Everyone Says, 'Just Build.' I'm Not So Sure.

By: Palash Jain

In the last 52 days, Anthropic’s Claude team dropped 50+ major updates. OpenAI and Google matched that pace. A tech journalist who's covered startups for twenty years said he couldn't remember a company shipping this much, this fast. My LinkedIn feed was a blur of launches, each one treated like it changed everything.

I watched it all feeling a strange déjà vu. Because I've been here before.

At Broadcom’s VMware, a Fortune 10 cloud computing company, I led a platform team that went from a single release a month to 30+ per a month. This was unprecedented velocity for an enterprise infra company and boy, we were proud of it. And then we started noticing that customers were confused. Sales couldn't explain what had changed. Marketing was still catching up on the last release while three new ones shipped. By the time anyone absorbed what we'd built, the needs had already moved on. The velocity was real. The value was not.

So, we slowed down. We figured out what people needed. And the work got better.

The AI industry feels like it's in the phase my team was in: impressive speed, unclear destination. Nobody wants to slow down because slowing down feels like falling behind.

The numbers don't settle the question, but they make you pause. Nearly $250 billion in venture capital went into AI in 2025—55% of all global VC investment. Add Big Tech's infrastructure spending and you're approaching $500B+. Enterprise AI revenue that same year? About $37 billion. That's a lot of money going in for not a lot of value coming out. That’s almost a 10:1 ratio.

And the capital isn't spreading. More than 60% of AI funding went into mega rounds above $100 million, mostly to a handful of companies. This is where I keep coming back to the SaaS era. SaaS worked—sustainably—because the iPhone created a distribution layer. Profits didn't pool at the top. They spread across developers, app ecosystems, accessory makers, global supply chains. Everyone participated in the upside. That's what made it sustainable. But AI hasn't found its iPhone. There's no hardware product, no distribution mechanism turning this investment into broadly shared economic returns. I find this model hard to sustain. That's not pessimism. That's finance 101.

But honestly, a bigger dilemma is the productivity question.

Everyone says, "just build." Increase productivity. President and CEO of NVIDIA, Jensen Huang, says pay people in tokens. The message everywhere is produce more, faster.

I read an amazing line by Elena Verna who leads growth at AI vibe coding platform, Lovable:

“If AI makes you 10x faster, nobody says, ‘Amazing. Please enjoy your afternoon!’ They ask why you're not doing 10x more.”

One of the most useful things I picked up during my Kellogg MBA is the distinction between value creation and value capture. Simple idea, but it reframed how I see everything now. If AI commoditizes creation—or in other words, if anyone can produce, generate, build—then creation isn't where the value lives anymore. The value has moved to the quality of human input.

So then why is everyone telling me to optimize for the commodity, telling me to build more, do more?

The things that make humans better at thinking—travel, spending time with friends, attending a concert, walking through a museum—we treat those as breaks from real work. But the research says otherwise. Travel measurably increases cognitive flexibility according to Psychology Today. Ten minutes of friendly conversation boosts executive function as much as doing crossword puzzles according to research from the University of Michigan. According to Harvard Medical School, strong social ties build cognitive reserve—the brain's actual ability to cope and perform. And yet nobody counts these things as productivity.

We say technology, advancement, and war is for humanity. But what I see right now is wealth consolidating, opportunity concentrating, and the very jobs that gave people a sense of purpose getting replaced.

But here's the thing: I don't think that's inevitable. I think it's a choice we're making by treating AI as something that goes everywhere, all at once.

Let me propose a contrarian argument, just for a moment. We've had powerful technologies before; think nuclear energy, breakthrough medicine such as penicillin, heavy machinery, and missile systems. We didn't let any of them loose and say, "just apply it to everything and see what sticks." We were intentional. We asked: where does this create genuine value? Where does the benefit outweigh the cost? Where does it belong—and where doesn't it?

What if we treated AI the same way? Not as a universal accelerant, but as something we apply with care, for specific purposes where it genuinely matters scientific research, healthcare, education (the biggest, in my opinion), climate, policy, etc. The places where human limitation is the actual bottleneck, not human laziness.

Not everything needs to be faster. Some things need to be slower to be good. And I say this coming from a background that lived by the fail fast mindset. Understanding what a customer struggles with takes time. Getting the small things right—the experience, the details that make the customer feel proud of using it—takes time.

Sometimes the most productive thing you can do is sit with a problem long enough to understand it before you ship anything.

Read More Tech Talks: Why the Builder of an AI Editing Service Says He Isn’t Replacing Editors Data, Deals, and Government. What’s New with Nuclear? The Wall-less Empire: The Illusion of Open AI Markets The SpaceX Diaspora: How Rocket Science Is Transforming Startup Culture