Stanford’s CS183B, taught by Sam Altman with Y Combinator, is the closest thing to a canonical startup curriculum. Twenty guest lectures from founders and investors who built Airbnb, Stripe, PayPal, Facebook, LinkedIn, and the firms that funded them. Below is every lecture as a carousel: the core claims first, readable in seconds, with a link to the full talk under each.
Lecture 1: How to Start a Startup
A great startup needs an idea, a product, a team, and execution, and the idea is the part everyone underrates. Sam Altman and Dustin Moskovitz open the course by arguing you should build something a small number of people love, not something a lot of people sort of like.
Lecture 1 - How to Start a Startup (Sam Altman, Dustin Moskovitz)
youtube.com/@ycrootaccess
23 Sept 2014 · Startups
44 min watch → 80 sec
The lecture lays out the core conditions for startup success: a strong idea, a product users love, a capable team, and strong execution. It also argues that founders should start companies only when they feel compelled by a mission and a timely market opportunity.
A hypergrowth startup succeeds mainly through four factors, a great idea, a great product, a great team, and great execution. Luck still matters a lot.
The first job is to build something a small number of users love, and the best ideas usually start small, unpopular, and narrow.
2 / 18
Start a company only if you feel compelled to solve one specific problem.
Mission sustains the decade of effort, focus, and recruiting.
3 / 18
Starting a startup is less glamorous and less flexible than most people expect.
Do it because the problem matters and you are suited to solve it.
4 / 18
10x
Execution matters more than the idea.
Execution is also 100 times harder. Bad ideas still fail.
5 / 18
A startup idea includes more than the product.
1
Market size matters.
2
Market growth matters.
3
Growth strategy matters.
4
Defensibility makes replication hard.
6 / 18
A strong startup idea answers, "Why now?"
It could not work two years earlier, and waiting two years would be too late.
7 / 18
Choose a rapidly growing market over a large slow-growing market.
Founders usually cannot create demand for a market that does not want to exist.
8 / 18
Students often spot fast-growing new markets early.
They understand new technology and changing user behavior better than older investors or operators.
9 / 18
The best startup ideas usually sound bad at first.
They start small, win a narrow market, then grow with a much bigger future market.
10 / 18
Most successful pivots move toward something the founders personally needed.
Airbnb began with Brian Chesky needing rent money and having extra space.
11 / 18
If you are not building something you need, get extremely close to customers.
Work in their office or talk to them multiple times a day.
12 / 18
Good startup ideas are usually easy to explain in one sentence.
Clone ideas like "X with beautiful design" usually fail.
13 / 18
Build something a small number of users love, not something many users merely like.
Deep enthusiasm spreads. Weak enthusiasm rarely becomes love later.
14 / 18
Recruit early users manually, not through ads.
Ben Silbermann personally recruited early Pinterest users for feedback and advocacy.
15 / 18
A startup's early loop is simple.
1
Build the product.
2
Talk to users.
3
Change the product.
4
Measure retention and active use.
5
Track revenue and net promoter score.
16 / 18
Ignore most early distractions.
1
Ignore PR.
2
Ignore conferences.
3
Ignore advisers.
4
Ignore partnerships.
5
Protect your health.
17 / 18
Read the full source
Lecture 1 - How to Start a Startup (Sam Altman, Dustin Moskovitz)
youtube.com
YC Root Access
18 / 18
Full talk: https://www.youtube.com/watch?v=CBYhVcO4WgI
Lecture 2: Ideas, Products, Teams and Execution
The best ideas look bad at first, so a good idea that sounds obviously good is usually already crowded. Sam Altman on why mission-oriented ideas win and why the founders matter more than the market at the start.
Lecture 2 - Team and Execution (Sam Altman)
youtube.com/@ycrootaccess
25 Sept 2014 · Startups
46 min watch → 95 sec
Sam Altman lays out how early startups should choose co-founders, hire slowly, manage and fire employees, and execute with focus, speed, and intensity. The central operating rule is to protect momentum through growth, clear priorities, and fast decisions.
A startup lives or dies on product quality, founder quality, and execution. Ideas alone create no value, and founders set culture, speed, and standards.
Choose co-founders and early hires with extreme care. Momentum comes from focus, intensity, speed, and high quality.
02 / 20
Students often spot fast-growing markets early.
What you and your peers start using can predict the future better than formal market analysis.
03 / 20
A startup CEO has 5 core jobs.
01
Set the vision.
02
Raise money.
03
Evangelize the company.
04
Hire and manage the team.
05
Set the execution bar.
04 / 20
Execution means choosing the 2 or 3 top priorities and saying no to almost everything else.
Only the CEO can set how hard and how well the company operates.
05 / 20
Momentum is a startup's lifeblood.
Software startups must keep growing. Hardware startups must not let the ship date slip.
06 / 20
When momentum stalls, teams get demoralized, conflict rises, and the company starts to unravel.
Recovery usually comes from small wins, often in sales or product.
07 / 20
0.99 vs 1.01
A viral coefficient below 1 dies, above 1 grows exponentially.
A tiny improvement in the right place can decide the company.
08 / 20
Choosing a co-founder is one of the highest-stakes startup decisions.
The safest pattern is a co-founder you already know well.
09 / 20
9 of about 75
YC companies added a random co-founder between interview and batch start.
All 9 teams fell apart within a year.
10 / 20
A bad co-founder is worse than being a solo founder.
A strong co-founder is still better. Sam Altman estimated all top 20 YC companies had at least 2 founders.
11 / 20
YC's best founding team size is 2 or 3.
01
1 founder is usually weak.
02
2 founders works well.
03
3 founders works well.
04
4 founders sometimes works.
05
5 founders is usually bad.
12 / 20
Remote co-founding is a major early-stage disadvantage.
Fast communication and cycle time matter, and Sam Altman could not name a top software startup with separated co-founders.
13 / 20
The best early co-founders share these traits.
01
They are resourceful.
02
They are tough.
03
They stay calm and decisive.
04
They are creative under pressure.
14 / 20
For software startups, a technical founder is usually essential.
Non-technical founders who plan only to hire and manage engineers usually fail.
15 / 20
Early hiring should be slow and selective.
One mediocre employee in the first 5 to 10 hires can poison culture or kill the company.
16 / 20
5 months
Airbnb spent this long interviewing its first employee.
It hired only 2 people in its first year and wrote cultural values before hiring anyone.
17 / 20
Hire early only for desperate need.
01
More employees raise burn.
02
More employees add complexity.
03
More employees create tension.
04
More employees slow decisions.
18 / 20
When you are actively hiring, recruiting becomes a top founder priority.
Great candidates have many options and may take a year to convince.
19 / 20
20 / 20
Full talk: https://www.youtube.com/watch?v=CVfnkM44Urs
Lecture 3: Before the Startup
Startups are so counterintuitive that your instincts actively mislead you. Paul Graham argues you don’t need to know anything about startups, you need to know about your users, and that the way to have startup ideas is to become the kind of person who has them.
Lecture 3 - Before the Startup (Paul Graham)
youtube.com/@ycrootaccess
30 Sept 2014 · Startups
48 min watch → 95 sec
Paul Graham argues that startups reward user understanding and real product demand, not fundraising tricks or institutional game-playing. He says college is usually the wrong time to start one, and that the best preparation is to learn deeply and work on interesting problems.
Startup advice is often counterintuitive, so ordinary business instincts cause avoidable mistakes. Startup success comes mainly from understanding users and making something they want.
Gaming the system stops with users, because they only care whether the product solves their problem. College is usually the wrong time to start, unless a side project is already taking off.
2 / 20
The best startup skill is understanding users and making something they want.
Fundraising tricks and startup mechanics matter less than product demand.
3 / 20
Users cannot be gamed for long.
Investors can be fooled for a round or two, but users only care whether the product solves their problem.
4 / 20
Young founders often play house.
1
They raise money.
2
They rent an office.
3
They hire friends.
4
They neglect making something people want.
5 / 20
Fundraising is not startup success.
The best investor pitch is fast growth you can show.
6 / 20
Growth hacks are mostly a distraction.
Durable growth comes from making something users love, then telling users about it.
7 / 20
A founder with strong user love raises money more easily than a founder with a flat usage graph.
Product demand matters more than fundraising expertise.
8 / 20
Trust your instincts about people more than your instincts about startups.
Impressive people who trigger misgivings often become serious problems later.
9 / 20
Choose cofounders like close friends.
