Jul 8, 2025 · 38m · y-combinator
How to Spend Your 20s in the AI Era · Y Combinator
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In a live podcast recording, Y Combinator partners advise ambitious young engineers and founders on navigating the AI era by building authentic utility, adopting extreme agency, and targeting high-value niche opportunities rather than chasing hype or credentials.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The partners hold 53.9% of the talking time here. How this is scored →
The partners as informed peer 6.4 Guest teaching 4.3 Guest disagreement 1.7 The partners pushing back 0.9
speaking balance: gold is the partners, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 23:57 Gary rejects superficial aura farming and fake tech simulacrum
Gary aggressively condemns fake hype in tech, pointing to disgraced figures like SBF and Elizabeth Holmes to argue that anything detached from real utility is a lie.
Hardest push from the partners ▶ 7:08 Harj dismantles the 'last window to get rich' panic narrativeHarj directly challenges the premise of the panic, referencing Paul Buchheit to argue that if true AGI arrives, accumulating human currency becomes meaningless anyway.
Biggest teaching moment ▶ 2:55 Gary reveals art history majors have lower unemployment than CS gradsGary surprises the panel by quoting official New York Fed labor data showing CS majors suffering more than double the unemployment rate of art history graduates.
The partners hold their own ▶ 35:07 Harj outlines historical precedent for niche market dominationHarj demonstrates commanding institutional knowledge by breaking down how massive giants like Airbnb and Stripe originated in minuscule, dismissed niches.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The partners as informed peer | Guest teaching | Guest disagreement | The partners pushing back | Why |
|---|---|---|---|---|---|---|
| Welcome to Lightcone Podcast Live | 6 | 5 | 1 | 1 | The panel discusses whether the AI boom represents the last window to achieve wealth. Gary Tan brings up surprising New York Fed data regarding CS unemployment compared to art history, prompting Harj Taggar to cite Bryan Caplan's signaling theory of education. | |
| Outdated CS Curricula vs. Agency & Vibe Coding | 6 | 3 | 1 | 1 | Diana Hu and Jared Friedman note how universities ban modern coding tools like Cursor, comparing it to early internet bans on Google search. Harj pushes back against panic-driven startup founding by referencing Paul Buchheit's perspective on human money in an AGI era. | |
| Hyper-Growth AI Startups & Real Utility vs. Hype | 7 | 4 | 2 | 1 | The hosts and Gary Tan analyze hyper-growth trajectories in AI, highlighting companies reaching massive revenue milestones with tiny teams. Harj explains how value shifted from domain relationships back toward high-level engineering execution. | |
| Gaining Domain Expertise as 'Undercover Engineers' | 7 | 5 | 1 | 0 | Gary and the hosts discuss how young founders can gain deep domain knowledge by acting as undercover engineers. Jared cites three recent YC batch founders successfully building vertical AI for dentists without any prior dentistry experience. | |
| Overcoming Student Mindsets, Investor Idols, and Scammy Programs | 6 | 5 | 3 | 1 | Diana urges students to discard formulaic checklist mentalities, emphasizing that startup building is an open landscape. Gary fiercely criticizes university entrepreneurship retreat programs that incentivize deceptive behavior and fake-it-till-you-make-it posturing. | |
| Social Media, 'Aura Farming,' and Building in Public | 5 | 6 | 4 | 2 | Harj asks about the balance between building and cultivating an online audience. Gary passionately dismisses superficial social media 'aura farming' as fake simulacrum, before outlining a practical method of using bi-weekly Loom product demos to drive development. | |
| Audience Q&A: Should I Drop Out for a Startup Offer? | 7 | 2 | 1 | 2 | An audience member asks if they should drop out after receiving a startup job offer. Jared offers direct guidance based on his own experience dropping out, advising the student to avoid fear-based FOMO decisions. | |
| Audience Q&A: Leaving Corporate Jobs & Finding Co-Founders | 6 | 4 | 1 | 0 | An audience member asks about the financial and career transition from a corporate job to a startup. Gary candidly describes his early personal debt mistakes, and Jared identifies co-founder life-stage alignment as the primary gating factor. | |
| Audience Q&A: Starting Super Niche in the AI Era | 8 | 5 | 1 | 0 | An audience member inquires about the necessity of launching in hyper-niche markets. Harj and Diana cite early playbooks from Airbnb, Stripe, and Coinbase, while Gary explains how high-IQ reasoning models serve as potent wedges into niche datasets. |
Statements from this episode (18)
Garry Tan: CS majors face higher unemployment than art history majors
“The craziest stat, I think this came out of the New York fed in February of this year computer science majors You know, obviously this is not the people in this room. This is just, like, out of, like, you know, a normal distribution of all computer science maj…”
Harj Taggar: Humans cannot outcompete AI on instructions; agency will matter
“It's pretty clear now in the AI world that, like, the AI is very good at following instructions, and it's probably going to be hard for humans to compete with the AI on just, like, Following instructions reliably, in which case people here need to think about …”
Friedman: Universities prohibit students from learning essential future AI tools
“They're quite literally prohibiting the students from learning the tools that they are going to need in the future.”
