The Self Is The Platform

On Consumer AI, and why individual leverage is eating collective coordination.

Sep 18, 2025

Two parallel shifts are occurring simultaneously as AI reshapes consumer technology: individuals rivaling groups, and private rivaling public. One person can now outperform the collective (scaling personal capability beyond team coordination). And private AI interactions are overshadowing public platforms (intimate 1:1 or 1:few beats broadcast 1:many). These intersecting forces are fundamentally inverting the mechanics of scale.

Technology has always been about leverage — extending what individuals can do alone and amplifying what groups can do together. And for the past two decades, the big technology story was collective scaling: coordination via networks.

The most valuable startups of the era built platforms that connected millions of people: Facebook, Instagram, Airbnb, Uber, Slack. Marketplaces went from zero to scale and startup to IPO. Personal software made people incrementally better, but to go further, you needed more people: assistants, meetings, and platforms to coordinate with even more people.

The equation was simple: more people meant more value. Power accrued to those who organized networks of people, aggregated their assets, and coordinated behavior at scale.

But AI has inverted that equation. Now, the individual can rival the collective. The single-player ceiling has shattered. Scaling yourself is now the most accessible, fastest-growing source of leverage. This is the great leverage inversion.

The individual is eating the collective.

If the last era was about platforms organizing people at scale, this one is about tools scaling the individual. Personal leverage can now beat social leverage.

Each of us becomes a Platform of One: a central node surrounded by an ecosystem of enhancements and extensions that multiply what we can learn, build, and do — across work, play, and self-optimization. The self is the platform.

The most popular AI-native tools like ChatGPT, Cursor, and Midjourney expand individual capacity exponentially. This extends to hardware like Meta's new Ray-Ban glasses for hands-free AI assistance. They’re all what I see as supertools (in many respects a new breed of prosumer tool). These products still benefit from network effects — but not the coordination type that dominated the last era.

Instead, they thrive on data network effects (collective usage creates better models that benefit everyone) and personalization effects (the tool becomes more valuable to you specifically by learning your patterns and workflows). These AI-enabled amplification effects feel unbounded. (They can still accrue brand and social effects without leading with user-to-user coordination).

The rise of software creators, spurred on by the unrelenting wave of vibe-coding platforms, adds another dimension: now anyone can build simple software for themselves. The likes of Lovable and Bolt and more show how solo supertools now beget more toolmakers, self-reinforcing the cycle of capability expansion.

Consumer AI is personal AI.

Consumer AI is foremost personal AI, and private or semi-private by nature.

Everyone’s awaiting the AI social network, but the reality is that, right now, the most novelty and the most leverage is coming from new tools built for one. Social will come, but not as we know it. It'll have to be unique enough to beat the pull of existing social networks, the retreat to private spaces, and the allure of individual enhancement. (Most of the attempts so far have been derivative or gimmicky).

Even social needs are being supplemented with private, 1:1 AI relationships. Solo companion experiences via the likes of Character.ai and the embodied Friend necklace offer connection without coordination.

Yet there's a poetic paradox: the AI supertools we use alone and in private are the same ones reshaping culture so publicly. Solo is eating social.1

The fastest way to change how a group behaves is to give each person private superpowers. Walk into any office or library or even coffee shops now and you see it: people deep in one-on-one conversations with AI — asking for opinions, searching down rabbitholes, writing, designing, coding. That’s just the passive seep into culture. The active comes when people share results, threads, even entire conversations.

We act in private, then amplify in public.

Borrowed networks, borrowed interfaces.

Single-player tools are private by nature, which makes product-led growth slow. And in past eras, tech stabilized relatively fast so teams could shift focus. But AI breaks the rhythm. Every few weeks we get model upgrades that force a rebuild.

It’s a brutal tradeoff: chase the models or build the moat. It’s incredibly hard to do both. Building a startup has always been described as “flying the plane while building it.” In the AI era, the laws of aerodynamics keep changing mid-flight too.

One solution: borrow existing infrastructure when you can’t do it all from scratch.

We're building new supertools but relying on old networks to spread them. Twitter, YouTube, LinkedIn, Instagram, Substack: these networks act like a Distribution Commons — like cultural utilities carrying new products and behaviors into the mainstream. Reviews, screenshots, demos, and launch videos spread faster than the product alone. This is how solo tools still scale virally.

We're embedding new tools in familiar interfaces — email, texting, calendars, browsers. Extensions and integrations inherit years of user habits, making adoption effortless. We see this power in tools like Poke, a viral AI assistant that lives in iMessage. The next wave will AI-enhance the apps we already use daily — photos, notes, weather, health — the full breadth of our personal utility stacks.

Of course both strategies create platform dependencies. Owners can bleed you out, cut you off, clone you, or if they’re really impressed — buy you. But embed well enough, long enough and users get attached to how you improve their lives. And that loyalty can survive platform lockouts, migrations, or even full rebuilds.

Long live the solo supertool era.

Individual capability ceilings keep rising with every AI foundational model release. So the 'maxed out individual' state seems perpetually deferred. And scaling solo leverage is too vast and valuable a market to move past quickly.

This is the solo supertool era: scale the individual first, then build the networks.

This transformation will be economic and cultural as much as technological. As individuals gain leverage, headcount becomes drag and coordination becomes cost. Consulting, middle management, and organization tools face compression. Industries built on individual capability enhancement will expand — across productivity, creativity, education, finance, fitness, health, and everyday tasks.

Importantly, I'm not writing off the emergence of new interfaces, multiplayer use cases, and novel coordination networks — at all. They’ll emerge as they always do. But I'm willing to bet that this individual enhancement phase will be longer, deeper, and more transformative than past cycles, reshaping tech and culture. And I expect supertools that win this era to become the center of these networks.

When one becomes many.

At the logical extreme of the great leverage inversion, the boundary between individual and collective erodes. The self becomes a distributed system: one consciousness orchestrating many capabilities. The individual rivals the collective by becoming the collective. The self is the platform.

Perhaps the distinction becomes centralized agency versus distributed agency.

The individual can now scale internally before needing to scale externally. Why hire a team when you can deploy AI versions of yourself? Why build organizational systems when your personal AI toolkit executes at enterprise scale? In this world, teams become collaborations of many platforms of one — orchestrations of distributed selves tackling complexity beyond what any single enhanced individual can (or simply when humans choose the spiritual rewards of shared space and creation).

The individual is now the base unit of scale — not the org, not the network.

Long live supertools for one — and for the many selves we can now become.

If you liked this essay, consider sharing with a friend or community that may enjoy it too. (If you share / post a screenshot on socials, tag me — I’m mostly here and on Twitter.)