A Third Path, Paathuklam | Prakash Suresh

Sridhar Vembu - the man who bootstrapped Zoho into a billion-dollar company without a single dollar of venture capital - recently told programmers to “start considering alternative livelihoods.”

When someone with that track record says something that blunt, you don’t dismiss it. You sit with it.

Sitting with it also means thinking it through. He’s right about the disruption. But more than the argument, I want to talk about the posture - because how you hold this moment might matter more than what the moment is.

The Smartest People Are Worried

Sridhar’s argument came in two parts. First, he pointed out that SaaS stocks are getting hammered - and for good reason. The industry was always vulnerable, spending more on sales and marketing than on engineering. AI-assisted coding is the pin popping that bubble.

Then came the second, sharper claim. A non-coder built a full iOS app. Anthropic built a C compiler with Claude. His conclusion: if you write code for a living, it’s time to think about what comes next.

On January 30th, Anthropic released 11 open-source plugins for Claude Cowork - covering sales, legal, finance, customer support. Each could connect to a company’s existing systems and automate entire workflows.

Investors saw the implication immediately: why pay for Salesforce, LegalZoom, or FactSet when an AI agent with the right plugin could do the job? Software stocks began their freefall.

Days later, Anthropic launched Opus 4.6 with “agent teams” - AI models coordinating other AI models to break down complex tasks autonomously. CNN ran the headline: “The AI that shook software stocks gets a big update.”

That same week, Sridhar posted the line that cut deepest:

Those who write code for a living should start considering alternative livelihoods.

The evidence is piling up fast.

ServiceNow’s stock has dropped 50% in the past year - despite reporting better-than-expected earnings. Salesforce is down 40%. Adobe, 35%. Roughly $830 billion has been wiped from software stocks since late January. Bloomberg is calling it the “SaaSpocalypse.”

Salesforce quietly moved Heroku - the platform that defined cloud deployment for a generation of developers - into “sustaining engineering” mode, stopping enterprise sales to new customers. That’s not a pivot. That’s a wind-down.

These aren’t speculative fears. Companies are reporting good numbers and still getting punished. The market is pricing in a world where traditional software gets disrupted by AI - not eventually, but now.

The fear is real. It’s rational. And I think it’s also incomplete.

Intelligence Has Never Been Saturated

Here’s the pattern that keeps repeating: every time we’ve increased the supply of intelligence, demand has grown faster.

Spreadsheets didn’t kill accountants. They created more accounting.

Suddenly every small business could model scenarios, track expenses, forecast revenue. The number of people doing “accounting work” exploded - most of them weren’t accountants.

ATMs didn’t kill bank tellers. Banks opened more branches because operating costs dropped, and tellers shifted to advisory roles. The industry grew.

The internet didn’t reduce the need for writers. It created a world where every company needs content, every product needs documentation, every person has a publishing platform. There are more people writing professionally today than at any point in history.

The pattern is consistent: when you make intelligence cheaper, you discover uses for it that were previously unthinkable. Not slightly more uses. Dramatically more.

GPT-4 launched at $60 per million output tokens. Today, equivalent capability costs under $1. That’s a 99% price collapse in under three years.

Demand didn’t fall. It exploded. OpenAI went from $1B to $12B+ in ARR while slashing prices every quarter.

A task that cost $1,000 in AI compute now costs under $4. At that price, companies don’t just automate what humans were doing. They do things that were never viable at human-labor pricing. Analysis that would have required a $200K analyst for a year now runs for $50 in an afternoon.

That’s not replacement. That’s expansion.

What Changes Is the Work, Not the Demand

But expansion at the macro level doesn’t automatically calm the individual engineer wondering if their job is next. Raju Vegesna - Zoho’s Chief Evangelist - put it this way:

The farmer used to work the land. Now he watches a screen while the combine works. The pilot used to fly the plane. Now he monitors instruments while the plane flies itself. The programmer used to write the code. Now he increasingly watches the screen while the system writes it. The job didn’t disappear. The role transformed.

Every wave of automation removes execution and leaves supervision. As Gopi Krishna Lakkepuram, who spent over a decade at Microsoft and now builds Hyperleap AI, put it: AI replaces the task, not the engineer.

Yes, the C compiler is impressive. But someone had to know what a C compiler is, why it matters, what edge cases to test for, and when the output was wrong.

That’s engineering judgment. No model has that yet.

The engineer who only knew how to write syntax was always at risk - long before AI showed up. The one who understood why the code needed to exist? That person just got a massive upgrade.

And look at what’s happening. A creator who built a Bhagavad Gita app with AI didn’t replace a programmer. He created demand that didn’t exist before.

He was never going to hire a developer for that project. AI made it possible, and now it exists.

Sridhar himself used Gemini as “an extremely intelligent economic philosopher” to think through policy questions. That’s not code replacement - that’s a new use of intelligence entirely.

Every single one of these examples is additive.

The Real Bottleneck Shifts

When intelligence becomes abundant, the scarce input changes.

