ExpertsFMAmjad Masad on Rethinking College for the AI Era
The a16z Show · September 23, 2026
5 Selected Lessons:
reasoning_pattern · 2 heuristic · 1 framework · 1 priority · 1
What Amjad looks for in young talent: depth, persistence, and likability
Source
The a16z Show — Amjad Masad on Rethinking College for the AI Era (September 23, 2026) · at 43:56 ▸
When this applies
when evaluating any junior candidate for high potential, regardless of industry
The takeaway
Rank depth-of-pursuit and genuine intrinsic motivation above raw breadth or contrarian intelligence, but require baseline interpersonal likability as a gating factor.
Why it works
Depth in one area (following a thread 'all the way through') signals intrinsic motivation and persistence, traits that predict future execution better than breadth or contrarian cleverness alone; likability signals the social capital needed for others to want to support them.
One level up
Technical or intellectual brilliance without the social trust to attract support is capped in its ultimate impact.
Where it stops holding
This stops working when a role genuinely requires broad generalist capability rather than deep specialization, or in contexts where likability signals can be faked/gamed.
Detail
hierarchy:
- 1. Depth: did they follow a thread/project all the way through, able to teach you something
- 2. Contrarian but respectful: having real opinions without dismissing others as idiots
- 3. Interpersonal skills / authentic likability: people need to want to be around and support you
rationale: 'Most people lack' the ability to go deep and most 'get bored easily'; depth plus likability predicts someone worth betting on.
In their words
I think the first thing is that they dove really deep in something. They can teach me something
Interest-anchored project-based learning enabled by AI
Source
The a16z Show — Amjad Masad on Rethinking College for the AI Era (September 23, 2026) · at 10:53 ▸
When this applies
when teaching any skill to a learner who has their own pre-existing interests
The takeaway
Anchor skill acquisition to the learner's own stated interest and let them discover the tool's utility by building, rather than teaching the tool divorced from application.
Why it works
People learn and retain material far better when its utility is tied to something they already care about, rather than abstracted into a generic syllabus; AI now removes the cost of customizing a path for each learner.
One level up
When a previously impossible-to-scale personalization becomes technologically cheap, what was once an aspirational ideal becomes the only defensible default.
Where it stops holding
This stops working when learners have no clear personal interest to anchor to, or when foundational skills must be sequenced in ways incompatible with ad hoc project pursuit.
Detail
procedure:
- Identify the student's genuine interest (e.g., basketball, chess)
- Let them use the skill (coding, stats) as a tool to build something they care about
- Use AI to remove the need for a teacher to manually customize a syllabus per student
- Treat 'project-based learning' as the default standard, not a special buzzword
rationale: Historically it was impractical for every student to get a customized curriculum; AI now makes it not just possible but the obviously correct default.
In their words
The great thing about today, you know, with AI is every single student can go and do that on their own, and they don't need a customized syllabus or exercise given to them, they just need the freedom and confidence to say, yes, you're allowed to do this.
Startup commitment as 'premature optimization'
Source
The a16z Show — Amjad Masad on Rethinking College for the AI Era (September 23, 2026) · at 17:11 ▸
When this applies
when anyone is about to commit years to a path mainly because peers are succeeding at it (status/envy-driven commitment)
The takeaway
Test for whether a commitment is driven by positive pull (an idea with its own gravity you can't avoid) versus negative push (envy, peer pressure, external capital availability) before locking in.
Why it works
Committing resources to a path before the underlying motivation (genuine interest/conviction) is validated wastes effort on something 'not eventually relevant or consequential' — the programming analogy of optimizing code before knowing it matters.
One level up
Resource allocation decisions should wait for genuine signal of fit, not for the mere availability of resources (capital, opportunity) to force the decision early.
Where it stops holding
This stops working when waiting for 'gravity'/conviction causes someone to miss a genuinely time-limited market window, or when the person's self-assessment of motivation is unreliable.
Detail
approach: Ask: is this idea pulling me forward with unavoidable momentum, or am I pushing myself into it because of envy/peer pressure/available capital?
signals:
- Idea 'explodes into a startup on its own' despite resistance
- Motivation framed as envy of peers' success
- Pre-committing to an idea before genuine interest is tested
- VCs offering large sums pre-idea, creating predatory lock-in
rationale: Premature optimization (committing before relevance is proven) is 'the root of all evil' in both programming and career choices.
In their words
And then they pre-commit to an idea they're not even interested in and they raise money for it.
Itch-following as a recursive build/explore cycle
Source
The a16z Show — Amjad Masad on Rethinking College for the AI Era (September 23, 2026) · at 19:37 ▸
When this applies
when an operator or founder notices an unrelated curiosity pulling their attention away from current priorities
The takeaway
Treat curiosity ('the itch') as a signal worth following even without a visible payoff path, because the payoff (new business line, new insight) emerges from following through, not from pre-planning.
Why it works
Oscillating between building and exploring lets accumulated knowledge from exploration surface new problems worth building for, and building surfaces new things worth exploring — a compounding loop.
One level up
Suppressing unexplained personal curiosity in favor of pure goal-direction removes the raw material future opportunities are built from.
Where it stops holding
This stops working when there's no slack/resources to explore without payoff, or when the person lacks the skill to convert exploration into execution (turning an itch into a built thing).
Detail
approach: Notice an unexplained itch (interest) -> pursue it without justifying the payoff -> let it recursively connect to building something -> often results in a genuinely new capability or business line.
signals:
- Interest arises for no clear reason
- Investigation reveals an unsolved gap (e.g., LLMs bad at chess)
- Exploration naturally shifts into building (training models)
rationale: 'A lot of kids get the itch beaten out of them' through institutional pressure to just follow instructions, losing this generative capacity.
In their words
I realize that AI is actually now capable of doing ML
Institutions should supply resources to unconventional student projects, not just tolerate them
Source
The a16z Show — Amjad Masad on Rethinking College for the AI Era (September 23, 2026) · at 8:46 ▸
When this applies
when a junior person inside any institution starts an unsanctioned project with real traction
The takeaway
Don't just permit high-potential deviation from the norm — actively route resources toward it once it shows traction.
Why it works
If the institution only tolerates deviation rather than actively supporting it, talented people are forced to hide their best work, wasting the institution's own resources and signal.
One level up
The value of flexibility in a system is realized only when the system rewards divergence, not merely avoids punishing it.
Where it stops holding
This stops working when the 'unconventional project' is actually harmful or when resource allocation to outliers starves the core mission supporting everyone else.
Detail
rule: Students pursuing unconventional, promising projects should get college resources (compute, labs, visibility) proactively offered, not forced to hide.
threshold: Trigger point: when a student's side project is 'taking off' — i.e., showing external validation like investor interest.
rationale: Hiding talent wastes both the student's and institution's potential; visibility lets the institution amplify genuine breakthroughs.
In their words
There's this kid, Eric, I think he's Richard's friend, who has to take like investor calls from the bathroom because he's like hiding in the school. They should be like, you know, coaching him through it or enabling it in some way, bringing investment demo day or something, this kind of thing.