1
Work with people you genuinely like.
2
Work with people you respect.
3
Know how they act when interests conflict.
10 / 20
Starting a startup changes you unpredictably.
Early confidence or insecurity says little about who becomes a strong founder.
11 / 20
If you doubt you can start a startup, the only real test is to try.
Do not run that test while you are still in college.
12 / 20
Good startup ideas rarely come from trying to invent startup ideas directly.
Brainstorming usually produces plausible-sounding bad ideas.
13 / 20
Many strong startup ideas began as side projects.
1
Yahoo began that way.
2
Google began that way.
3
Facebook began that way.
4
Apple began that way.
14 / 20
A reliable way to find ideas is to live in the future.
Get to a technology's leading edge, then notice what seems obvious to you but missing to others.
15 / 20
Domain expertise matters more than startup expertise.
Larry Page knew search deeply. Brian Chesky and Joe Gebbia brought design and execution to Airbnb.
16 / 20
A non-technical founder helps most through domain expertise, operations, or sales.
In a pure technology startup, someone still has to build the product.
17 / 20
Business school is usually poor startup preparation.
It trains managers for large companies, not product builders in early-stage startups.
18 / 20
The best preparation is to learn powerful things, work on interesting problems, and do it with people you like and respect.
That is how startup ideas and cofounders usually emerge.
19 / 20
20 / 20
Full talk: https://www.youtube.com/watch?v=ii1jcLg-eIQ
Lecture 4: Building Product, Talking to Users, and Growing
Talk to users before you build, not after. Adora Cheung breaks down how to find your first users, run the build-measure-learn loop, and separate the metrics that matter from the ones that flatter you.
Lecture 4 - Building Product, Talking to Users, and Growing (Adora Cheung)
youtube.com/@ycrootaccess
2 Oct 2014 · Startups
52 min watch → 85 sec
A practical framework for going from zero users to growth by defining the problem, immersing in the industry, building a real MVP, talking to users, measuring retention, and choosing sustainable growth channels.
Early startup work needs long, uninterrupted blocks, because compressed focus improves product judgment. A startup should start with a one-sentence problem, founders who care, and direct user verification.
Before coding, learn the industry, pick a narrow first customer, and storyboard the experience. An MVP solves the problem with the smallest feature set. Early growth comes from users, retention, and manual iteration.
02 / 20
Building in secret often creates a launch spike, then immediate user loss.
Products built without real feedback usually miss the problem.
03 / 20
Validate the startup problem before coding.
01
Write one sentence.
02
Pick a problem you care about.
03
Verify other people have it.
04 / 20
Learn the target industry by doing real work.
01
Do the job yourself.
02
Study competitors deeply.
03
Read S-1s, financials, earnings calls.
05 / 20
Doing the work exposes the inefficiency worth attacking.
Homejoy came from seeing booking and cleaner scheduling break inside house cleaning.
06 / 20
A narrow first customer segment makes product decisions easier.
Serve teenage girls or soccer moms first, not everyone at once.
07 / 20
Storyboard the full user experience before writing code.
Map discovery, use, and post-use feedback as one flow.
08 / 20
An MVP, minimum viable product, is the smallest feature set that solves the user's problem.
Anything extra slows learning.
09 / 20
Get your place cleaned for $20 an hour.
Homejoy positioning
10 / 20
Product positioning should fit in one sentence.
Users ignore long explanations, but a clear functional benefit lands fast.
11 / 20
Get first users from direct sources.
01
Founders and friends.
02
Local communities.
03
Online communities.
04
In-person outreach where users already are.
12 / 20
Early growth often requires unscalable tactics.
Street fairs or intercepting people already solving the problem can get zero to a few users.
13 / 20
Direct conversations beat surveys for product feedback.
Surveys overrepresent people who love or hate the product, not the middle.
14 / 20
Retention is the main macro metric for product quality.
Retention means the share of users who keep coming back over time.
15 / 20
Reviews and NPS can signal product quality before enough retention data exists.
NPS means Net Promoter Score.
16 / 20
Manual processes should come before automation.
Homejoy accepted only 3% to 5% of cleaner applicants, then automated the steps that predicted quality.
17 / 20
There are three sustainable growth types.
01
Sticky growth comes from retention.
02
Viral growth comes from referrals.
03
Paid growth comes from acquisition spend.
18 / 20
Paid growth works only when CLV exceeds CAC and payback is safe.
CLV is customer lifetime value. CAC is customer acquisition cost.
19 / 20
Read the full source
Lecture 4 - Building Product, Talking to Users, and Growing (Adora Cheung)
youtube.com
YC Root Access
20 / 20
Full talk: https://www.youtube.com/watch?v=yP176MBG9Tk
Lecture 5: Competition is for Losers
Competition is for losers: if you’re in a great business, you’re the only one doing what you do. Peter Thiel argues that monopolies drive progress and that every successful startup escapes competition rather than winning it.
Lecture 5 - Competition is for Losers (Peter Thiel)
youtube.com/@ycrootaccess
7 Oct 2014 · Startups
50 min watch → 3 min
Peter Thiel argues that startups should avoid competition by building monopoly-like businesses. He explains how value creation differs from value capture, why small markets are the best starting point, and why durable market control matters more than growth alone.
Value Creation vs. Value Capture
A valuable business creates X dollars of value for customers and captures Y percent of that value as profit, where X and Y are independent variables
Competition destroys profit by bidding away created value, while a monopoly, meaning a business with dominant control of a market, keeps high margins and accumulates capital
02 / 17
Revenue vs. Profit
$195B vs. $50B
US airlines produced $195 billion in 2012 revenue but stayed marginally profitable, while Google produced just over $50 billion yet had far higher market value
Creating a large market is not the same as capturing value from it
03 / 17
Businesses misdescribe their markets to hide reality: weak businesses define the market absurdly narrowly to sound unique, while monopolies define it very broadly to hide dominance
A restaurant claiming to be the only British restaurant in Palo Alto faces real competition from nearby locations and cuisines
04 / 17
Google held about 66 percent market share in search, a clear monopoly position, but could describe itself as part of global advertising or the broader technology market
This broad framing made its dominant position look small and competitive, avoiding regulatory scrutiny
05 / 17
Startups should begin by dominating a very small market, then expand outward in concentric circles, because a large market on day one usually means heavy competition and weak differentiation
06 / 17
Successful monopoly entry strategies started with tiny, defensible markets
01
Amazon began as an online bookstore, not all retail
02
eBay began with niche collectibles like Pez dispensers
03
PayPal first spread among about 20,000 eBay power sellers
04
Facebook started with about 10,000 Harvard students and reached 60 percent penetration in 10 days
07 / 17
Clean tech startups in 2005 to 2008 often pitched themselves as serving trillion-dollar energy markets
Entering a huge market as a small player usually means many competitors and little ability to control pricing or distribution
08 / 17
Strong monopoly businesses usually combine four traits
01
Proprietary technology that is at least 10 times better on an important dimension
02
Network effects where the product gets better as more people use it
03
Economies of scale from high fixed costs and low marginal costs
04
Brand that reinforces the monopoly position
09 / 17
Software businesses are especially suited to monopoly because software has near-zero marginal cost, can spread quickly, and can combine all four monopoly traits in one product
10 / 17
PayPal in March 2001
75%
After 27 months in business, about three quarters of PayPal's value came from cash flows in 2011 and beyond, not near-term results
Long-term defensibility matters more than current momentum for tech companies
11 / 17
The most valuable companies are not just first movers but last movers, because most of a tech company's value comes from cash flows many years in the future
Durability and the ability to generate cash a decade out matter more than short-term growth
12 / 17
Innovation often creates enormous value for society without making inventors rich, because scientists usually capture 0 percent of the value they create
Many competitive industries like railroads, airlines, and disk drives improved rapidly while profits were competed away
13 / 17
One historical way to capture value from innovation is a vertically integrated monopoly, meaning one company controls many linked parts of the system
Ford and Standard Oil did this, as did Tesla and SpaceX later by integrating suppliers and distribution more tightly than rivals
14 / 17
Competition can become a psychological trap because people treat crowded pursuits as proof of value
Copying what many others are doing often leads to ferocious rivalry over small differences and small rewards
15 / 17
Competition can improve performance on the metric being contested, but it can also stop people from asking whether the contest itself is worth winning
Peter Thiel
16 / 17
17 / 17
Full talk: https://www.youtube.com/watch?v=5_0dVHMpJlo
Lecture 6: Growth
Growth is the single metric that best captures whether a startup is working. Alex Schultz, who scaled Facebook’s growth, on retention as the foundation, why the magic moment matters, and how virality actually compounds.
