Founder Depression

If you ask a founder how her startup is going, the answer is almost always some version of “Great!”

There is a huge amount of pressure as a founder to never show weakness and to be the cheerleader in all internal and external situations.  The world can be falling down around you—and most of the time when you’re running a company, it is—and you have to be the strong, confident, and optimistic.  Failing is terrifying, and so is looking stupid.

Founders end up with a lot of weight on their shoulders—their employees and their families, their customers, their investors, etc.  Founders usually feel a responsibility to make everyone happy, even though interests are often opposed.  And it’s lonely in a way that’s difficult to explain, even with a cofounder (one of the things that works about organizations like Y Combinator is that you have a peer group you can lean on for support).

So a lot of founders end up pretty depressed at one point or another, and they generally don’t talk to anyone about it.  Often companies don’t survive these dark times. 

Failing sucks—there is no way to sugarcoat that.  But startups are not life-and-death matters—it’s just work.

Most of the founders I know have had seriously dark times, and usually felt like there was no one they could turn to.  For whatever it’s worth, you’re not alone, and you shouldn’t be ashamed.

You’ll be surprised how much better you feel just by talking to people about the struggles you’re facing instead of saying “we’re crushing it”.  You’ll also be surprised how much you find other founders are willing to listen.

Bitcoin Price Pressure

The price of bitcoin keeps trending down.  [1] This causes a lot of people to declare the end of bitcoin.

It’s important to understand that the default price pressure of the bitcoin ecosystem is down.  There are a lot of reasons for this, so I’ll just give a few examples. 

When most merchants take bitcoin for a purchase, they immediately sell for dollars (they can’t usually pay their vendors in bitcoin, they can’t pay tax in bitcoin, they don’t want to be exposed to the volatility, etc.)

When miners mine bitcoin, they have to sell some of them to pay for the electricity in dollars (or, more likely, RMB).  As the difficulty goes up and the price of bitcoin goes down, they have to sell a larger and larger percentage of what they mine.  In fact, as far as I can tell, mining is currently unprofitable with any reasonable cost of electricity.  It still makes sense to mine if you’re living in a dorm and don’t pay for electricity, or if you can’t pass a KYC check and are willing to pay a big premium to get bitcoin.

The recent US and international tax changes and tax filings also caused sell pressure.  This will continue to be the case until the US government takes bitcoin for taxes.  If I ran the government, I would never do this—the US needs the dollar to remain the world reserve currency.

There are plenty of other reasons, but the point is that bitcoin is different from many other financial markets, where the default pressure is up. [2]

This doesn’t mean bitcoin is doomed.  It just means that for it to succeed, we’ll need significant external buy pressure.  As I wrote awhile ago, I think the key thing we need for this is people actually using bitcoin for transactions instead of speculation (and merchants willing to hold bitcoin balances). [3] Unfortunately, transaction volume still appears to be trending down. [4] Anecdotally, I hear from merchants who start taking bitcoin that after an initial spike they see almost no volume.

The other way to get enough buy pressure would be if many people started deciding they want to hold bitcoin as a hedge or a speculation.  This spurs occasional bubbles, but we haven’t yet seen it work long term.

A declining bitcoin price does not mean bitcoin is failing.  And even if bitcoin itself fails, I think the blockchain will be one of key technical innovations of this time period.  In fact, in the most recent set of YC interviews, we saw more great blockchain companies than bitcoin companies.  Maybe Ripple or Dogecoin ends up winning. [5] Or maybe something that hasn’t been invented yet.

But it’s hard to imagine a future where the blockchain concept isn’t really important.



[2] Paul Buchheit first made this point:



[5] Speaking of dogecoin, as of this writing, it is the number 3 crytocurrency measured by daily volume, and if the market cap went up $10MM (about 25%), it would be the number 3 by market cap also.

Employee Equity

Startup employees often do not get treated very well when it comes to stock compensation.   New ideas float around occasionally, but lawyers are usually averse to trying new things, and investors don’t feel that they have enough incentive to try something new for employees.

There are four major problems: 

1) Employees usually don’t get enough stock. 

2) If an employee leaves the company, he or she often can’t afford to exercise and pay taxes on their options. 

3) Employee options sometimes get unfavorable tax treatment.

4) Employees usually don’t have enough information about the stock or options. 

Here are some proposed solutions:

1) Startups should give employees more stock.  Value is created over many, many years.  Founders certainly deserve a huge premium for starting the earliest, but probably not 100 or 200x what employee number 5 gets.  Additionally, companies can now get more done with less people.

