a16z

Amjad Masad on why young people should be heretics

Amjad Masad· Founder & CEO of Replit at Replit
·~47 min·English·a16z
AgentsLLMAI CompanyBusiness StrategyTraining
TL;DR

In a conversation about reinventing college for the AI era, Amjad Masad argues that the real job of young people is to question deeply held beliefs — and that how we learn, build, and start companies all have to change to protect that.

01Core Mental Model

The Glitch in the Perfect City

Amjad Masad argues the real utility of young people is to be heretics — to question deeply held beliefs — and that today's college does the opposite by beating outside thought out of them.

the utility of young people is to question deeply held beliefs is to be heretics. And in today's college actually the opposite. It kind of beats any outside thought out of them

— Amjad Masad, a16z
Key Insight
The story reframes rebellion as a civilizational function, not a discipline problem: if every culture is a monoculture, the person who ignores its rules is the only one who can find what it cannot see.
02Design of the Academy

A College That Isn't Prescriptive

His ideal college is not prescriptive: freedom to choose what you study, space to keep building when an idea grips you, and peers who question rather than conform.

I'd want to go to a college that's not prescriptive about what I learn.

— Amjad Masad, a16z
Key Insight
The Jobs calligraphy example hides a stronger claim — that the value of 'useless' study is only visible in hindsight, so any system that prunes it in advance is optimizing for a future it cannot predict.
03How Learning Works

Build First, Then Learn the Tool

Learning should start from something you want to build, not a syllabus, because you keep skills you can use rather than knowledge whose purpose you never see — and AI now gives every student that path.

I'm motivated by what skills can it give me in order to achieve what I want to build.

— Amjad Masad, a16z
Key Insight
This inverts the usual order of school — utility first, theory as needed — and quietly makes AI the enabling technology: a tutor per student is what turns 'build to learn' from an ideal into a default.
04The Choice at 18

The Unfair Binary

Young people face an unfair choice — a rigid four-year degree or a $50k check to found a startup at 18 — when what most need is two low-pressure years to explore before committing.

I think the saying is premature optimization is the root of all evil. That's how bad people think it is

— Amjad Masad, a16z
Key Insight
The programmer's proverb does real work here: a startup raised on envy is optimization pointed at the wrong objective, and the price is a decade committed before the founder knows what they actually want.
05Following Curiosity

Vibe Research

A random itch — late-life chess — led Masad to discover that AI can now do machine learning itself, what he calls vibe research, and to start building it into Replit.

you could call it vibe research, you could call it auto research. I think that's Karpathy's term.

— Amjad Masad, a16z
Key Insight
The point isn't chess — it's that curiosity is a search with no visible gradient to the payoff. Masad is betting Replit on the recursion he lived: give anyone the tools and their itches will find the businesses.
06The Future of Work

The Self-Driving Company

Masad believes the shape of companies is about to change: agents can absorb the bureaucracy in the background, freeing people for real autonomy while AI judges their impact — a company that increasingly runs itself.

I've changed how I run the company and we call this idea of like a self-driving company eventually like the company will like run itself.

— Amjad Masad, a16z
Key Insight
The claim reframes org structure as a technology problem, not a management philosophy: earlier flat-org experiments (Medium's holacracy, GitHub's no-managers) blew up for lack of a substrate, and Masad is betting agents are the substrate that finally makes autonomy scale.
07How to Disrupt

Respect the Past, Then Change It

Masad's rule for entering any established company or field is to genuinely respect what has been built, then feel free to change as much of it as you can — even when it makes people angry.

what I try to tell people at the company is like we respect the past, but we're unapologetically going to change it.

— Amjad Masad, a16z
Key Insight
The two halves are load-bearing together: respect earns the standing to change things, and without it disruption reads as mere disrespect. Masad offers his own 'don't learn to code' reversal as evidence that being early and unpopular is the same position as being right, only earlier.
08The Human Element

Act in a Way You'd Be Proud Of

Masad rejects utilitarianism for virtue ethics: cultivate principles through reading and introspection, then act so you would be proud regardless of who is watching.

I sort of like found my moral compass by reading wildly and introspecting

— Amjad Masad, a16z
Key Insight
His reasoning is itself anti-fragile: as the world races toward the singularity, consequentialism fails because you can't predict outcomes, so a character you can carry into any future beats a rulebook written for this one.