Dialogue with the founder of Y Combinato: Discuss the factors of the company's success and failure

Act bravely and seize opportunities in time, otherwise you may miss the opportunity.

Compilation: Deep Tide TechFlow

*Note: This article is included in the Shenchao TechFlow topic "YC Entrepreneurship Course Chinese Notes" (updated daily), dedicated to collecting and sorting out the Chinese version of YC courses, and the third article is the online course "A Conversation" by YC founder Paul Graham with Paul Graham". *

Geoff Ralston：

Paul Graham started Y Combinator in 2005, Paul, before Sass was a thing in the mid-'90s, you thought people would want something to do with Sass. So, how did you come up with the idea for Yahoo to acquire Via dotcom?

Paul Graham：

In the mid-90s, it was widely believed that writing software was the same as writing software that ran on the client. For those who remember this period, client usually refers to the Windows operating system.

However we don't want to learn how to write software for Windows, we are only familiar with Unix. Therefore, we were desperate to find a way to write software without having to write software for the Windows platform.

We started exploring on our own, initially we tried emailing updates to change the site. But then we realized, if you can use the SMTP protocol, why can't you use the HTTP protocol to update the website? By running the software on the server and controlling it through a link in the browser, we finally found a solution.

The idea seemed so strange at first that we weren't even sure if it was possible. But we're still trying to write a clunky website builder that works by clicking a link. It's really rough, but it works well enough to build websites.

So far, we have successfully avoided the dilemma of needing to write software on the Windows platform.

Geoff Ralston：

Of course, sometimes innovative ideas do come from thinking differently. I mean, many websites now follow a similar workflow, so there is a lack of innovation. But if you can build a whole new website or product from scratch, that's truly innovative. Do you have such a plan?

Paul Graham：

This idea is very common, but in fact it is often the result of the analysis of historians, because they can see all the kinks in the development to make things seem obvious.

I always hate it when people describe startup ideas as light bulbs. Because not only is it the most clichéd metaphor, it's also wrong.

You don't just have the thought and realize it, right? It's actually more about you having this idea... that you're able to do something that no one else has done before, and it might be a bad idea, but you're too lazy to learn, so you decide to do it, right? Like Zuckerberg, started Facebook.

It's like we're going to see what happens, right? It's hard to believe that this will always be a startup.

Geoff Ralston：

It's good for everyone. Hope some of your ideas seem really implausible.

Paul Graham：

I love when an idea seems implausible. This has happened a lot in my hacking career, but it has become more common during my years with startups. I was talking to some people yesterday and they mentioned two things they could do.

One of them looks a little... naughty.

It doesn't mean to abuse anyone, but just to take advantage of something that seems impossible, right? For example, let the software run on the server, but the user feels that the software is running on his own computer, but in fact it only interacts with the user through the browser.

This kind of thinking is indeed a bit naughty.

But I recommend going with this idea because it seems too surreal and comical.

Geoff Ralston：

It seems like there are so many ways to discuss a great idea, but it always gets a bit complicated when you try to match the idea with reality...like, is the idea a bit naughty? Or is it less intuitive or less obvious? Also, it's hard to take any particular idea and say it has potential.

Paul Graham：

Actually, you don't know that's another thing. The future of startups is hard to predict, and it's an area of uncertainty. Even if an agency like Y Combinator does a good job of picking startups, maybe only 5 out of every batch of 150 startups will become giants or something like that.

What defines a giant depends on your point of view and how you define it. Anyway, even if you were able to pick perfectly, there is no guarantee that the batch size will be only five startups, you may need to pick 20.

Being a hacker is a good thing because business opportunities are hard to predict, but if you are a hacker, you can try it out just for fun.

Laziness does help sometimes, but this meme escapes into the wild and is considered a bad thing, probably better because being lazy is a good thing people do even though they think being lazy is a bad thing.

Geoff Ralston：

However, sometimes for some reason to do a thing but choose to remain silent. Although we are used to calling it lazy, it is actually not quite that. We tend to avoid doing seemingly pointless things, which makes us more graceful, especially for those of us who naturally hate pointless actions. It was for this reason that in 1995 I experimented with every idea I had to build my own web business and did all the work myself. It was a scary situation because I didn't know what to do and I had to find other people to help, but I didn't know if they would be good partners.