It’s very difficult to put precise numbers on this because the specifics of every situation matter so much.  Perhaps the best way to think about it is to try to come up with a total compensation package with the same expected value (using the company valuation of the last round, or a best-efforts guess if it’s been a long time since the round) as the employee would get at a big company like Google.  As an extremely rough stab at actual numbers, I think a company ought to be giving at least 10% in total to the first 10 employees, 5% to the next 20, and 5% to the next 50.  In practice, the optimal numbers may be much higher.

One problem is that startups try to have very small option pools after their A rounds, because the dilution only comes from the founders and not the investors in most A-round term sheets.  The right thing to do would be to increase the size of the option pool post-A round, but unfortunately this rarely happens—no one wants to dilute themselves more, and this leads to short-sighted stinginess much of the time.

Option pools are complete fiction; boards can increase them whenever they want.  It should never be used as a reason for not making a grant.

2) Most employees only have 90 days after they leave a job to exercise their options.  Unfortunately, this requires money to cover the strike price and the tax bill due for the year of exercise (which is calculated on the difference between the strike and the current FMV).  This is often more cash than an employee has, and so the employee often has to choose between walking away from vested options he or she can’t afford to exercise, or being locked into staying at the company.  It’s a particularly bad situation when an employee gets terminated.

This doesn’t seem fair.  The best solution I have heard is from Adam D’Angelo at Quora.  The idea is to grant options that are exercisable for 10 years from the grant date, which should cover nearly all cases (i.e. the company will probably either go public, get acquired, or die in that time frame, and so either the employee will have the liquidity to exercise or it won’t matter.)  There are some tricky issues around this—for example, the options will automatically convert from ISOs to NSOs 3 months after employment terminates (if applicable) but it’s still far better than just losing the assets.  I think this is a policy all startups should adopt.

As an aside, some companies now write in a repurchase right on vested shares at the current common price when an employee leaves.  It’s fine if the company wants to offer to repurchase the shares, but it’s horrible for the company to be able to demand this.

3) Tax treatment on ISOs sounds good, but there’s an issue with AMT (Alternative Minimum Tax) and employees often end up paying more tax than they were expecting.   NSO gains are taxed as ordinary income.  RSU gains are also taxed as ordinary income.

Tax optimization is a second-order issue, and for an immediate solution, I think extending exercise windows to 10 years is the most important thing to do.  But longer-term, we should figure out a way for employees to be taxed on their stock compensation the same way as founders (whether or not capital gains should be taxed less than ordinary income is a separate discussion, but in any case I think the tax treatment should be the same). 

I think this may be doable.  Ideally, employees would just get restricted stock (not RSUs), and then when they sold it’d be taxed as long-term capital gains.  The problem is that as the company grows, the stock has a non-trivial present value, and if an employee were granted stock then they would then owe immediate tax on the value of the grant.

I think there are a lot of ways to fix this.  The easiest would be if the IRS would agree to not tax illiquid private stock until it gets sold, and then tax the gain from the basis as long-term capital gains and the original value as ordinary income.

Another might be to create a new class of employee stock.  Today, in an early-stage company, common shares are usually worth much less than preferred shares.  It might be possible to create a class of shares with less rights than common and thus worth even less.  The idea would be to convert these shares into common on an acquisition or IPO, but before that, they would be non-transferable and have no value.  If it were possible to create a class of stock that the IRS agreed had next to zero value, it might be possible to grant employees this sort of stock, have them owe a tiny bit of tax on it now, and then have normal long-term capital gains treatment years later when the startup goes public. 

4) Most startups do a bad job of helping employees think about the value of their options.  At a minimum, any startup should tell a prospective employee what percentage of the company the equity grant represents (number of shares is meaningless).  Some startups are very hesitant to do this—they don’t want to disclose the number of shares outstanding.  Employees should demand to know what percentage of the fully-diluted shares their stock options represent, and be very suspect of any startup that won’t tell them.

A specific question worth asking is some version of “so if I have 0.5% of company and it gets acquired tomorrow for $100 million dollars, will I get $500,000?”  There are many ways for this not to be the  case—there can be a huge liquidation preference, for example—that most employees don’t know to think about.  So it’s worth asking about a specific scenario.

While you’re asking questions, another good one to ask is “how much money did the company lose last month, and how much is in the bank?”  This is better than asking how much runway the company has, because most founders calculate that off of a plan that assumes revenue growth which does not always materialize.


I have two other thoughts about stock-based compensation at startups.

First, I think employee stock and options should usually not be transferrable.  It causes considerable problems for companies when employees sell their stock or options, or pledge them against a loan, or design any other transaction where they agree to potentially let someone else have their shares or proceeds from their shares in the future in exchange for money today.

