluralism, the need for new intermediate institutions between the individual and the state.

Then I saw this from Subrata Singh where a panchayat is inviting citizens to volunteer a few hours a week, to support day-to-day work.

Both are responses to shared institutions that often feel unresponsive, impersonal, or hollowed out.

Vitalik says: when institutions don’t work, build new containers elsewhere that are more coherent, more selective and more aligned. Create spaces people can opt into, and let pluralism emerge through choice.

The panchayat says: perhaps the work is to thicken the container we already share. To rebuild capacity by presence, by showing up, staying in conversation, and offering time and attention where we are.

One treats pluralism as the freedom to choose one’s community. The other treats pluralism as something practised, relationally, with people you didn’t choose.

The truth I am drawn to is that relational capacity grows from staying, from being willing to enter imperfect institutions and make them a little more human from the inside – even/especially when the work is slow, unglamorous, and awkward.

We may well need a thousand societies to bloom. But we also need more people willing to ask: “How can I help, here?”

---

Originally written for LinkedIn on 29 December 2025. View original →

The future is unlikely to be designed and built. It will be grown.

As the year closes, this idea has kept surfacing – partly because I want it to be true, and partly because it keeps turning up in what I’m reading, in conversations and in the ether.

I had these two pieces open – one a working paper on Rewilding Civic Life (drawing lessons from the UK’s Big Local experiment) and the other an essay on systems transformation moving from “big pushes” to “a thousand small shifts”.

Big Local is a story about what happens when you protect space long enough for people to act together, get things wrong, learn, and try again. And with real power and a long horizon, and without demanding that everything justify itself on a quarterly reporting framework.

Kattel zooms out to say that this is what development looks like now because the systems that matter shift through the accumulation of many small experiments that change the relationships inside and between systems. Less architect and more gardener.

Either I’m trapped in a well-curated filter bubble, or there’s a slightly eerie coherence in what a lot of people are independently noticing. I genuinely don’t know which.

But if the future is grown rather than built, it changes what we should reward. It changes what leadership looks like. And it changes what philanthropy is for; less “funding the solution”, more “funding the conditions”.

If that’s even roughly right, what kind of leadership do we need, and what should funders stop doing so that the growing can actually happen?

PS: There’s also the work on relational state capacity, which argues that what we call “capacity” is incomplete without relationships. That a society’s ability to act depends not just on institutions or infrastructure, but also on the texture of everyday interactions between citizens and those who act in the name of the state.

And evidence from rural communities showing that trust, agency, and belonging don’t grow primarily through dialogue (even dialogue across difference) but when people do things together that matter. Action seems to create the relational surplus that trust later draws from.

---

Originally written for LinkedIn on 22 December 2025. View original →

Maybe geography is connection, and distance is abstraction.

The Internet was supposed to make geography obsolete. “Everywhere yet nowhere.” The promise was that it would cross borders, break down barriers, and destroy distance. And it did, but what we lost in the process was place.

Place is where homeostasis happens. Where the boy can work and still care for his widowed mother. Where neighbours can meet face-to-face. Where trust accumulates through repeated encounters.

The interesting pattern: both Blue Dot and Roundabout are using digital infrastructure to restore geography as a constraint. The 15 km radius, the neighbourhood and the local steward who knows the context.

They could choose to scale globally, but connection at scale without place becomes… what we have now? Feeds that feel extractive. And metrics that are hollow but claim abundance.

Is the future of healthy digital infrastructure less about conquering distance and more about serving proximity?

PS: Check out https://www.areawise.app/ from Amit Bansal Areawise

---

Originally written for LinkedIn on 2 December 2025. View original →

Yes, I often lament about abstraction winning over connection. But here’s an example of abstraction actually serving connection, and it’s worth understanding why.

In Dharwad, Karnataka, an employer who needed 20 welders was 800 metres away from an ITI full of trained welders. Neither knew the other existed. The district didn’t have an unemployment problem as much as it had an invisibility problem. All the supply and demand in the world, and nothing moving because the relational tissue was too thin.

The Blue Dot (ONEST) project uses digital infrastructure (tools of abstraction! Maps! Protocols! Verified credentials!) to make local jobs and local job-seekers visible to each other. Students see openings within 15 km of home, and employers see verified candidates across the road. Simple enough. But what makes it different from most tech “solutions” is what it chose not to do.

It removes the wrong friction, neighbours who never meet, while preserving the right friction: locality, human verification, placement officers accompanying volunteers door-to-door, nightly calls to fix what broke, trust built face-to-face.