Hearing about your experience reminds me of the time you slept on the floor of Robert's house. You met Trevor Blackwell and Robert Morris online, but how did you decide to bring them on as co-founders? What is the process like? When did you make up your mind to make them your partners?

Paul Graham：

They are very aggressive co-founders with great programming skills. Still, I recall that Robert wasn't always involved in the entire startup process. Whenever he participates in a board meeting, or we receive any takeover offer, the board must seriously consider whether to accept it. That meant we had to deal with some nasty takeover offers.

For example, "How about we're going to give you two million dollars in stock?" And Robert would often say, "Let's put it to a vote, and I have to be honest with you, at least we can stop this work." Still , we would still choose to accept these deals at the time. I had a deal with Robert that if he made $1 million online, I would reward him with a pair of earrings.

But Robert didn't like wearing earrings, so after the deal, Trevor Frog and I took him to Harvard Square to buy a pair of earrings, which resulted in Robert wearing a new pair of earrings. Now that there are pictures of Robert wearing the earrings online, it won't be too long before we need to make a point. While it never occurred to me to be specific about it, I think he looks fine with the earrings on.

The reason Robert was chosen as a partner was because he was my accomplice in everything, not because I was always number one and he was just my second-in-command. When he sets out to accomplish certain tasks, I'm also his accomplice, as I was in the 1988 crisis caused by the Internet worm incident.

Geoff Ralston：

Did you know that Robert Morris was famous all by himself?

Paul Graham：

He's the guy who invented buffer overflow. I remember when he told me the idea, I said to him, "Wow, what a cool idea, you should totally do this."

He was the first famous hacker to actually get into trouble, indicted in 1986, becoming the first convicted felon under the Computer Fraud and Abuse Act.

So, if you're looking for a co-founder, look at people who have been sued for their work. Funny how this seems to be common for the FBI and law enforcement, but not a big deal in day-to-day law enforcement.

Motivations like sex, drugs, money, and revenge are considered, but Robert does it out of curiosity and doesn't make that list. In fact, it's hard for them to figure out what's going on, even within the government. So, I chose Robert as a partner because I would do anything with him. We had a lot of plans and he was a really good programmer who could type as fast as he could.

Although the C language is verbose, he was still able to quickly edit the source code and recompile it to achieve the effect he wanted.

At Harvard, undergraduates could only have accounts on the Science Center's official undergraduate computing system, and Aiken Lab's computer science department had real computers. So when he wanted to set up his account on the appropriate machine, he chose to switch over to a separate user in front of a machine as superuser, and then switch back. He is very curious and wants to be able to set up the account on the correct machine.

Throughout the whole thing, however, the main character involved is Robert. Robert is an interesting character because he was fired for reconnecting Harvard with the Internet. When Robert was a college student, Harvard University used to be one of the early Internet nodes, but lost Internet connection due to bit decay. RTM (Robert) spends the semester getting Harvard reconnected to the internet. Although he underperformed academically and was even expelled for a year, we later viewed this experience as a recruiting technique.

We put up a poster on the Harvard campus asking if there were people who had been fired from a program because we wanted to hire them. Eventually, we found a programmer who was really good, and he made a lot of money because we were acquired that year. He's enjoying himself so much now that he probably won't be back until he's 25.

A month later, and we'd been at the company for a full month, Robert became a rebel. He said, “We’ve been working on this company for a full month and we’re not done yet.” I started to think what kind of territory are we getting into? Maybe we need more programmers.

So I asked, "Well, Robert, who was the smartest person you met in grad school?" He replied, "Trevor." I was surprised, "Really, Trevor?" Because Trevor didn't seem like a very smart guy.

But actually, Trevor was very smart, so we recruited him. He quickly joined our team and showed super efficient hacking skills. If Robert says he's the smartest, he's definitely the smartest.

However, this brings with it a profound lesson: when selecting employees or partners, find people you trust, and then find people they trust. Some qualities are more important than others, smart people can judge other smart people, but trustworthy people cannot judge other trustworthy people. In fact, trustworthy people are often fooled by untrustworthy people.

Like our architect Kate, she's incredibly trustworthy but always fooled by the dastardly schemers who take advantage of her. So if you want them to be trusted people to join your team as co-founders, you have to make that judgment a priority.

The entire network works like this, and there has to be someone who can tell if people are trustworthy or not. For me, it's not my forte. At YCombinator, Social Radar Jessica is in charge of this.

In the end, I also hired staff, which was done via the website.