Lecture 6 - Growth (Alex Schultz)
youtube.com/@ycrootaccess
9 Oct 2014 · Product
47 min watch → 95 sec
Growth comes from retention, not tactics alone. The key levers are a retention curve that flattens above zero, a clear North Star metric, a fast path to the product's magic moment, and disciplined use of channels like virality, SEO, and notifications.
Retention drives growth. A product whose retention curve flattens above zero has product-market fit, and a curve that falls to zero does not.
Do not build a separate growth team before product-market fit. One North Star metric and one magic moment align growth around durable value.
2 / 20
A retention curve shows what share of one signup cohort still uses the product over time.
If the curve flattens above zero, some users keep coming back.
3 / 20
10,000
customers can be enough to estimate a cohort retention curve.
Check each day what percentage are still monthly active.
4 / 20
Good retention depends on category.
E-commerce can work at 20% to 30% monthly retention. Social usually needs around 80% early retention.
5 / 20
Do not add virality, paid acquisition, or growth hiring before product-market fit.
If users do not stay, those tactics only accelerate churn.
6 / 20
A North Star metric is one company-wide measure of durable value.
Examples include monthly active users, sends, nights booked, and gross merchandise volume.
7 / 20
A North Star metric improves decisions by rewarding long-term activity or transactions.
It prevents teams from optimizing for registrations or reported revenue.
8 / 20
A magic moment is the first product experience that makes the value obvious.
Growth improves when more users reach that moment quickly.
9 / 20
Growth tactics should remove barriers to the magic moment.
1
Show Facebook users their friends quickly.
2
Show eBay users the collectible they want.
3
Show Airbnb users a compelling place to book.
10 / 20
20%
confirmed registrations fell after eBay changed affiliate payouts.
Activated confirmed registered users fell only about 5%, then accelerated.
11 / 20
The eBay affiliate change worked because affiliates sent searchers to relevant item pages.
Users saw what they wanted before being asked to register.
12 / 20
Virality has three parts.
1
Payload is how many people each share reaches.
2
Frequency is how often users send it.
3
Conversion rate is how often recipients sign up or act.
13 / 20
Hotmail grew because every email carried a signup link.
That created high frequency and high conversion, despite low payload per message.
14 / 20
A viral loop matters only after retention works.
Even K-factor above 1 fails if new users quickly churn.
15 / 20
12 hours
Facebook's community translation system took to translate French.
It eventually supported 104 languages, 70 translated by the community.
16 / 20
Keyword research comes before search engine optimization.
In the UK, users searched 'cocktail recipes', not 'cocktail making'. In the US, 'drink recipes'.
17 / 20
100x
SEO traffic increase after Facebook added a directory to public profiles.
Google could reach pages directly instead of through buried navigation.
18 / 20
Email, SMS, and push notifications work only when deliverable and relevant.
Triggered messages tied to meaningful events beat generic newsletters and protect channel access.
19 / 20
20 / 20
Full talk: https://www.youtube.com/watch?v=n_yHZ_vKjno
Lecture 7: How to Build Products Users Love
Users love products that meet an expectation you set, so under-promise on the parts you can’t control. Kevin Hale on how first impressions, tone, and small design choices decide whether people fall for a product.
Lecture 7 - How to Build Products Users Love (Kevin Hale)
youtube.com/@ycrootaccess
14 Oct 2014 · Product
48 min watch → 1.5 min
A product users love comes from strong first impressions, direct contact with users, and support handled by the people who build the software. The talk ties growth to conversion and churn, and shows how retention improves when teams reduce friction and stay customer intimate.
Startup growth is the gap between conversion, the share of visitors who become users, and churn, the share of users who leave.
Products users love come from support-driven development, customer intimacy, and retention habits that made Wufoo grow five years by word of mouth.
02 / 20
The values that win the first user and first dollar are the values that scale into a large company.
Early habits become company habits.
03 / 20
First impressions include the first email, first login, first support reply, and even error pages.
Each moment can become the story users repeat.
04 / 20
A lovable product needs two layers of quality.
01
Taken-for-granted quality means basic functionality works.
02
Enchanting quality means extra details make use pleasurable.
05 / 20
Small details can create emotional attachment.
01
Playful form copy matters.
02
Thoughtful tooltips matter.
03
Error pages matter.
04
Designed documentation matters.
06 / 20
For developer tools, documentation is part of the product experience.
Stripe improved onboarding by inserting each logged-in user's API credentials into code examples.
07 / 20
Distinctive launch ideas spread better than generic promotions.
Wufoo's battle-axe API contest drew 25-plus applications, including iPhone, Android, and WordPress projects.
08 / 20
85%
Divorce prediction accuracy from 15 minutes of conflict.
John Gottman also reached 94% from an hour-plus discussion of hopes and dreams.
09 / 20
Customer support reveals product relationship problems.
01
Pricing and payments create money fights.
02
Performance creates time fights.
03
Competition creates jealousy fights.
04
Partnerships create in-law fights.
10 / 20
Stonewalling, refusing to respond, is one of the worst support behaviors.
Unanswered users are likely to churn, especially early.
11 / 20
7 to 12 minutes
Wufoo's support response time, from 9:00 a.m. to 9:00 p.m.
Ten people supported 500,000 users and 5 million form respondents.
12 / 20
75.8%
Support requests completed the emotional-state field.
Messages became calmer, with fewer exclamation marks, curse words, and all caps.
13 / 20
Direct user exposure improves design quality.
Jared Spool says teams need two hours every six weeks. Wufoo's builders got 4 to 8 hours weekly through support.
14 / 20
A product's knowledge gap is the distance between what users know and what they must know to use it.
Close it by teaching more or requiring less knowledge.
15 / 20
30%
Of Wufoo engineering time went to internal tools, FAQs, tooltips, and better help flows.
One documentation redesign cut support 30% overnight.
16 / 20
Retention needs visible ongoing care.
Wufoo showed a personalized 'since you've been gone' feature list and wrote handwritten thank-you cards every Friday.
17 / 20
Customer intimacy means becoming the best overall solution by knowing users closely.
Small startups can pursue it without large budgets.
18 / 20
Wufoo grew five years by word of mouth, raised about $118,000, had 10 people at acquisition, and returned about 29,000% to investors.
It focused heavily on product experience and support.
19 / 20
Read the full source
Lecture 7 - How to Build Products Users Love (Kevin Hale)
youtube.com
YC Root Access
20 / 20
Full talk: https://www.youtube.com/watch?v=sz_LgBAGYyo
Lecture 8: How to Get Started, Doing Things that Don’t Scale
The startup things that work early are the ones that don’t scale: recruiting users by hand, doing manual work software will later do. Stanley Tang and Walker Williams on getting the first users and earning press.
Lecture 8 - How to Get Started, Doing Things that Don't Scale, Press
youtube.com/@ycrootaccess
16 Oct 2014 · Startups
52 min watch → 90 sec
The talk covers three early startup skills, testing demand with fast experiments, doing manual work that does not scale to find product-market fit, and using press only when it supports a clear business goal.
Early-stage startups should treat ideas as experiments, launch the simplest test fast, and let real customer behavior decide whether demand exists.
Product-market fit, a product that reliably satisfies a real market need, matters more than scalable systems. Press works only for a specific business goal, not status.
02 / 19
Doing things that do not scale is an early startup advantage.
Manual work wins first users, improves service, and reveals what to build.
03 / 19
150 to 200
small business owners DoorDash interviewed first.
Many said they lacked delivery infrastructure and turned down orders.
04 / 19
DoorDash tested demand with PaloAltoDelivery.com in one hour.
The site listed restaurant PDF menus and a personal cell number for orders.
05 / 19
Strong demand can show up through a bad product experience.
DoorDash kept getting calls although customers read PDF menus and ordered by phone.
06 / 19
Word of mouth can validate a painful problem.
DoorDash sent one dorm email, did almost no marketing, and still got referrals.
07 / 19
Manual work taught DoorDash what to build later.
01
Founders delivered orders themselves.
02
Founders handled customer support themselves.
03
Founders dispatched drivers manually.
04
Those tasks informed later driver assignment algorithms.
08 / 19
Cheap tools replaced infrastructure at the start.
01
Square processed payments.
02
Google Docs tracked orders.
03
Apple's Find My Friends located drivers.
09 / 19
Optimize for speed over elegance early.
Teespring's CTO duplicated the codebase and database in 3 to 4 days, not about a month.
10 / 19
Solve only the next order of magnitude.
Build for 100 users when you have 10, then 1,000 when you have 100.
11 / 19
There is usually no silver bullet for early user acquisition.
First users often come from 100 emails a day, calls, and personal networks.
12 / 19
Charging early customers is usually better than giving the product away free.
Free users can fake demand and value the product less than paying users.
13 / 19
Champions are customers who actively advocate for a product.
Founders create champions by giving memorable service and personally fixing bad experiences.
14 / 19
$130,000 to $140,000
monthly revenue before Teespring founders stopped doing support themselves.
Running support exposes the most important product problems.
15 / 19
One unhappy customer can erase momentum created by many happy customers.
Always make problems right, even when the fix is expensive.
16 / 19
Press is not a scalable growth engine.
It may help reach the first 100, 200, or 1,000 users, then stops working without genuinely new stories.
17 / 19
Getting press is a repeatable process.
01
Define a story people want to read.
02
Get warm introductions to relevant reporters.
03
Pitch clearly.
04
Follow up with launch details and assets.
18 / 19
Read the full source
Lecture 8 - How to Get Started, Doing Things that Don't Scale, Press
youtube.com
YC Root Access
19 / 19
Full talk: https://www.youtube.com/watch?v=oQOC-qy-GDY
Lecture 9: How to Raise Money
Investors decide on the founders and the story more than the spreadsheet. Marc Andreessen, Ron Conway, and Parker Conrad on what actually moves a fundraise and the mistakes that kill one.
Lecture 9 - How to Raise Money (Marc Andreessen, Ron Conway, Parker Conrad)
youtube.com/@ycrootaccess
21 Oct 2014 · Startups
50 min watch → 2.5 min
Marc Andreessen, Ron Conway, and Parker Conrad explain how startup fundraising works, what investors actually look for, how to match each round to risk reduction, and why investor quality matters more than maximizing valuation.
Venture capital is a game of outliers: about 4,000 venture-fundable companies seek funding each year, about 200 get funded by a top-tier firm, about 15 reach $100 million in revenue, and those 15 generate about 97% of the year's venture returns.
Investors back startups with an extreme strength that could make them an outlier, not startups that merely check every box. The strongest way to raise money is to make the business itself compelling.
02 / 20
Fundraising should match cash to risk reduction, each round should finance specific milestones that remove specific risks.
Product risk, launch risk, customer risk, and revenue risk are the types of risk each round should address.
03 / 20
Founders should choose investors for long-term trust, judgment, and useful networks.
The investor relationship can last 10 to 20 years and matters most when the company hits stress.
04 / 20
97%
of venture returns come from the 15 companies per year that reach $100 million in revenue
Out of 4,000 venture-fundable companies seeking funding each year, only about 15 become these outliers.
05 / 20
Investors back startups with an extreme strength that could make them an outlier, not startups that merely check every box without a remarkable advantage.
A single exceptional quality matters more than being good across all dimensions.
06 / 20
Seed investors often judge founders in the first minute on five qualities:
01
Leadership ability
02
Product obsession
03
Decisiveness and speed
04
Communication skill
05
Whether the product solves a personal problem the founder actually experienced
07 / 20
A founder should explain the product in one compelling sentence.
If an investor still does not understand what the company does after the first sentence, the pitch is already failing.
08 / 20
Procrastination hurts startups, because progress depends on making decisions quickly about hiring, firing, product, and execution.
Seed investors watch for decisiveness as a signal of whether the founder can execute.
09 / 20
Bootstrapping as long as possible can be an advantage.
Businesses that could survive without outside capital are often exactly the businesses investors most want to fund.
10 / 20
Raising venture capital is usually easier than the work that follows, such as recruiting engineers, selling to enterprise customers, or creating viral growth in a consumer product.
11 / 20
Founders should not ask investors to sign an NDA at the start.
Early fundraising depends on trust, and an NDA signals distrust before the relationship begins.
12 / 20
Founders should get investor commitments in writing immediately.
Investors may later forget that they committed, forget the valuation, or forget promises such as finding a co-investor.
13 / 20
At the seed stage, the right investors matter more than squeezing for the highest price.
Strong seed investors make trusted introductions that shape later rounds.
14 / 20
$9M
cap brought heavy demand for Parker Conrad's seed round, while $12M to $15M found little interest
The round came together in about a week once it dropped below the $10 million psychological threshold.
15 / 20
Founders should avoid selling too much of the company early.
Seed rounds often land around 10% to 15%, series A rounds often sell about 20% to 30%, and excessive dilution can demotivate the team and make later investors walk away.
16 / 20
For capital-intensive startups, fundraising requires unusually precise planning.
Each round must fund the exact milestones needed without creating too much dilution or forcing the company into bad later financing.
17 / 20
Debt and lease financing can supplement venture capital for capital-intensive companies, but they demand stronger operational discipline.
Loan covenants are easy to trip and can cost founders the company.
18 / 20
A good investor brings four qualities that matter beyond the check:
01
Domain expertise in the company's market
02
Introductions to customers and future hires
03
Ethics and integrity under stress
04
Staying power to support the company through multiple rounds
19 / 20
Read the full source
Lecture 9 - How to Raise Money (Marc Andreessen, Ron Conway, Parker Conrad)
youtube.com
YC Root Access
20 / 20
Full talk: https://www.youtube.com/watch?v=uFX95HahaUs
Lecture 10: Culture
Culture is what your team does when you’re not in the room, and it’s set by who you hire early. Brian Chesky and Alfred Lin on building the Airbnb culture and why founders should obsess over it from day one.
Lecture 10 - Culture (Brian Chesky, Alfred Lin)
youtube.com/@ycrootaccess
24 Oct 2014 · Business
50 min watch → 85 sec
A company culture should be designed around mission-specific values, reinforced through daily behavior, and used to hire, fire, and make hard decisions. Airbnb and Zappos show how early values shape long-term endurance, brand, and team performance.
Company culture is everyday values and actions that guide decisions, trust, and limits in pursuit of a mission.
Strong culture is a long-term advantage. Airbnb designed it early, hired slowly, and built mission-specific values like "champion the mission."
02 / 18
A practical culture test is simple. Values must help the company pursue its mission.
Values detached from the mission do not guide useful decisions.
03 / 18
Mission-specific values beat generic values.
01
Zappos used "deliver wow through service."
02
Zappos used "be humble."
03
Generic words like honesty and teamwork are too broad.
04 / 18
Useful core values take depth, not slogans.
01
Zappos started with 37 candidate values.
02
Zappos reduced them to about 10.
03
Employees gave input.
04
The process took a year.
05 / 18
Culture must become a daily habit, like fitness or customer service.
Companies that treat culture as an occasional initiative drift out of shape.
06 / 18
Patrick Lencioni's team-dysfunction pyramid fails in sequence.
01
Lack of trust blocks healthy conflict.
02
Lack of conflict blocks commitment.
03
Lack of commitment weakens accountability.
04
Weak accountability blocks results.
07 / 18
11.08%
Return from 1997 to 2003 for Fortune 100 Best Companies to Work For.
That was almost twice the stock-market returns of the S&P 500 and Russell 3000 shown.
08 / 18
Airbnb's founders became the company's initial DNA in 2008.
They worked like a tightly aligned family and spent nearly all their time together.
09 / 18
Founders are like parents, and the company is like a child.
Aligned founders create healthier culture. Dysfunctional founders create dysfunctional company behavior.
10 / 18
"Belong anywhere."
Airbnb's enduring mission.
11 / 18
Airbnb was not fundamentally about booking rooms. It was about helping people belong and come together.
That mission defined what the company was building.
12 / 18
A 2012 London riots story showed Airbnb could create friendship, not just transactions.
Host Sebastian said 7 former guests checked on him before his own mother did.
13 / 18
Airbnb hired for ability and values fit.
01
World-class ability was required.
02
Separate core-values interviewers sat outside the candidate's function.
03
Technical excellence could not override the culture decision.
14 / 18
Brian Chesky personally interviewed Airbnb's first 300 employees.
He asked whether a candidate would still take the job with 10 years left to live.
15 / 18
Airbnb rejected a Rocket Internet clone with 400 people and $90 million.
Airbnb had 40 employees and $7 million, but saw mercenary culture as the bigger risk.
16 / 18
Culture and brand are two sides of the same coin.
Internal beliefs show up externally, and feature-only companies become commodity utilities.
17 / 18
18 / 18
Full talk: https://www.youtube.com/watch?v=RfWgVWGEuGE
Lecture 11: Hiring and Culture, Part 2
Hire slowly and only people you’d want to work for. Patrick and John Collison of Stripe and Ben Silbermann of Pinterest on how they hired their first teams and protected culture as they grew.
Lecture 11 - Hiring and Culture, Part 2 (Patrick and John Collison, Ben Silbermann)
youtube.com/@ycrootaccess
28 Oct 2014 · Business
51 min watch → 2.5 min
Founders from Stripe and Pinterest explain how early hiring shapes company culture, how to spot undervalued talent, and how onboarding, transparency, and team structure must evolve as a startup scales.
Hiring and Culture
Culture is the behaviors a startup reinforces through who it hires, what work it does, what it shares, and what it celebrates
Early hires become multipliers who carry standards into many later hires as founders can shape fewer decisions directly
02 / 19
The first 10 hires should be chosen for traits that compound culture
Integrity, low ego, curiosity, ability to finish things, and unusually high care for product quality and details
03 / 19
Early startup recruiting works best by finding undervalued talent, often people early in their careers or overlooked by the market
Convince them with a mission that is ambitious, specific, and honest about difficulty
04 / 19
Culture is more like gardening than architecture
Founders plant initial values through early hires, then keep reinforcing what works while removing behaviors that do not fit
05 / 19
Pinterest's early hiring emphasized five traits in candidates
01
Worked hard and had high integrity
02
Had low ego and broad curiosity
03
Had deep excellence in at least one area
04
Included quirky candidates with unusual hobbies that signaled creativity
05
Could finish things and ship work
06 / 19
Stripe's founders asked whether a candidate was probably the best among their friends at that craft
A more practical test than the vague idea of a 10x employee
07 / 19
Early recruiting rarely comes from inbound applications
Founders often need to seek people out through friends of friends, Craigslist, tech talks, coffee meetings, barbecues, and users of the product
08 / 19
A startup is easier to recruit for when the mission is large and legible
Ambitious people are more attracted to a hard, important goal than to a narrow or overly niche project
09 / 19
Founders should calibrate what excellence looks like before interviewing, especially in functions they do not understand well
Ask proven experts which traits, questions, and backgrounds actually predict success
10 / 19
For the earliest hires, working with candidates on a real project for about a week can reveal far more than standard interviews
It is hard to fake skill, judgment, and collaboration for that long
11 / 19
References are most useful when they measure what a person is like to work with
Questions that force ranking, such as top 1%, top 5%, or top 10%, produce more honest signal than generic praise
12 / 19
Transparency can be a core cultural advantage when employees share a mission and have broad access to information
They can make better decisions without waiting for founders
13 / 19
Transparency does not scale by dumping all information on everyone
It requires tooling and norms so people can find what matters and so public criticism does not discourage sharing ideas
14 / 19
New hires become effective faster when they do real work immediately and get rapid feedback on both job performance and cultural fit
They also need to understand the long-term mission instead of only the task in front of them
15 / 19
As companies grow, onboarding must become a designed process with clear managers, team context, and company priorities
Follow up at 30 days to see whether the person and team are actually productive
16 / 19
Large organizations stay nimble by creating autonomous, cross-functional teams that own their resources and know their top priority
They can move without waiting on a monolithic central process
17 / 19
Early employees can grow into leadership roles, but companies need to let people try management without trapping them there
Some will become strong leaders and others will contribute more as individual experts
18 / 19
Read the full source
Lecture 11 - Hiring and Culture, Part 2 (Patrick and John Collison, Ben Silbermann)
youtube.com
YC Root Access
19 / 19
Full talk: https://www.youtube.com/watch?v=H8Dl8rZ6qwE
Lecture 12: Building for the Enterprise
Selling to enterprises is a different sport: long cycles, real buyers, and trust before features. Aaron Levie of Box on how to build a company when your customer is a company.
Lecture 12 - Building for the Enterprise (Aaron Levie)
youtube.com/@ycrootaccess
30 Oct 2014 · Business
46 min watch → 1.5 min
Enterprise software startups can win by using cloud, mobile, and user-led adoption to enter companies through a narrow product wedge, then expand with security, scale, and better design. The talk maps the market shifts and tactics that make this possible.
Cloud computing, mobile devices, and user-led adoption lowered the cost and friction of selling software to businesses.
Businesses spend about $3.7 trillion on enterprise IT yearly, versus $35 billion on mobile apps and $135 billion on digital advertising.
02 / 17
Cloud computing let startups deliver software online, instead of each customer buying servers and installing software in its own data center.
That cut deployment from months to immediate access for developers and employees.
03 / 17
Standardized cloud platforms let one startup serve a two-person company and General Electric, with 300,000 employees.
That expanded the market beyond only the largest enterprises.
04 / 17
In 2004 and 2005, online file sharing was painful because storage was expensive, browsers were weaker, and institutions often gave users tiny storage limits.
That created an opening for cloud file sharing.
05 / 17
Box started in 2005 as a consumer file-sharing product, then shifted to enterprise customers.
It had too many paid features for consumers and too little security and control for businesses.
06 / 17
Mobile devices shifted enterprise software from IT-led buying to user-led adoption inside companies.
Employees could introduce tools before central IT formally approved them.
07 / 17
2 billion
smartphones were already in use.
Nearly 3 billion internet users forced companies to support work across any device, location, and network.
08 / 17
Digital behavior reshaped every industry.
01
Retail needed omni-channel commerce software.
02
Healthcare needed personalized, distributed care software.
03
Media needed on-demand distribution and targeted marketing software.
09 / 17
A startup can enter the enterprise by letting individual users adopt the product first.
After usage spreads, sell the company on security, control, and scalability.
10 / 17
Enterprise startups should look for asymmetries, advantages incumbents cannot or will not match.
One example is platform-agnostic products when suite vendors prefer closed integration.
11 / 17
Business-model asymmetries can work too.
Zenefits gave HR software to startups and earned commissions from insurance companies instead of charging customers directly.
12 / 17
Early customers should be outliers at the edge of their industry.
Companies already operating in the future reveal new use cases first, including drones in construction, farming, and oil and gas.
13 / 17
Customer feedback should be translated into a simpler product vision, not copied literally.
Customers describe problems well, but often not the best product design.
14 / 17
Enterprise products should be modular, not heavily customized.
APIs, application programming interfaces, let other software connect without rewriting the core product.
15 / 17
Enterprise products spread faster when they follow three rules.
01
Solve one painful problem extremely well.
02
Use consumer-grade design that feels simple.
03
Use salespeople for guidance and deployment support.
16 / 17
17 / 17
Full talk: https://www.youtube.com/watch?v=tFVDjrvQJdw
Lecture 13: How to be a Great Founder
Great founders are defined more by relentless resourcefulness than by any single trait. Reid Hoffman of LinkedIn on the contrarian bets, speed, and networks that separate the founders who win.
Lecture 13 - How to be a Great Founder (Reid Hoffman)
youtube.com/@ycrootaccess
4 Nov 2014 · Startups
50 min watch → 1.5 min
Reid Hoffman argues that great founders are not all-purpose geniuses. They succeed by building strong co-founding teams, choosing the right networks and location, running on a clear investment thesis, and balancing persistence with adaptation.
Great founders are not superhuman at every skill. They have a few real strengths, learn quickly, and build teams that cover weaknesses.
Startups usually do better with 2–3 co-founders. Founders win by navigating paradoxes, running from an investment thesis, and reducing every risk except one core bet.
02 / 19
The image of a founder who is elite at everything is misleading.
No one is universally strong at product, strategy, management, and fundraising.
03 / 19
A startup usually has better odds with 2–3 co-founders than with a solo founder.
Trusted partners can cover each other's weaknesses across many required skills.
04 / 19
Co-founder choice should be based on trust and collaboration.
A messy founder breakup a year into the company is often fatal.
05 / 19
Great founders navigate both sides of each tension.
01
They are persistent and flexible.
02
They are visionary and data-driven.
03
They are internally and externally focused.
04
They do the work and delegate it.
06 / 19
An investment thesis is a clear explanation of why a startup should work.
It states what others believe, why you disagree, and what evidence changes confidence.
07 / 19
A contrarian startup idea is not just unusual.
Informed people disagree for real reasons, and the founder knows something they do not.
08 / 19
500,000 to 1 million
Users critics thought LinkedIn needed before a professional network became useful.
More than two-thirds of informed people Reid Hoffman asked expected failure.
09 / 19
LinkedIn grew because early users joined before full utility existed.
They believed the vision or wanted to experiment, which helped the network reach scale.
10 / 19
Persistence versus pivoting should follow evidence, not emotion.
Ask whether market evidence is increasing or decreasing confidence in the investment thesis.
11 / 19
$12 million
PayPal's burn in August 2000, with no revenue.
That lowered confidence and forced immediate problem-solving, not blind persistence.
12 / 19
PayPal reduced risk before launch by killing PalmPilot payments and switching to email payments.
Almost nobody at a dinner table had a PalmPilot.
13 / 19
Reduce every risk except one core uncertainty.
01
Choose a team that covers weaknesses.
02
Build where talent and customers fit.
03
Solve distribution before polishing features.
04
Keep financing alive long enough to learn.
14 / 19
Startup location should match the networks the company needs.
Founders should go where talent, customers, and operating know-how fit the problem.
15 / 19
Silicon Valley is strong for software, mobile, networks, and marketplaces, but not every startup belongs there.
Some businesses depend on stronger networks elsewhere.
16 / 19
Groupon fit Chicago because its early model depended on building massive sales forces.
A fashion startup should also follow fashion networks, not a generic startup hub.
17 / 19
Founders should be unbalanced while building a company.
Startups have many ways to die, so winning usually requires years of exceptional focus and effort.
18 / 19
19 / 19
Full talk: https://www.youtube.com/watch?v=dQ7ZvO5DpIw
Lecture 14: How to Operate
Founders should do the unscalable, high-leverage work themselves and delegate the rest with clear ownership. Keith Rabois on operating a company: barrels versus ammunition, and editing over creating.
Lecture 14 - How to Operate (Keith Rabois)
youtube.com/@ycrootaccess
6 Nov 2014 · Business
47 min watch → 80 sec
A company scales by maximizing output through simplification, selective delegation, strong operators, transparent metrics, and obsessive attention to details. The talk frames management as editing people, priorities, and systems into a high-performance machine.
Operating a company means maximizing output, not visible effort. Founders raise output by simplifying priorities, reallocating people, and fixing small details.
Teams scale through barrels, people who ship ideas end to end. Delegation depends on task-relevant maturity, not one fixed management style.
2 / 18
Early-stage company building should feel messy.
Too much process usually means innovation is too slow.
3 / 18
Execution improves when priorities shrink to 1 to 3 repeatable goals.
1
People cannot track complicated goals.
2
One top priority forces A+ work.
3
People otherwise default to easier B+ work.
4 / 18
Managers improve decisions by asking a few high-value questions.
Andy Grove estimated removing unnecessary steps improves performance 30% to 50%.
5 / 18
A founder should act like an editor.
The highest-leverage management work is cutting complexity and clarifying ambiguous thinking.
6 / 18
One company voice should reach products, recruiting, support, packaging, and public communication.
The whole company should feel created by one mind.
7 / 18
Delegation should match task-relevant maturity, experience with that exact task.
1
Low consequence, low conviction: fully delegate.
2
High consequence, high conviction: intervene and explain why.
3
Use different oversight for different people.
8 / 18
Use less red ink over time.
Leaders should make fewer corrections each month and quarter as standards are internalized.
9 / 18
Teams scale through barrels, not just ammunition.
A barrel ships an idea from conception through launch and brings others along.
10 / 18
A barrel usually reveals itself in two ways.
1
Keep expanding scope until it breaks.
2
Coworkers outside reporting lines seek that person for help.
11 / 18
Promote someone only if their learning curve matches the company's growth curve.
Otherwise, hire above them.
12 / 18
A founder should design the company dashboard personally.
The dashboard defines success and should be intuitive enough for daily use across the company.
13 / 18
Transparency should include broad access to operating information.
1
Company metrics.
2
Board-deck reviews.
3
Meeting notes.
14 / 18
Metrics need paired indicators.
One number alone can hide damage, like more hiring with worse quality.
15 / 18
The best metrics insight often comes from anomalies, not averages.
Unexpected behavior can reveal growth channels or hidden problems.
16 / 18
“
PayPal found 54 eBay power sellers had handwritten 'please pay me with PayPal' into listings.
That anomaly became a major growth channel.
17 / 18
18 / 18
Full talk: https://www.youtube.com/watch?v=6fQHLK1aIBs
Lecture 15: How to Manage
The hard thing about management is that there’s no formula for the hard things. Ben Horowitz on making layoffs, demotions, and the wartime decisions no playbook prepares you for.
Lecture 15 - How to Manage (Ben Horowitz)
youtube.com/@ycrootaccess
11 Nov 2014 · Business
50 min watch → 85 sec
Management decisions change company culture through how every employee interprets them. The lecture uses demotions, raises, stock options, and Toussaint Louverture's leadership to show why leaders must think beyond the room.
A CEO makes better critical decisions by predicting how each choice will be interpreted across the company, not just in the room.
Demotions, raises, and equity rules shape culture because people infer what the company rewards, tolerates, and considers fair.
02 / 18
A CEO should judge a decision by every audience it reaches, including employees who only hear about it later.
Interpretation changes behavior.
03 / 18
Demoting an underperforming executive can look unfair if that person keeps 1.5% equity while engineers hold 0.1% or 0.2%.
Equity signals status.
04 / 18
A demoted executive often loses authority, because coworkers may not follow yesterday's head of sales as today's lower-level manager.
05 / 18
An executive demotion teaches the company what failure means at the highest-paid level.
It answers whether effort preserves status, or results do.
06 / 18
Giving raises when employees ask teaches that pay goes to the strongest negotiator, not the strongest performer.
07 / 18
A fair raise process has 4 parts.
01
Reviews run every six months or quarterly.
02
Managers review actual work.
03
Coworkers give input.
04
No off-cycle exceptions.
08 / 18
A transparent compensation process reduces fear that pay depends on closeness, appearance, or social access.
Everyone knows the same method applies.
09 / 18
90 days
is when most startup stock options expire after an employee leaves.
Vested equity can still be lost.
10 / 18
$2.5 million
in cash can be needed to exercise options that appear worth $10 million after departure.
Wealthier employees keep upside more easily.
11 / 18
Before 2004, APB Opinion No. 25 made 10-year exercise windows risky for earnings, which could block an IPO or acquisition.
12 / 18
A 10-year exercise window is fairer to leavers, but it also rewards someone who leaves with options plus a new job.
Stayers keep only current equity.
13 / 18
Startup option policy is a cultural choice.
Either give leavers time to exercise, or say stock matters mainly if employees stay through success.
14 / 18
Toussaint Louverture built the only successful slave revolution in history by deciding from allies', enemies', and Haiti's future viewpoints.
15 / 18
Toussaint Louverture built Haiti's future culture through 4 choices.
01
He banned pillage and sexual misconduct.
02
He recruited the best defeated officers.
03
He let former owners keep land.
04
He required owners to pay workers.
16 / 18
Exceeded the United States
describes Haiti's export revenue under Toussaint Louverture.
Ben Horowitz presents this as moral and economic success.
17 / 18
18 / 18
Full talk: https://www.youtube.com/watch?v=uVhTvQXfibU
Lecture 16: How to Run a User Interview
A good user interview digs for what people actually do, not what they say they’d do. Emmett Shear of Twitch on the questions that surface real problems and the ones that only confirm your bias.
Lecture 16 - How to Run a User Interview (Emmett Shear)
youtube.com/@ycrootaccess
13 Nov 2014 · Product
46 min watch → 95 sec
User interviews work best when they focus on the right people, real behavior, and underlying problems instead of feature ideas. Examples from Twitch show how interviews with users, competitors, and non-users can shape product direction.
User interviews work best when they identify who a product is really for, because picking interviewees matters as much as asking good questions.
Early interviews should probe current behavior, pain points, goals, and context, not feature requests. Competitor users and non-users often reveal more growth opportunities than current users alone.
02 — 19
Founders can skip interviews only when they are the exact target users.
That usually creates a narrow side project, not a large company.
03 — 19
Kiko Calendar lacked product direction partly because its founders were not calendar users.
They also did not interview people who actually used calendars.
04 — 19
Justin.tv fit a 24/7 live reality show well.
It struggled beyond that use case because the team lacked insight from broader users.
05 — 19
The right interview pool can include more than users.
01
Students use the app.
02
School IT administrators choose tools.
03
Parents may pay.
06 — 19
Ask, "How do you do this today?"
Concrete tools, habits, collaborators, and review steps reveal whether the problem is real.
07 — 19
6 to 8
Interviews per similar user segment before new insights usually slow.
Sample different extremes, not one narrow group.
08 — 19
A weak signal is a workflow with no major blocker.
If nothing hurts, the problem may not justify a new product.
09 — 19
Do not validate features by asking, "Would you use this?"
Show a minimal version, like a browser extension instead of rebuilding the whole app.
10 — 19
$5
Payment is a stronger validation signal than verbal enthusiasm.
Handing over money is harder to fake than saying yes.
11 — 19
Analytics and on-site feedback show drop-offs in an existing flow.
They do not reliably reveal which problem a startup should solve next.
12 — 19
Twitch learned more from competitor users than from existing Justin.tv users.
Competitor users reported blockers like weak revenue sharing and unstable video in Europe.
13 — 19
Non-users exposed adoption barriers that blocked Twitch growth.
01
Computers were too slow.
02
Training lasted 12 hours.
03
Edited YouTube videos felt better.
04
Broadcasting practice risked revealing strategies.
14 — 19
Twitch removed adoption blockers instead of copying requested features like polls or scrolling text.
It bought computers, worked with broadcasting software companies, and integrated with Xbox and PlayStation 4.
15 — 19
The first people Twitch converted were often the same people it had interviewed.
Returning with solutions built trust and proved the company had listened.
16 — 19
Common interview mistakes distort product decisions.
01
Showing the product too early.
02
Asking about favorite features.
03
Interviewing whoever is easiest to reach.
17 — 19
Recording interviews keeps the interviewer engaged.
Recordings also give the company direct evidence for aligning product decisions over about 3 years.
18 — 19
Read the full source
Lecture 16 - How to Run a User Interview (Emmett Shear)
youtube.com
YC Root Access
19 — 19
Full talk: https://www.youtube.com/watch?v=qAws7eXItMk
Lecture 17: How to Design Hardware Products
Hardware punishes mistakes because you can’t ship a patch, so the constraints shape everything. Hosain Rahman of Jawbone on designing physical products and the supply chain behind them.
Lecture 17 - How to Design Hardware Products (Hosain Rahman)
youtube.com/@ycrootaccess
18 Nov 2014 · Product
47 min watch → 1.5 min
A hardware product should be built as a full system of device, software, and data, guided by a clear user problem and refined through staged validation, trade-offs, and iteration. Wearables are framed as context engines that can organize connected-device experiences.
Connected devices work best when designed around the user, because a 24/7 wearable can act as a context engine for thermostats, cars, and media systems.
Durable hardware needs full-stack execution, and products should be designed as one system across sensors, algorithms, apps, cloud, and integrations.
02 / 19
The central design tool is the why, a precise user problem people urgently need solved or cannot live without once they have it.
03 / 19
Jawbone's product process has six stages.
01
Exploration tests broad ideas.
02
Validation proves reality.
03
Concepting defines hero experiences.
04
Planning makes trade-offs.
05
Development refines details.
04 / 19
Exploration rewards tinkering and Demo Fridays.
An idea continues only if it merits another $50,000 of internal exploration.
05 / 19
Validation asks whether a concept is technically and commercially real.
Teams prove how it works, estimate cost, and check whether batteries or other enablers exist now.
06 / 19
Concepting uses a cross-disciplinary product experience team to define hero experiences, quality bar, and roadmap.
The team spans industrial design, software design, audio, writing, and graphics.
07 / 19
Planning and development turn concept into a launch plan through explicit trade-offs.
Teams balance features, battery life, manufacturability, schedule, and retail timing until value clears a minimum threshold.
08 / 19
Small hardware details can create disproportionate emotional impact.
Jambox startup sound, rubber feel, color, and Up sleep-graph animation could take months to resolve.
09 / 19
User research works best when it studies behavior and context, not product requests.
People rarely ask for a new category directly, but they reveal how they live, share, and use tools.
10 / 19
The Jambox created a new market by solving mobile listening as a shared, portable experience.
11 / 19
Wireless speakers
78%
of the overall speaker market by Christmas 2013.
The category was 0% in fall 2010.
12 / 19
Each release should be a stepping stone from today's technology toward a broader future vision.
Jawbone used an experience continuum to decide what ships now versus later.
13 / 19
Up used a three-part framework.
01
Track, sensors collect data.
02
Understand, software adds context.
03
Act, the system prompts behavior when timing matters.
14 / 19
Raw health data is useless without context.
A heart rate of 75 means little unless the product explains why it matters and what action to take.
15 / 19
Cross-functional pods help one team own a feature across the whole system.
Each pod combines product experience, hardware, software, and data expertise.
16 / 19
Constraints improve hardware design by forcing simplification.
Storyboards turn those constraints into concrete experiences for design and engineering teams.
17 / 19
As hardware teams scale, communication becomes the main organizational challenge.
Regular cross-functional reviews let sourcing, manufacturing, firmware, materials, sensors, and design solve system conflicts together.
18 / 19
Read the full source
Lecture 17 - How to Design Hardware Products (Hosain Rahman)
youtube.com
YC Root Access
19 / 19
Full talk: https://www.youtube.com/watch?v=F4K_qVlYQkg
Lecture 18: Legal and Accounting Basics for Startups
Cheap mistakes in incorporation, equity, and bookkeeping become expensive years later. Kirsty Nathoo and Carolynn Levy of Y Combinator on the legal and financial setup founders get wrong.
Lecture 18 - Legal and Accounting Basics for Startups (Kirsty Nathoo, Carolynn Levy)
youtube.com/@ycrootaccess
20 Nov 2014 · Startups
48 min watch → 1.5 min
Startup legal and accounting basics include forming a Delaware corporation, splitting and vesting founder equity, handling SAFEs and dilution, paying payroll correctly, assigning IP, and keeping records organized for fundraising and hiring.
Startups should usually form as Delaware C corporations and keep records complete, because investors expect standard documents and one bad conversion created $500,000 in legal bills.
Founder stock should usually split roughly equally and vest over 4 years with a 1-year cliff. Founders and employees must be treated as workers of a separate company.
02 / 19
A corporation is a separate legal entity that protects founders from personal liability.
If the company is sued, claimants usually pursue company assets, not founders' bank accounts.
03 / 19
Forming a Delaware corporation requires more than filing.
01
Adopt bylaws.
02
Appoint directors and officers.
03
Assign inventions and code.
04
Make the company own its IP.
04 / 19
Founders must separate personal actions from company actions.
The company, not the founder personally, must own assets and sign agreements.
05 / 19
Founder ownership is not created by conversation alone.
Each founder signs a stock purchase agreement and buys shares with cash or contributed intellectual property.
06 / 19
An 83(b) election is a tax filing for restricted stock that vests over time.
Missing it, or losing proof, can damage taxes and kill an investment deal.
07 / 19
Founder stock should usually vest over 4 years with a 1-year cliff.
Vesting stops someone who leaves early from keeping a large ownership stake.
08 / 19
0%
A founder keeps 0% if they leave before 1 year under a 4-year vesting schedule with a 1-year cliff.
After 1 year and 1 day, the founder keeps 25%.
09 / 19
Founder equity should usually be split roughly equally.
YC saw zero top companies with significantly disproportionate founder splits.
10 / 19
Early fundraising often uses SAFEs or convertible notes, not priced stock rounds.
Investors give money now for stock later, when a priced round sets valuation.
11 / 19
4x
A $100,000 SAFE on a $5 million cap gets about four times as many shares as $100,000 invested in a $20 million priced round.
A valuation cap rewards early risk.
12 / 19
45%
Founders can give up about 45% before growth if $2 million of SAFEs convert at a $6 million cap and new investors buy 20%.
Low caps plus large SAFE totals create heavy dilution.
13 / 19
Fundraising and acquisitions often demand these records.
01
Incorporation papers.
02
Stock purchase agreements.
03
Signed 83(b) elections.
04
Financing documents.
05
IP assignment records.
14 / 19
Startup money should pay business expenses only.
Payroll, rent, hosting, and customer acquisition qualify. Personal spending misuses company funds and is not tax deductible.
15 / 19
Workers must be handled like real employees of a separate company.