I think it’s fair that if founders sell stock, they should offer an opportunity to employees that have been at the company for more than a certain number of years to sell some portion of their shares.  And some companies offer an employee liquidity program even when the founders don’t sell any shares themselves.  But otherwise, I think it’s reasonable for employees to wait for an acquisition or IPO.

Second, I think it’s time to consider other vesting ideas.  The standard at startups is 4 years with a 1-year cliff.  So you get 25% of your options after you’ve been there for a year, and 62.5% if you leave after 2.5 years.

It’s possible that 4 years is now too short—companies are often worth more than they were 10 years ago, but they take longer to reach liquidity.  I’ve seen some startups offer 5 or 6 year vesting schedules.  To compensate for this, they offer above-market grants.

Another structure I’ve seen is back-weighted vesting.  For example, 10% of the grant vests after the first year, and then 20%, 30%, 40% in the following years.  Again, the startups I know that do this tie it to above-market grants, and I think it helps them select for employees that really believe in the company and want to be there for a long time. (Also, companies that have vesting schedules like this usually do it for founders too.)

Finally, a third structure I’ve seen is a new way of thinking about refresher grants.  For a company using options, it’s nice to grant employees options early while the strike is low.  It’s possible to give “forward-dated grants”—i.e., you can give a high-performing employee a refresher grant today where 1/3 of it starts vesting immediately and the other 2/3 starts vesting when their initial grant is fully vested.  This guarantees them a low-strike price and presumably a relatively large grant in a few years.  Dustin Moskovitz at Asana does something like this, and I think it makes a lot of sense.

These are just a few of the ideas I’ve seen about new ideas for employee vesting.  But I think they’re worth considering—the default 4-year grant does not seem to be the best option.

The Worst Part of YC

We are going to send out YC summer 2014 interview decisions (both yes and no) before 10 pm PDT tonight.

The worst part of our job at YC is rejecting companies.  It leaves me feeling down for many days after our application process.  I experienced plenty of rejection from investors while running my own startup, and I remember well how bad it is.  Starting a startup is such a hard thing that we wish we could help everyone doing so.

Some day, we hope to be able to fund almost all the good startups.  In the meantime, we still have capacity constraints as we figure out how to scale, and that means we’re stuck saying no to potentially good companies.

The best startups often look bad at this stage, and we make mistakes.  We could easily miss something great.  If you’re working on something that users love, you like working on it, and you have a plan for how to build a business around it, then please don’t let us deter you.  (It's sometimes useful to ask friends for an outside perspective on how you're doing.)

Although we had nearly 20% more applications than a year ago, it was really striking how much higher the average quality of applications was for this batch compared to any previous batch.  Most of the partners independently mentioned this to me.

Anyway, to everyone we’re unable to fund: Best of luck, and don’t give up.  We love to see founders and companies reapply—companies that look bad now can look great with 6 months of progress.

Startups, Role Models, Risk, and Y Combinator

The YC application deadline is this Friday, and you can apply here.

People often tell us they think they want to start a company but just aren’t sure, so I thought I’d share some thoughts.  Although it’s true that most people aren’t well suited to start startups, a lot of people that could be great at it are afraid to make the leap.  They look at super successful founders who now seem impossibly impressive.

The first time I met the Airbnb founders they were clearly smart and fairly impressive, but nothing like what they are today.  We met at a coffeeshop in Mountain View, and they were stumbling over their words and talking about how things weren’t going that well.  Now they are taking over the world.  This improvement is not a special case—the same thing happened for the Collison brothers at Stripe, and the founders of Homejoy, Weebly, Coinbase, Teespring, Pebble, and on and on and on.

Here’s the secret: everyone starting a startup for the first time is scared, and everyone feels like a bit of an imposter.  Even the most successful founders doubt themselves and their startups many times in the early days.  But founders improve very quickly.

So when you’re thinking about whether or not you can start a startup, remember that you shouldn’t compare yourself to these people now.  They became much more impressive in the course of running of their startup, and so can you.

Starting a startup is very hard and very painful.  Success usually requires a level of determination and commitment for which most people don’t have a mental model.

For example, when Adora Cheung was starting Homejoy, she would work all day as a cleaner to learn the business, drive an hour back to Mountain View, stay up as late as she could coding, then drive back to San Francisco at ~3am to beat traffic, sleep in her car, and do it again.  She also gave both her apartment and her car to early cleaners so that they could partner up with Homejoy.  We don’t want to delude anyone about what running a startup is like—it’s a rational decision to decide you don’t want to start a startup.  

But there are lots of great reasons to start a company, and a lot of people are willing to accept the pain. The unfortunate situation is when people who want to start startups don’t actually get started—they feel like the great startup founders are too impressive, or they don’t know what to do, or it’s too risky.

It’s really not that risky—in general, few things are as risky as they seem.  And Y Combinator makes it even less risky—we don’t invest much money, but it’s enough to live on (even with a family in most cases).  If the startup doesn’t work out, one of the advantages of the alumni network is that most YC founders find something interesting to do next.

The only thing you have to know how to do is build something people want.  If you can do this, and you are sufficiently relentless, you can probably create more value and have more impact than you could in a regular job.  YC can teach you nearly everything else—in fact, most of what we do is give startups one-on-one advice.

Founders are usually amazed by how much they get done over the three months of YC, and how much they change.  The structure of YC helps startups focus on the few things that matter, and a group of people that mostly start out feeling like they don’t belong transform to some of the best founders out there.

If you’re still on the fence about applying for the Summer 2014 YC batch, we hope you’ll make the leap!  And don’t worry if you’re not as far along as you’d like of if your application isn’t polished enough.  We’ve gotten very good at looking past this, and also it doesn’t hurt you if you don’t get in the first time you apply (we rejected Drew Houston from Dropbox the first time he applied).  We fund companies at all stages, from just the faintest idea to post-Series B. 

Here is some advice other people have written about how to apply:

How to Apply to Y Combinator by Paul Graham

Last Minute Advice for YC Applicants by Garry Tan

Harj Taggar on Quora answering "What is the best advice for a startup applying to Y Combinator?"

Michelle Crosby on her YC experience

Drew Houston's Dropbox Application

Harry Zhang's (Lob YC S13) Advice for YC Applicants 

 Y Combinator Applicant Advice by Zain Shah

What I've Learned From Female Founders So Far

On the whole, I got a great response to my request for feedback about how YC could encourage female founders.  It's clear there are two separate problems: 

1) Some women already starting startups aren't interested in doing Y Combinator.

2) Some women who could be great founders don't start startups. 

I realize it's always a bit ridiculous for a guy to talk about what it's like for female founders, but I'm interested in doing whatever I can to help, because the venture business has definitely been unfair to women.  The women on our team also care deeply about this issue, and  can do more than I can to address it.

For point #1, one of the most consistent messages was that we need to make it clear that we care about the issue and want to fund more female founders.  So I'll say that now: we want to fund more women.  And we'll keep saying this in the outreach we do.

We want to fund more women because it's the right thing to do, but we're not doing this for diversity's sake alone.  We want to fund more women because we are greedy in the good way--we want to fund the most successful startups, and many of those are going to be founded by women.

Many are also going to be founded by people of different races, different religions, from different countries, straight, gay, in their 20s, or in their 50s.  All of those apply to people in the current YC batch.  In fact, they all apply to the YC partnership as well.  Again, we don't do this for the sake of diversity. We do it because we want to get the best people, whatever they're like.

In the current YC batch, 24% of the companies we funded have one or more female founders, and there will be a lot of companies out of those with the potential to serve as role models.  We hope that as the number of female YC alumni continues to rise, more women will feel YC is a place that supports and respects them.

Another message was that we should do more to make women feel welcome.  Many emails pointed out that our website shows nearly all men; we'll fix that.  We'll also continue to work with our most successful female founders to talk about their experiences and mentor women that could be future founders.  We'll continue to ask women to come speak at dinners. In this batch, two of my four favorite speakers were women (Adora Cheung and Julia Hartz).  And we're working on something to improve the quality of Hacker News comments.

A very common request was for us to have women in the interviews we do before funding companies.  In the last batch, we had a woman in 2 of the 3 interview tracks.  We now have more female YC partners, so for this upcoming batch, we'll have a woman in every track.

Nearly all women who emailed me suggested that we keep the exact same bar for women as for men (anything else wouldn't be fair to the incredible women we fund every batch), but many pointed out that women are often good in different ways and at different things than men--for example, that men and women express confidence differently--and that we should make sure our criteria catch that.

A specific issue that came up is a belief that we look for founders that look like Mark Zuckerberg.  Actually that meme began as a self-deprecating joke. We funded a guy once who looked like Mark but ended up doing badly, and when PG was asked by a reporter how to fool him, he said that apparently this was one way. His real point was that looking like Zuckerberg means nothing--that you can look remarkably like him and still fail miserably.  I think it's more accurate to say we look for founders that have some of the qualities that have made Zuckerberg so successful.

Finally, I heard a lot of support for events like the Female Founders Conference and a belief that they could help change the industry.  And if YC continues to fund more women, many people believe VCs will follow.

For point #2, I think we can do a lot to reach young women earlier and help teach them about startups and coding.  Many women pointed out that you don't have to be a coder to be a founder.  That's definitely true, and it was a good reminder for me personally.  But I think it's good to at least present learning to code as an option worth considering.

As we do more events, we'll continue to reach out to women.  Kat, our director of outreach, Jessica, our founding partner, and I will all specifically work on this.  For example, we're thinking about holding a hackathon later this year.  It'd be great to have a lot of women attend.

We're also going to ask some of our successful female founders to do more outreach.  I believe we have already funded at least one female founder/CEO who will produce a multibillion dollar company. She and others are outstanding role models.

There's lots of work still to do, but we're on it.  I hope other investors will join us.

New RFS -- Breakthrough Technologies

We’d like for Y Combinator to fund more breakthrough technology companies—companies that solve an important problem, have a very long time horizon, and are based on an underlying technological or scientific breakthrough.  Not many people try to start these companies, so starting a company that will require a huge amount of time and money is an automatic competitive advantage.  SpaceX and Tesla are great examples of what is possible.

It used to be the case that governments funded a lot of development of breakthrough technologies.  The bad news is that they have mostly stopped; the good news is that the leverage of technology is such that now small startups can do what used to take the resources of nations. [1]

We think the YC model works well for these companies.  We invest with infinite time horizon and are not afraid of risky-looking companies. [2] We understand software, which will be central to many of these companies.  We are good at getting companies to focus on solving real problems for real customers, and not just developing technology for its own sake.  And our model helps companies figure out a right-sized initial project achievable with a small amount of time and money—great companies get built with a series of small wins that compound over time, and early momentum is critical.  A common failure mode of many ambitious companies is to bite off an initial project that is far too big and expensive.  Finally, we know a lot about raising money, which will be a big part of the challenge for many of these companies as they mature.

Here is a list (we’ll add to it over time) of some areas we’re particularly interested in, but more generally, we’ll pay attention to any area where technology can make the world much better.

Energy.  There is a remarkable correlation between the cost of energy and quality of life.  Throughout history, when the cost of energy has come down a lot (for example, with the steam engine) the quality of life goes up a lot.

Cheap energy would do a huge amount to reduce poverty.  New energy sources could also help the environment, the economy, reduce war, ensure a stable future, make food and water more abundant, and much more.

We believe economics will dominate—new sources must be cheaper than old ones, without subsidies, and be able to scale to global demand.

Nuclear energy can hit the bid, and possibly so can renewables.  But pricing is the first order question.

In addition to generation, we’re also interested in energy storage and transmission.  10x better batteries would enable great new things, as would the ability to easily move energy around. 

AI. Relative to the potential impact, it doesn’t seem like enough smart people are working on this.

A lot of smart people talk about AI with a combination of awe and fear, both for good reasons.  But it feels like it could be one of the dividing lines in the history of technology, where before and after look totally different.

Robotics. Robots will be a major way we get things done in the physical world.  Our definition is pretty broad—for example, we count a self-driving car as a robot.  Robots are how we’ll likely explore space and maybe even the human body.

Biotech. It’s still early, but it seems like we’re finally making real progress hacking biology.  There are so many directions this can go—fighting disease, slowing aging, merging humans and computers, downloading memories, genetic programming, etc.   We are certain that this is going to be a surprising, powerful and controversial field over the next several decades—it feels a little bit like microcomputers in the 1970s. 

Healthcare. Healthcare in the United States is badly broken.  We are getting close to spending 20% of our GDP on healthcare; this is unsustainable.

We’re interested in ways to make healthcare better for less money, not in companies that are able to exploit the system by overcharging.  We’re especially interested in preventative healthcare, as this is probably the highest-leverage way to improve health.  Sensors and data are interesting in lots of different areas, but especially for healthcare.

Food and water. At some point, we are going to have problems with food and water availability.  Technology can almost certainly improve this.  Great innovations are possible—we will need another advancement on the scale of what Norman Borlaug did. 

Education. If we can fix education, we can eventually do everything else on this list.  The first attempts to use technology to fix education have focused on using the Internet to distribute traditional content to a wider audience.  This is good, but the Internet is a fundamentally different medium and capable of much more.

Solutions that combine the mass scale of technology with one-on-one in-person interaction are particularly interesting to us.

This may not require a “breakthrough technology” in the classical sense, but at a minimum it will require very new ways of doing things. 