Most platforms replace relationship while this one restores it. Most systems optimise for scale without limit, while this one has a built-in boundary radius of 15 km as a choice.

There’s a moment in the piece where a boy feels relief, realising he can work nearby and still care for his widowed mother. That’s what it looks like when technology remembers what it’s for.

Heavy social infrastructure. Light technology. A container, not a platform.

Abstraction doesn’t have to disembed, extract, and scale indefinitely. It can re-embed people into local relationships. This is rare, but it is possible. And we should study it closely.

Story in the comments by Pankaj Mishra from a couple of months ago. Gaurav Gupta EkStep Foundation

---

Originally written for LinkedIn on 1 December 2025. View original →

Two pieces, read together, make a case I want to sit with.

Damasio on what education is for:
“Hunger, thirst, satiation, pain, heat, cold, well-being, pleasure, malaise. All of these are feelings and all of these… indicate to you, in no uncertain terms, something about the state of your life right now.”
“The ultimate practical goals of education are very simple: reduce harm to the individual, enhance the individual’s welfare.”

WIRED on Alpha School:
“She told her mom that staff at the school said she didn’t earn her snacks and wouldn’t get them until she met her learning metrics.”
“She broke down and sobbed that she’d rather die than keep going.”

Education evolved to serve life. Here, life serves metrics.

---

Originally written for LinkedIn on 27 November 2025. View original →

Abstraction defeats social change in at least three ways.

1. Natasha Joshi shows the epistemological abstraction: lived understanding > quantifiable data > “impact metrics.”
The push to express care in market terms, losing what matters in the translation. We devalued what we couldn’t measure (Loss of quality).

2. Deepak Varuvel Dennison shows the knowledge abstraction:
embodied, local, oral wisdom > documented institutional knowledge > digitised corpus > AI training data.
Whole worlds of human understanding are simply absent. We digitised only what was dominant (Loss of diversity).

3. Jocelyn Skillman shows the relational abstraction:
village care > professional therapy > AI companion.
Each step trying to solve the scarcity created by the previous abstraction. We automated the empty shell that remained (Loss of humanity).

We’re at the point where LLMs trained on incomplete, abstracted knowledge are being used to deliver care that’s already been abstracted twice, measured by metrics that can’t capture what matters.

This is what late-stage abstraction looks like. We’ve repeatedly chosen abstraction over connection across every domain, building a technological future that’s systematically ignorant of most human wisdom and structurally incapable of providing what humans actually need.

The promise was abundance. Is the reality a hollow simulation?

---

Originally written for LinkedIn on 24 November 2025. View original →

Fabulous conversation with Aniket Doegar from Haqdarshak and Archana Pillai at NSRCEL this afternoon.

Aniket talked about how he began obsessed with numbers — the validation of “How many?” that all of us in India grew up believing was the definition of scale. And then, over 10–15 years, he realised something quieter and more durable: that the deepest impact he ever made was with the 95 children he taught in his first two years as a Teach For India fellow. And that the problems worth solving take a decade of listening, patience, and course-correction, not a quarterly sprint.

Scale need not be a race to big numbers and can be a long, meaningful apprenticeship to the problem. Staying curious, building a culture that can hold the weight of growth, and being willing to learn from the field rather than from fixed plans.

---

Originally written for LinkedIn on 21 November 2025. View original →

The leadership contradiction no one names clearly: we ask leaders to embody uncertainty while designing systems that punish them for it.

First: McKinsey & Company’s study of Fortune 500 CEOs. Superficially, enlightened language: “curiosity,” “learning mindsets,” leaders who “admit they don’t know everything.” But read it again. Even the celebrated vulnerability, “bring your worst self”, is nominally a strategic problem-surfacing to maintain the machine. And even McKinsey can’t quite hide that it’s breaking people.

Second: A piece by about poker players that, inadvertently, maps our entire leadership moment and points out that when the game is life and not just numbers, connection matters more than optimisation.

I return to the tension of connection versus abstraction (again and again, SORRY!) because we need frameworks to see patterns, but when abstraction becomes untethered from the actual people, places, and histories, it loses what makes systems real.

Do we have it backwards?

The wisdom isn’t in doing more, faster, with greater optimisation, but in knowing when the framework helps and when to let it go. The heroic paradigm is collapsing under its own weight. Do we have the language and practice ready for what comes next?

---

Originally written for LinkedIn on 18 November 2025. View original →