a16zReplit CEO Amjad Masad on What Young People Should Learn in the AI Era
CHAPTERS
- 0:00 – 1:00
Young people as heretics: the real purpose of youth (and what college gets wrong)
Amjad argues that young people’s core social value is to challenge deeply held beliefs—yet modern college often trains conformity. He frames this as a civilizational loss: institutions should protect and encourage first-principles questioning, not “beat it out” of students.
- •Youth’s role is to question assumptions and disrupt monocultures
- •College often optimizes for credentialing, obedience, and employability narratives
- •The old promise (do well in college → stable life) is breaking down
- •Societies (including Silicon Valley) develop blind spots without heretical experimentation
- 1:00 – 5:36
“The City and the Stars”: a sci-fi parable about disruption and curiosity
Amjad recounts a sci-fi story about a perfectly managed AGI-run city where a “glitch” produces someone truly new—who then starts asking forbidden questions. The parable sets up why curiosity and boundary-testing are essential for progress.
- •AGI-governed utopia breeds stagnation and norm enforcement
- •Memory “compression” and retraining create sameness; novelty requires a glitch
- •The protagonist’s disruptive questioning reveals what the culture suppresses
- •The story becomes a lens for what education should cultivate
- 5:36 – 8:46
Designing the ideal 18–22 experience: freedom, meandering, and exceptional peers
Asked what he’d design for an 18–22 experience in the AI era, Amjad emphasizes non-prescriptive learning, time to build, and surrounding yourself with unconventional people. He highlights “meandering” as a feature, not a bug, of meaningful education.
- •Choose an environment with many paths, not rigid requirements
- •Steve Jobs’ calligraphy example: ‘random’ interests compound later
- •Give builders space—don’t punish intensity with attendance bureaucracy
- •Peer quality matters: conventional environments create conventional outcomes
- 8:46 – 10:52
Side quests should be supported, not hidden: using institutional resources to build
Gagan describes how many students already juggle work and projects, but institutions still don’t endorse it. The group argues colleges should actively resource ambitious side quests—so students don’t have to “hide investor calls in the bathroom.”
- •Students increasingly live “side quest” lives; policy hasn’t caught up
- •Schools should provide labs, compute, and mentorship for real projects
- •Success shouldn’t create adversarial dynamics with the institution
- •The best ecosystems bring more resources to what’s working
- 10:52 – 14:32
Curiosity-driven, project-based learning—now unlocked by AI personalization
They argue project-based learning should be the default: start with a student’s interests and back into skills. AI changes the practicality—customized learning paths can be self-directed, with tools acting as tutors and accelerants.
- •Project-based learning connects skills to real utility (e.g., stats via basketball)
- •Many learners are motivated by skills toward building, not abstract knowledge
- •AI enables individualized curriculum without heavy institutional overhead
- •Education should validate autonomy and confidence, not just compliance
- 14:32 – 17:17
The counter-critique: ‘professionalizing too soon’ and the false college-vs-startup choice
Erik raises the concern that optimizing for career outcomes too early can crowd out personal and intellectual development. Gagan argues young people are pushed into an unfair binary—rigid college or immediate startup fundraising—when most need a low-pressure exploration period.
- •Some value of college is social/intellectual maturation beyond technical skills
- •The current system forces a binary: structured school vs VC-backed startup
- •Most teens shouldn’t make a 10-year commitment at 18; exploration matters
- •A “middle ground” model: time to explore with resources available when needed
- 17:17 – 19:40
Premature optimization in careers: why envy-fueled startups can be predatory
Amjad compares early startup fundraising to premature optimization: kids pre-commit to ideas they don’t deeply care about, often driven by social comparison. He also critiques “pre-idea” funding as structurally predatory—locking young founders into cap tables too early.
- •Starting a startup can be premature optimization—optimizing before you understand the problem
- •Envy and social pressure can replace genuine curiosity as motivation
- •‘Pre-idea’ million-dollar checks can trap founders before they’re ready
- •The right idea should feel unavoidable—momentum should pull you in
- 19:40 – 24:13
Entrepreneur + intellectual: following itches from chess to ‘vibe research’ and new products
Amjad explains how intellectual exploration and building form a loop: learning feeds building, which creates new questions to learn. He uses his chess obsession and attempts to train models as an example of curiosity turning into product direction—especially as AI lowers the barrier to ML research.
- •Oscillation between ‘read/explore’ and ‘build/ship’ is mutually reinforcing
- •Chess led to coding experiments, model training, and insights about LLM limitations
- •AI is making ML more accessible (“auto research/vibe research”)
- •Seemingly random interests can produce market and product ‘alpha’
- 24:13 – 27:00
Learning to follow vs learning to lead: autonomy, trust, and the self-driving company
They discuss how schooling over-indexes on compliance, while the future may demand autonomous, high-agency contributors. Amjad argues AI could change organizational structure by making “bureaucracy invisible,” enabling more autonomy and new forms of coordination.
- •People need enough disagreeableness to be original, but not alienating
- •‘Following’ can mean learning systems and collaboration, not just obedience
- •AI agents may become the organizational glue that makes autonomy scalable
- •Amjad’s vision: firms evolving toward a ‘self-driving company’
- 27:00 – 36:38
Discernment, boundary-testing, and ‘meta…maxing’: when to trust elders and when to break rules
The conversation turns to judgment: respecting what exists while still changing what’s broken. They explore how to handle “troublemakers,” the role of sandboxes (CTFs, bounty programs), and how repeated reps build discernment about risk, upside, and downside.
- •Respect the past while being unapologetic about changing it
- •Create environments with fewer status games (no grades/tests) to reduce rebellion dynamics
- •Channel hacking/boundary-testing into safe sandboxes and cybersecurity pipelines
- •Discernment comes from reps—making choices, failing, and updating heuristics
- 36:38 – 44:02
Beyond skills: becoming a good person, moral development, and communicating beyond the in-group
They argue “values formation” can happen in high-trust, positive-sum ecosystems—if interpersonal skills and responsibility are taught explicitly. Amjad emphasizes the need for well-adjusted adults who can communicate with broader society, not just tech insiders, and outlines his own path through moral philosophy toward virtue ethics.
- •High-trust cultures reward reputation and reliability; ethics become practical
- •Silicon Valley is now an institution—its builders must interface responsibly with the world
- •Reading widely (liberal arts, ethics) helps form an internal moral compass
- •Amjad leans toward virtue ethics: act in ways you’re proud of regardless of observation
- •Principles should be minimal/axiomatic and adaptable in a fast-changing world
- 44:02 – 47:03
What Amjad looks for in young talent: depth, persistence, and likability
In closing, Amjad shares what makes young people stand out: deep mastery in something they can teach others, and the ability to follow a thread beyond immediate reward. He adds that contrarianism without respect is a dead end—interpersonal skills and authentic likability matter.
- •A strong ‘spike’—they’ve gone unusually deep and can teach something
- •Evidence of persistence and intrinsic motivation over external rewards
- •Contrarian opinions are valuable, but arrogance repels collaborators
- •Baseline interpersonal skills and genuine likability increase trust and opportunity