Internet Infrastructure.  We can’t imagine life without the Internet.  We need to be sure it keeps working—this includes everything from security to free and open communication to infrastructure.  The Internet is a transformative power, and we’re particularly interested in applications that transform the big underpinnings of society (bitcoin is a great example!).  The Internet lets people around the world coordinate action—there are almost certainly important businesses to be built around this concept.

Of particular interest to us are ways to use the Internet to fix government—for example, crowdfunding social services.

An important trend is the API-ification of everything.  As more and more businesses are accessible with a web API, the Internet becomes more and more powerful. 

Levers.  We’re interested in technology that multiplies the efforts and productivity of individuals.  Robots are a great example, but this also includes areas like new programming languages, powered exoskeletons, augmented reality, etc.

Science.  Science seems broken.  The current funding models are broken and favor political skill over scientific genius.  We need new business models for basic research.  There are a lot of areas where scientific developments can have huge commercial applications—materials, neuroscience, climate engineering, and cheaper/better ways to get to space, just to name a few—and we’d love to figure out a way for it to happen.  Bell Labs worked a long time ago but would probably not work in today’s world.

Transportation and housing.  About half of all energy is used on transportation, and people spend a huge amount of time unhappily commuting.  Face-to-face interaction is still really important; people still need to move around.  And housing continues to get more expensive, partially due to difficulties in transportation.  We’re interested in better ways for people to live somewhere nice, work together, and have easier commutes.

As a side note, you shouldn’t start a company just because it’s on this list.  Our hope is that someone already working on a company in one of these areas that might not have otherwise applied to YC will now consider it.  The great majority of the startups we fund will continue to be the sort of Internet and mobile companies we’ve funded in the past, so if that’s what you wanted to do before this post, keep doing it.  Traditional-looking startups like Google and Facebook are obviously as important as any company one could imagine, and clearly are breakthrough technologies. 

[1] To be clear, we are not interested in funding patent trolls.  We only want to fund businesses that actually solve problems and create value. 

[2] Related to long time horizons, if a company needs to raise a billion dollars of funding over the course of its life, that doesn’t scare us—in fact, that’s a plus.

The Founder Visa (again)

Nearly 5 years ago, Paul Graham first proposed the founder visa.  There has been a lot of discussion since, but nothing has happened. 

Maybe he was too ambitious in asking for 10,000 startup visas per year.  So here is a proposal for the US government: please let Y Combinator help allocate up to 100 visas to founders per year.  We’ll continue to take applications for funding from around the world, and work with whatever process you’d like—we just need to be able to get the founders visas quickly (None of the current paths works well enough for this, but a slight reworking of the O1 visa around criteria and timing could be sufficient.).  If the test works with us, you could expand it to other investment firms.  We’re happy to be the beta tester, and we’re confident we’ll prove that it’s a good idea.

100 visas a year is nothing.  But 50 new startups a year could be a huge deal. Many will fail, of course, but one could be the next Google, Facebook, Airbnb, or Dropbox. Though this is almost an immeasurably small number of visas, it could have a measurably large effect on the number of jobs created in the United States.

Startups are what the US is the best in the world at.  We figure out new businesses faster than anyone else.  It would be disastrous if that stopped being the case.

If founders from elsewhere want to pay taxes and create jobs in the US, we should let them.  Other countries are already encouraging this.  If you believe that intelligence and determination are evenly distributed, less than 5% of the best founders are born in the US.  But it’d be great if many of them started their companies here.

This is just a start. We are also in need of broad-based immigration reform, and I believe more immigrants will help our country.  But I also understand that the founder visa got tangled up with full-scale immigration reform, which may take a long time.  This is an easy way to have an immediate effect, and it’s good to move the ball down the field with small, incremental experiments.

Let us show you what we can do with 100 visas.  This will be measurable, and in 5 years, we can tell you exactly how many jobs get created.

Fundraising Mistakes Founders Make

There’s a lot written about what you should do when you raise money, but there hasn’t been as much written about the common mistakes founders make. Here is a list of mistakes I often see: 