Raw coding ability - the capacity to translate a spec into working syntax - is becoming a commodity. But knowing which problem to solve, how to frame it, and whether the output is right? Those are becoming more valuable, not less.

The returns are flowing to people who can direct intelligence, not just provide it. Knowing what to build - understanding the problem well enough to direct the solution - matters more than generating it.

This has always been true, but the gap is widening fast. And that gap is where the opportunity lives.

That’s less about technical skill and more about posture. James Clear put it well:

Entrepreneurship is just the ability to trust that you’ll figure it out because you’re constantly facing some challenge or problem. No matter where you’re at on the curve.

That’s the skill that transfers. Not any specific language or framework, but the confidence that you can figure out whatever comes next. The tools change. The problems change. The ability to adapt doesn’t become less valuable - it becomes the whole game.

No Time for Caution

Everything above is the argument. This is the part that actually matters.

Adaptability requires a clear head.

Naval Ravikant makes a point that sounds simple but cuts deep: being unhappy is extremely inefficient. An anxious mind doesn’t sleep well, reacts with anger, makes impetuous decisions. It gets caught in the busy trap - running from one thing to the next without ever prioritizing.

A peaceful mind makes better decisions. - Naval Ravikant

If we’re in panic mode right now - doom-scrolling layoff threads, catastrophizing about career extinction - we are actively making ourselves worse at the one thing that matters: navigating what comes next. Panic is not just unpleasant. It’s strategically bad.

And it doesn’t just affect individuals. Organizations panic too.

The instinct to tighten control - mandating presence, restricting flexibility, reaching for oversight instead of trust - is the institutional version of doom-scrolling. It feels productive. It’s not. The best work has always come from conviction, not compliance.

There’s a scene in Interstellar where Cooper has to dock with a space station that’s spinning out of control. His robot companion TARS says “It’s not possible.” Cooper’s response: “No. It’s necessary.”

He doesn’t pretend the danger isn’t real. He doesn’t minimize the spinning. He analyzes the spin rate, matches it, and docks.

And notice - Cooper doesn’t dock alone. TARS, an AI, helps him match the spin. The AI said it was impossible. The human decided it was necessary. And then they did it together.

The soundtrack for that scene is called “No Time for Caution.” The title is the message. There’s no time for caution, but there’s also no time for panic. Only for calm, precise action on the variables you can actually control.

That’s the posture this moment demands. Not denial, not panic - analysis and execution.

What skills do you have?

What can you build now that you couldn’t before?

What problem have you always wanted to solve but couldn’t justify the cost?

The answers are available. But only to a mind that’s calm enough to see them.

A Third Path

With that clarity, let’s return to what Sridhar envisions. He offers two futures. In one, technology becomes trivial and we refocus on life, nature, community, and faith - people return to rural areas and live richer lives. In the other, centralized control leads to a dystopian concentration of power.

The optimistic vision is beautiful. And I think there’s a third path alongside it - one where intelligence becomes so accessible that it creates entirely new categories of work, new forms of value, new problems worth solving.

Not everyone retreats to the countryside. Some people discover that the thing they’ve always wanted to build is suddenly within reach.

Look at where the untapped demand lives. The Blume Ventures Indus Valley Report splits India into three economic realities.

India 1 - the elite 140 million, per capita income of $15,000, digitally empowered, already plugged into the global economy.

India 2 - 300 million aspirational middle class, $3,000 per capita, skilled but stagnating, vast potential being wasted.

India 3 - over a billion people, the informal workforce, excluded from the digital economy entirely. The report calls them “the unmonetised segment.”

That framing assumes intelligence stays expensive. If it doesn’t - if the cost of analysis, research, and expertise collapses the way compute costs have - then India 2 and India 3 aren’t unmonetised. They’re underserved. And underserved is a very different problem from unmonetisable.

The consultancy that used to serve five clients a year now runs hundreds of bespoke analyses a week. The teacher in a rural school now has a personal tutor for every student. The local business that could never afford market research now runs it over lunch for the cost of a coffee.

These aren’t India 1 stories. These are India 2 and India 3 stories - made possible only when intelligence gets cheap enough.

And here’s the thing - that’s not so different from what Sridhar describes. His vision of people returning to rural communities, living richer lives rooted in nature and faith - that’s also a world where judgment, context, and local knowledge matter more than raw technical skill.

The third path and his path might be the same path, viewed from different angles. One emphasizes what we build. The other emphasizes how we live. Both depend on the same shift: intelligence becoming abundant enough that what you know matters less than what you see.

Intelligence isn’t just code. It’s judgment, taste, context, problem selection, the ability to see what’s missing. The demand for that kind of intelligence has no ceiling - and a billion people in India alone have barely begun to access it.

As Gopi put it - “Don’t abandon the craft. Evolve it.”

In Kamal Haasan’s Vikram, when everything seems impossible, he doesn’t make speeches. He just says: paathuklam - we’ll handle it.

Not bravado. Not false optimism. Just the quiet confidence that whatever comes next, we’ll figure it out.

The demand for intelligence is unlimited. It always has been. The only thing that changes is what we mean by intelligence.

Paathuklam.

References: