Oh Good, the VCs Will Survive

As AI blitzes forward—reshaping industries and replacing roles—Marc Andreessen has some comforting news: venture capitalists will be fine.

In a recent podcast, the co-founder of Andreessen Horowitz (a16z) predicted that his profession might be one of the last to survive the AI wave. Why? Because it’s not just about numbers or models. It’s about “psychological analysis,” he says. It’s about relationships, pressure, instability, intuition. “You end up being a psychologist half the time,” he added.

How ironic that the man whose firm helped fund the very technologies now displacing workers is declaring that the most human job left... is his own.

The Luxury of Intangibility

Andreessen isn’t wrong. Venture capital, like many forms of high-stakes decision-making, depends on qualities that are hard to code: judgment, trust, timing, instinct. These are things that don’t fit neatly into a training set.

It’s not that VCs are immune to automation. It’s that their value lies in soft power—influence, persuasion, moxie—the very things most workers have never been paid for. And so, while CEOs from Shopify to Klarna boast about how AI has replaced hundreds of workers, Marc Andreessen offers a different message: Don’t worry, the gatekeepers will be fine.

Keystrokes for Dollars

In contrast to April’s 2025 Microsoft Work Trends Index report, The World Economic Forum’s Future of Jobs Report estimates that 40% of employers plan to reduce headcount in AI-automatable areas within five years. That includes roles in customer service, analytics, basic coding, and information processing (read: loan officer)—jobs that once represented the rungs of upward mobility.

At Shopify, CEO Tobi Lütke recently told staff: before asking for more people, prove AI can’t do it. At Klarna, CEO Sebastian Siemiatkowski says their chatbot is already doing the work of 700 people. At Microsoft, 30% of all code is now written by AI. And at Meta, Mark Zuckerberg expects that number to pass 50% soon.

Essentially, if your output can be measured in keystrokes you're replaceable. Unless you’re the one with the capital.

The Merits of Messy

Let’s pause for a moment and ask what Andreessen’s statement actually reveals. It’s not just about venture capital. It’s about the shifting boundary between human value and machine efficiency. The more a job relies on ambiguity the safer it is. Not because it’s more important, but because it’s harder to model.

Investors, psychologists, coaches, biologists, artists (some of them)—these roles endure not because they’re irreplaceable, but because they’re messy. And messy, for now, is still our domain. But messy doesn’t scale, and it doesn’t always pay well either.

A Fork in the Algorithm

Andreessen believes that picking startups is the “small part” of the VC job. The real value, he argues, lies in relationships. That’s probably true. But it raises an uncomfortable question: if the people who fund innovation believe their work is immune because it relies on human complexity, what does that say about the kind of innovation we’re funding?

It suggests a future optimized for the elite—preserved for those who own the ambiguity. Efficiency for the many, discretion for the few.

In a way, Andreessen’s confidence is oddly reassuring. Not because it guarantees anything for the rest of us—but because it reveals the dividing line between what we want AI to do, and what we want people to do.

If venture capital is safe, it’s not because it’s smarter, it’s because we haven’t figured out how to replicate messy. At least not yet. And if we ever do? It might just put a billionaire out of work.