• Over-optimizing the process
A lot of founders try to get way too fancy with tricks that they think will help them raise money.  It’s actually quite simple; if you have a good company, you will probably be able to raise money.  You’re better off working to make you company better than working on fundraising jiu jitsu.
The process is simple:
  1. Get intros to investors you want to talk to and reach out to them, in parallel, not in series - this is important, see (3).
  2. Explain to them why your company is likely to make them a lot of money. This usually includes the company’s mission, the product, current traction, future vision, the market, the competition, why you’re going to win, what the long-term competitive advantage will be, how you’re going to make money, and the team.
  3. Set up a competitive environment. You'll (unsurprisingly) get the best terms when multiple investors compete with one another for space in your round.  This is the one rule of "the game" that is really important--I'll talk about it more later on.
Some founders try things like carefully timing news articles, casually mentioning to one investor that they'll be having dinner with another investor, claiming their schedule is really packed except for one specific hour, and other tricks - but if you just build a good company, you generally won’t need to.
Many little things simply don't matter very much--for example, the "signal" sent when an early investor chooses not to participate in a later round. If the company is doing well stuff like this is easily overlooked, and if the company's not doing it will struggle to raise money anyway.
Unless you do it perfectly, game-playing will hurt you with most good investors. And you should be trustworthy and honest no matter what. Investors won't back you if they can't trust you.
• Over-optimizing the terms
Startups are usually a pass-fail course -- either you succeed or you don't.  If you fail, maybe you get acqui-hired, but that's happening less frequently and is usually little better than just getting a job at the acquiring company instead.
The important thing is to get good investors, clean terms, and not spend too much time fundraising. The biggest problem comes from chasing high valuations. Contrary to what many people think, at YC we encourage companies to seek out reasonable valuations. Valuations are something quantitative for founders to measure themselves on, and there are lots of investors willing to pay high prices, so they don’t always listen. But I’ll say it again: trying to get really high valuations is a mistake.

If you’re clearly in a position of leverage, it’s fine to push for a high valuation, but don’t jerk investors around. Just say what you want and don’t get into a lot of back and forth or term complexity. Also remember that very high valuations often push out good investors.
And don’t forget the prime directive of fundraising strategy: set things up so that you never do a down round. The badness of a down round is difficult to overstate; in fact, the threat of that is the best reason not to take a super high price when you’re offered one.  If you raise at such a price, everything has to go perfectly in order for your next round to be an up one.
• Failing to create a competitive environment
Ok, here is the one part of the game I really believe is critical.  You generally need to set up a competitive environment to get a good outcome in fundraising (or, for that matter, any big deal).
The hard part is getting the first offer. Once you have this, you have the leverage -- if other investors don’t act fast, you have an offer you can take, and they risk missing a potentially great opportunity (and maybe looking stupid to their partners, etc etc.) Until then, they can procrastinate and wait as long as they want. It’s remarkable how long it can take the first offer to come in, and how quickly the next ten can materialize.
So sometimes you have the hack the process a little bit to get this first offer. The best way is to find someone who loves what you’re doing and is willing to act. Although it’s ok to use that offer to get others, you should be nice to anyone willing to act first by prioritizing their offer, finding a way to get them into the round even if someone else leads it, etc.
There are a lot of other tactics for this that I should write a separate post on at some point.
Beware, though, that saying things like “our round is closing really fast” when you have no offers usually backfires. Investors talk and will call your bluff.
When you have a good competitive environment the leverage shifts to you - you will be astonished at how much things change. Firms that previously couldn’t meet you for three weeks will suddenly be able to schedule full partner meetings on a Sunday. And when multiple bidders really want to invest, a lot of the "non-negotiable" terms like 20% ownership and board seats go away.
• Coming across as arrogant, antagonistic, disrespectful, etc.
Somehow, a myth got started that investors like this and nerdy founders sometimes put on an affectation.  Don’t do it.  Be respectful (which includes things like not asking investors to make a decision after a first meeting unless you really are about to close your round).
Remember that investors are people too. They want to feel loved. The first time I raised money, I was hesitant to tell the investors I really liked that I really liked them because I thought I was giving up leverage. But it turns out telling the investors you really like that you especially want to work with them makes them more positively inclined to you, not less.
• Not hearing no
Investors don’t want to kill option value; founders are optimistic people.  This leads to investors saying a very nice version of "no" and founders hearing "with just a few more conversations, I may get to a yes."  Anything other than a term sheet is a "no", and all the reasons don’t matter.  Move on and talk to other investors.
• Not having a lead investor
A lot of founders put together party rounds comprised of dozens of investors and congratulate themselves that no single investor has much power over them.  But in practice investors have little power over companies that are doing well anyway, and what they actually have is no investor that is super invested in their success.
It turns out it’s really valuable to have one investor that you meet with every month and report progress to. This forcing function creates an operational cadence in the company that is a big net positive. It’s remarkable to me how much more frequently the party round companies go off into the weeds.
• Pitching poorly
A lot of founders get caught up in trying to follow a perfect template, and drone on and on about their competitors, the market evolution, etc.  They’re bored and it shows.
The way to pitch well is to focus on the parts of the business that truly excite you. That will shine through, and it will get the investors excited. Conveying your passion for the business is almost as important as what you say, and it’s almost impossible to fake.
Even if you’re an introvert, it will usually come through to a sophisticated investor. So start with the parts you’re really excited about.
Investors want to hear a good story, and that includes things like how you decided to work on this idea, why it matters, how you met your cofounders, etc. So don’t leave those parts out of the pitch.
Also, remember that smart investors are looking for the really big hits. So don’t do obviously dumb things like talk about potential acquirers in a seed round pitch - that will suggest you’re not trying to build a really big company.