01
Pay wages and payroll taxes.
02
Classify employee or contractor correctly.
03
Issue W-2s or 1099s correctly.
04
Assign IP to the company.
05
Follow employment rules.
16 / 19
Employees must be paid at least minimum wage.
In San Francisco, the lecture cites about $2,000 a month, plus workers' compensation insurance and US work authorization proof.
17 / 19
When firing someone, act immediately.
01
Communicate clearly.
02
Pay wages and accrued vacation immediately.
03
Cut off physical access.
04
Cut off digital access.
05
Repurchase unvested shares right away.
18 / 19
Read the full source
Lecture 18 - Legal and Accounting Basics for Startups (Kirsty Nathoo, Carolynn Levy)
youtube.com
YC Root Access
19 / 19
Full talk: https://www.youtube.com/watch?v=EHzvmyMJEK4
Lecture 19: Sales and Marketing, How to Talk to Investors
Early sales is the founder’s job, and it’s a learnable process, not a personality trait. Tyler Bosmeny on running that process, plus YC partners on pitching investors clearly.
Lecture 19 - Sales and Marketing; How to Talk to Investors (Tyler Bosmeny; YC Partners)
youtube.com/@ycrootaccess
2 Dec 2014 · Startups
49 min watch → 1.5 min
The talk covers how founders should handle early sales and fundraising themselves, from prospecting and discovery calls to closing contracts and avoiding free trials. It also lays out a short investor pitch structure and tactics for timing and running a fundraise.
In an early-stage startup, founders are the first salespeople.
Talking to users is selling, and product passion plus industry knowledge often matter more than formal sales experience.
02 — 20
Early startup sales is a funnel.
01
Prospect for people who will take a call.
02
Run discovery conversations about their problem.
03
Close with contracts and payment.
03 — 20
400+
companies Tyler Bosmeny contacted in Clever's first two months at Y Combinator.
Early sales is a numbers game because few companies buy from a startup with no users or revenue.
04 — 20
Prospecting
2.5%
of the market are innovators, the earliest buyers on Everett Rogers's technology adoption curve.
Prospecting means finding the few people who will even consider a startup.
05 — 20
Three effective prospecting channels are:
01
A founder's personal network.
02
Industry conferences where target buyers gather.
03
Concise cold emails to the right person.
06 — 20
Conference selling works when founders email attendees before the event and schedule meetings in advance.
Go where users already gather, and use the attendee list to fill the trip.
07 — 20
A cold email works better when it is brief and asks directly for a time to talk.
Say who you are and what you are building, without explaining everything.
08 — 20
The most effective sales calls are mostly listening, not pitching.
Let the customer talk about 70% of the time to reveal the real problem.
09 — 20
A strong sales conversation starts with questions like:
01
Why did you take this call?
02
How do you solve this today?
03
What would an ideal solution look like?
10 — 20
Relentless follow-up matters because interested buyers often ignore multiple emails and calls.
Push every lead toward a clear yes or no, because maybe consumes time without progress.
11 — 20
Closing often comes down to redlining, the back-and-forth editing of a contract.
Do not spend months arguing minor legal clauses when the business needs revenue and reference customers.
12 — 20
If a prospect wants one more feature first, treat it as a pass unless they will sign contingent on that feature.
One-off feature requests rarely become real sales.
13 — 20
Free trials are a trap for early startups.
Use an annual agreement with a 30- or 60-day opt-out instead of unpaid work without commitment or revenue.
14 — 20
Pricing and sales model have to fit each other.
A low-priced product cannot support repeated flights, many demos, and months of contract negotiation.
15 — 20
A 30-second investor pitch should say what the company does, market size, and traction.
Use plain language a parent could understand.
16 — 20
A 2-minute investor pitch adds unique insight, business model, team, and a specific fundraising ask.
If the product has not launched, show speed with concrete milestones.
17 — 20
Raise money when leverage is highest, ideally after traction appears.
If traction is limited, show a committed team, fast execution, and a launch plan that does not depend on investor money.
18 — 20
Investor meetings work best through warm introductions and parallel scheduling in the same week.
One founder should handle fundraising full-time so the rest of the company keeps building.
19 — 20
Read the full source
Lecture 19 - Sales and Marketing; How to Talk to Investors (Tyler Bosmeny; YC Partners)
youtube.com
YC Root Access
20 — 20
Full talk: https://www.youtube.com/watch?v=SHAh6WKBgiE
Lecture 20: Later-stage Advice
The problems that nearly kill a company change as it grows, and later-stage founders have to keep evolving. Sam Altman closes the course with what shifts once a startup is working.
Lecture 20 - Later-stage Advice (Sam Altman)
youtube.com/@ycrootaccess
4 Dec 2014 · Startups
48 min watch → 1.5 min
Later-stage startup success depends on shifting from product building to company building, with simple management, stronger alignment, better HR systems, and founder discipline around focus, burnout, and scaling decisions.
Most later-stage startup problems can wait until product-market fit, when a product strongly satisfies a market, and usually turn urgent around 25 employees or 12 to 24 months.
Around 25 employees, the founder’s job shifts from building a great product to building a great company, and alignment plus self-management become decisive.
02 — 20
A flat startup works until about 20 to 25 employees.
By 30 employees, everyone reporting to the founder often collapses into coordination failure.
03 — 20
A scaling startup needs one simple rule.
Every employee has exactly one manager, and every manager knows their direct reports.
04 — 20
Around 25 employees, founders must become company builders.
Many fail by not adapting to hiring, delegation, management, and organizational design.
05 — 20
Waiting too long to hire senior executives is a common founder mistake.
Experienced leaders can take over major functions, and founders usually wish they hired earlier.
06 — 20
Hero mode burns managers out.
Accept 2 or 3 weeks of slippage, then hire ahead of growth instead of absorbing everything personally.
07 — 20
Delegation works only when employees own decisions.
If smart people only gather data for the founder, the founder becomes the bottleneck.
08 — 20
Alignment drives productivity as headcount grows.
01
Repeat clear goals.
02
Repeat the roadmap.
03
Repeat company values.
04
Keep weekly and monthly communication rhythms.
05
Make top priorities identical across employees.
09 — 20
Square of headcount
Productivity tends to fall at this rate without systems.
Each new pair of people adds communication overhead.
10 — 20
A useful communication rhythm has 3 parts.
01
Weekly meeting for the CEO’s direct reports.
02
All-hands at least once a month.
03
Quarterly plan tied to annual goals.
11 — 20
Informal memory stops working as coordination load rises.
Founders need a personal system for long-term goals, daily priorities, follow-ups, and each person’s work.
12 — 20
Written operating rules and cultural values scale better than oral tradition.
Writing them down becomes high leverage by 100 or 1,000 employees.
13 — 20
Good HR around 25 to 45 employees means structure, not bureaucracy.
01
Clear reporting structure.
02
Career paths.
03
Frequent performance feedback.
04
Compensation bands.
14 — 20
Compensation bands prevent fairness crises.
Defined ranges for mid-level and senior engineers reduce chaotic one-off salary negotiations.
15 — 20
Employees should keep getting substantial equity, including refresher grants.
Founders should expect to grant roughly 3 to 5% of the company per year for about 10 years.
16 — 20
An option management system should replace Excel once the company scales.
Stock administration mistakes have cost companies and employees tens of millions of dollars.
17 — 20
Two people risks appear early in scaling.
01
At 50 employees, new HR compliance rules begin.
02
Diversity needs attention before later-stage issues.
03
A monoculture from the first 15 or 20 hires is hard to change.
18 — 20
Founders need self-management for a long journey.
Avoid acquisition talks unless you would truly sell for a relatively low number, because distraction kills momentum.
19 — 20
20 — 20
Full talk: https://www.youtube.com/watch?v=59ZQ-rf6iIc
The whole course, in one sitting
That’s all twenty lectures. Roughly sixteen hours of talks from the founders and investors behind Airbnb, Stripe, Facebook, and Y Combinator, scanned in about twenty-four minutes.
The through-line is consistent across every speaker: build something a small number of people love, get the team and the idea right before you worry about growth, and treat sales, hiring, and fundraising as learnable processes rather than talents you either have or don’t. None of it is a shortcut. It’s the set of principles the best founders keep coming back to.
Save the carousels you want to revisit, and when a lecture earns your full attention, the link to the original talk is under each one. The point was never to replace the course. It’s to let you decide, in an afternoon, which parts deserve the full sixteen hours.