Geoff Ralston：

What did you take away from hiring at your first startup? You know, if they don't have a great performance because they're working on a project, it seems understandable. It's a bit counterintuitive.

In YC, we often do things that are counterintuitive.

Paul Graham：

Startups in general are very, very counterintuitive. This is what YC is all about. If it were obvious what to do in the entrepreneurial process, we wouldn't have so much to teach, right? So it's our job to tell founders what they're ignoring. We tell them not to rush to hire people, but they tend to do it in a hurry and end up coming back and thinking, “Oh, I wish we had listened to your advice.” But really, all the guidance we give is is counterintuitive, not obvious.

Counterintuitive means it sounds wrong. So they do what they feel is wrong, and then hope to find and fix the problem early.

You know, it would make perfect sense to write an article about why startups are so counterintuitive. Then it's worth writing an article about.

Why is it so counterintuitive for startups? I have no idea.

While I could come up with some theories, I find the answers so complex and interesting that it's unlikely I'll be able to come up with a straightforward answer.

Geoff Ralston：

I mean, it's kind of counterintuitive that we're only talking about your writing.

When you start a business, you feel great, you write a ton about software that helps you grow, and you write an article telling people not to do things that don't scale.

Paul Graham：

Yes, I'm sure you've all read everything Paul wrote. However, there is a very important point here that applies particularly to the challenges most people face when starting a business early on. This idea is our motto at YC: "Don't start out doing too much, instead, do something disproportionate."

"Yeah, oh yeah, exactly." We might think we're doing something wrong or that our idea sucks, but it's not. We once did a software for an online store builder. You can use it to set up your own online store and start selling items. When we tell potential customers about our product, they usually answer, "Would you like to use our easy online store builder?" Most of the time, the answer is no. But they still need an online store, so we'll ask them, "Would you like to have an online store?" And they'll say, "Yes." Eventually we'd say, "OK, if we use our software to create an online store, you'll own it. Sounds great, right?"

While it may seem like a mess, it turns out to be a direct marketing method and we are also members of the Direct Marketing Association DMA. In every industry, a business usually has an internal name rather than a common external name. For example, the fast food industry is known as the fast casual service industry. The catalog business is known as the direct selling industry. So, we've also been a member of the DMA's catalog business and documented every catalog. Do you know the directory in the mail? We only need to write a letter to receive more catalogs. Now, we have a bookshelf filled with all the catalogs.

Geoff Ralston：

Yes, it looks like you're trying to build a new search engine. In this case, you need to set up your server, do all the searches manually, and present the results. Doing so will give you a better understanding of each link in the search process.

I'm thinking about what we've done with our mail service, which was also done manually initially.

We have a manual ad server and no real ad serving technology. We just wrote the code and added the ads to the server. It's not perfect, but it taught us an important lesson: if you think you've found a solution and built a lot of things for it, you're probably going to be wrong because you don't know all the details and problems yet.

Therefore, from this perspective, manually completing tasks can help us gain a deeper understanding of the entire process and find better solutions.

Paul Graham：

I will use our software to build other people's websites so I can get a better idea of how it performs. In fact, I'm the author of that website builder.

Therefore, I will use this software to think, but sometimes it is inconvenient. In the process of building a website for someone else, I also modify the software as needed, just like running server-based software. For example, instead of the CP command, I would use the ship command (that is, the mv command in Unix).

This approach allowed me to continue tinkering with the software and making it better while the site was being built. Therefore, when choosing a co-founder, you should choose those who you can trust and are most trustworthy. This can be helpful to many people, especially founders who are starting their own businesses on their own.

Geoff Ralston：

Why is it so difficult to be the sole founder?

Paul Graham：

There are a lot of different reasons, but I think the hardest part is morale. No one can keep morale high when things are bad, and even if you have a lot of them, they can't keep you happy all the time. Especially for start-ups, most companies either fail or succeed.

If you succeed, you will get rich.

But in the beginning, you only have 10 customers and want to grow at 20%. Next week you want another 10% increase. Growing 10% every week is an ambitious goal. But we just need to find one more customer. You can go out and find it yourself, right? So next week you have 11 customers, you have to have 1.1 customers, right?

This is basically a client, right? It doesn't matter how small the number is as long as you are steadfast and keep doing things and keep the growth rate good, because maintaining a steady growth rate means exponential growth, which also means your base will settle quickly .