• Not reference-checking major investors
Great investors can add a huge amount of value; bad investors can make your life miserable.  Before signing up to work with someone for the better part of a decade, spend an hour calling founders they have worked with to get a sense of what's in store for you.
• Lacking a clear vision
If you don’t seem to have any strong feelings or conviction, and you agree with every suggestion the investor makes about your business, you'll risk coming across as lacking a clear vision.  You should always listen to what someone smart has to say, but you should be firm on the things you really believe.
Founders with a clear vision can usually explain what they’re doing and why it matters in just a handful of words. Clear vision also usually entails at least one big new idea. Even if it’s a familiar problem, there should be something important the investor hasn’t heard before.
It’s ok to have some big unknowns, of course. You’re not expected to have all the answers, but you should have clear theses to start with.
• Not knowing key metrics
There are two questions I really look at in early stage investments:
  1. Does the team know what to do?
  2. Can the team do it?
The first question is addressed by the bullet point above.  The second is addressed by showing that the team cares about operational quality.  I’ve found that teams that execute well always know their numbers (or current status if in R+D mode) cold, and that it’s one of the best predictors of execution quality.  It’s surprising how many companies pitch investors without knowing this information.


Yesterday at lunch a friend asked me what tech trend he should pay attention to but was probably ignoring.

Without thinking much I said “artificial intelligence”, but having thought about that a bit more, I think it’s probably right.
To be clear, AI (under the common scientific definition) likely won’t work. You can say that about any new technology, and it’s a generally correct statement. But I think most people are far too pessimistic about its chances - AI has not worked for so long that it’s acquired a bad reputation. CS professors mention it with a smirk. Neural networks failed the first time around, the logic goes, and so they won’t work this time either.
But artificial general intelligence might work, and if it does, it will be the biggest development in technology ever.

I’d argue we’ve gotten closer in lots of specific domains - for example, computers are now better than humans at lots of impressive things like playing chess and flying airplanes. But rather than call these examples of AIs, we just say that they weren’t really that hard in the first place.  And to be fair, none of these really feel anything like a computer that can think like a human.
There are a number of private (or recently acquired) companies, plus some large public ones, that are making impressive progress towards artificial general intelligence, but the good ones are very secretive about it.

There are certainly some reasons to be optimistic. Andrew Ng, who worked or works on Google’s AI, has said that he believes learning comes from a single algorithm - the part of your brain that processes input from your ears is also capable of learning to process input from your eyes. If we can just figure out this one general-purpose algorithm, programs may be able to learn general-purpose things.
There have been promising early results published from this sort of work, but because the brain is such a complex system so dependent on emergent behavior it’s difficult to say how close to the goal we really are. We understand how individual neurons work pretty well, and it’s possible that’s all we need to know to model how intelligence works. But the emergent behavior of 100 billion of them working together on the same principles gets extraordinarily complex, and difficult to model in software. Or, as Nick Sivo says, "it's like reverse engineering the latest Intel processor with only the basic knowledge of how a transistor works."  It’s also possible that there’s some other phenomenon responsible for intelligence, and the people working on this are on the wrong track.

The biggest question for me is not about artificial intelligence, but instead about artificial consciousness, or creativity, or desire, or whatever you want to call it. I am quite confident that we’ll be able to make computer programs that perform specific complex tasks very well. But how do we make a computer program that decides what it wants to do? How do we make a computer decide to care on its own about learning to drive a car? Or write a novel?
It’s possible--probable, even--that this sort of creativity will be an emergent property of learning in some non-intuitive way. Something happened in the course of evolution to make the human brain different from the reptile brain, which is closer to a computer that plays pong. (I originally was going to say a computer that plays chess, but computers play chess with no intuition or instinct--they just search a gigantic solution space very quickly.)

And maybe we don't want to build machines that are concious in this sense.  The most positive outcome I can think of is one where computers get really good at doing, and humans get really good at thinking.  If we never figure out how to make computers creative, then there will be a very natural division of labor between man and machine.