Taggar: The 'Last Window to Get Rich' Logic Is Flawed Under AGI
“If this is actually the last window to make money and get rich, then you're basically implying that, you know, that we're going to get some definition of AGI or ASI, or whatever you want to call it. That's like a necessary condition for that to be true. Like, …”
Tan: 5-to-10-person AI startups are hitting $10M–$12M in net revenue
“The really big difference today is that the very best companies that we get to see day to day, they're like, I don't know, five people, 10 people I think, you know, each of us on stage and all of the YC partners are sort of collecting incredible startups that …”
Hu: Cursor scaled from $1M to $100M revenue in one year
“We've seen crazy growth unlike anything, only possible with, right now with AI, like all these companies that we work with, zero to twelve million in 12 months. The cursor story where they went zero to one in one year, the next one to a hundred. This is, like,…”
Friedman: AI has inverted SaaS growth, making B2B the hyper-growth sector
“It didn't used to be the case that B to B SaaS companies were the ones that had hyper growth. Like, there were some, like, consumer social companies that got hyper growth, but, like, B to B SaaS used to be this, like, you know, flotting, slow growing, like, ki…”
Taggar: College students often extract better AI model performance than PhDs
“The technical expertise is now actually really like the missing piece for a lot of these things. And we consistently see, at least in YC, that college students are actually at the forefront of this stuff, like actually understanding how to use the models and h…”
Harj Taggar: Incumbent software companies underwhelm businesses seeking AI solutions
“Because, like, AI has captured the, like, the imagination of everybody, everyone, like, wants to know what's possible and are consistently underwhelmed by what, like, the established software companies can offer them, but they're open to, like, college student…”
Garry Tan: Scammy founder programs produce future SBFs and Theranos fraudsters
“I worry that some of these programs are just literally trying to teach people to become more you know, SBFs and Theranos. And that's like, you know, that's a waste of time. Like, and you're going to go to jail.”
Friedman: Teaching entrepreneurship as a college course creates a cheap facsimile
“Anytime you try to bottle up entrepreneurship and like teach it as a college course, that's kind of what you end up with is like, basically like a sort of like cheap facsimile of entrepreneurship where like they teach you to like, you know, Follow a particular…”
Garry Tan: Plan two-week engineering sprints backwards from demo videos on X
“A much more powerful version of this is let's work backwards from what I want to put on X. I'm going to make a very simple Loom video showing off a feat of strength, a thing that I really want to share that I know our team can do, and then working backwards fr…”
Friedman: Dropping out of college due to startup FOMO leads to regret
“When college students are thinking about dropping out, and they're making it in a fear, they're making a fear-based decision, they have FOMO about startups happening in San Francisco, maybe their friend dropped out, and they're, like, worried that they're gonn…”
Tan: Evaluate startup jobs like an investor because median startups fail
“Like objectively, you should make literally a spreadsheet. You should, you know, go down and like evaluate it the way an investor would. And then the difference is the investor has a portfolio and you just have one life. And likewise, when you start a startup,…”
Garry Tan: Aspiring founders need 6 to 9 months of runway
“I mean, I think you want, I don't know, at least six, maybe nine months minimum of like, just, you know, being able to live on ramen in the cheapest possible way.”
Jared Friedman: The biggest bottleneck for corporate employees is co-founder timing
“And in my experience for people who are your age, who have like already graduated college or working at some company, that's the biggest limiting factor in practice for them actually doing a company is like they and their co-founder both need to be willing to …”
Taggar: AI startups command higher pricing by selling work rather than software
“Things that seem like they were niche before AI, you can get people to pay you a lot, lot more money for because they're not just buying software, they're buying like work from you.”
Tan: OpenAI o3 operates around 130 IQ, smarter than many past hires
“I think of O-three as basically about a 130 IQ, maybe O-three pro can be even smarter than that. When I really think about that, it's like, oh yeah, like a lot of the people who I've ever hired in my lifetime are like, yeah, O-three is smarter than that person…”