Geoff Ralston：

What factors can make a company successful and what factors can cause it to fail? Perhaps we could explore issues such as how employees understand their role and performance in the company. On startup school turf, there are thousands of companies trying to figure out what to do next and whether they are headed in the right direction or already mired in failure.

Paul Graham：

The main reason companies fail is often poor founder execution. As a startup, we often talk to people who focus on competitors. One of the advantages of YC funding many companies is that they have very large data sets. Do you know how many companies have been beaten by competitors? How many times since 1900? I tell startups that basically they have the same protection from competitors as a light aircraft would from a collision while flying through the clouds. Do you know what it means to protect? There is a lot of space. For example, the small Cessna has no radar and is difficult to see in the clouds. It's like running a 100-meter race and suddenly you meet another competitor on the track.

So what should you do? Run as fast as you can, because if they are stronger than you, they win, otherwise you win. I wouldn't try to start a business solely on my own as a founder because I don't think that's possible. Therefore, you need an excellent and solid founding team that understands and cooperates with each other.

We can give two types of answers. For investors, we can tell them how to choose startups, but only a part is useful for entrepreneurs. We can tell investors to pick smart people, but entrepreneurs are just as smart as they are. If they're already smart, how can we help them get smarter? Actually, being smart doesn't matter. More important is determination.

Suppose a person has an IQ score of 100 and a determination score of 100. If you start losing your resolve, then pretty soon you'll be an incompetent smart person. However, if you team up with someone who is super determined, you can gradually get smarter. Eventually you might find a guy with lots of taxi medals and still be very rich, or a job like garbage hauling, but you can learn a lot of smart stuff from it.

Geoff Ralston：

It depends on the strong bond you have with that person, you have to be able to move forward with him. Otherwise, they leave.

But, I think there is another important factor, you have to be able to play a certain role in the team, because your idea can be bad, and it will only get worse. So you have to be able to think about which direction we should go.

You have to have some kind of creativity that enables you to make the right choices, and...

Paul Graham：

Actually, you probably don't need much creativity. If you pay enough attention to your users, you can serve their needs the way scientists seek truth. Ultimately, your product idea will grow without you thinking too much about it, which is the result of evolution. So maybe you're either like Steve Jobs and have an intuition about what customers want, or you just need to have the skills to talk to customers and understand their needs, which aren't always obvious.

Well, you know Steve Jobs was successful because he was also a customer and besides that, he had a deep understanding of user needs.

For example, he may think that telephone jacks are obsolete and that there is no longer a need to provide everyone with more telephone jacks.

Steve Jobs may have been successful simply because his aspirations were so generic that they applied to everyone. He lives in the future. But aside from reading The New York Times, he wasn't much of a fan of the Internet, so he never really got his hands on the field.

Determination is extremely important in the business world, but that doesn't mean it's the same in the early days of a startup. The mistake you can make is not giving the user enough attention. Sometimes, an idea pops into your head, which is what you call a vision.

You spend a lot of time alone in coffee shops thinking and building your vision, and writing something complex and elaborate. However, if you're not communicating with potential customers, you're likely to run into sales problems, and that's why many are put off. It's better to find anyone in need who is willing to pay to solve the problem and help you find more people like it.

If you have a problem, all the better, but you need to get your product or service to market quickly, or you'll get awkward feedback for not getting it done in time. You may feel shame and discomfort when users tell you that your product or service is not as good as expected, but only in this way can you continue to improve and improve. Therefore, in the following process, you should actively keep in touch with users in order to continuously improve your product or service.

Geoff Ralston：

I remember when I joined YCombinator, they came out with Imagine K12, YC's version of edtech. When I was talking to Paul, he asked, "How is the Imagine K12?" Since he had helped us a lot, I replied, "It's going well"...

Then he asked me, "When are you going to launch the product?" I replied, "We're working on a PR plan, finishing software development, so maybe around February." But he suggested that I just launch it, why not start now Woolen cloth? I hesitated, but was moved by his words.

In the end, we launched the product within a week.

This is the mistake you can make, waiting too long can embarrass you. You have to act bravely and seize opportunities in time, or you may miss the opportunity.

Paul Graham：

We launched the YC website in less than a week, and of course we didn't have any software at first, just an ASCII form that people would fill out and email us back. Applications for the first two batches were submitted by email. We print out those emails and pass them on to our partners, and they grade the form, and that's our application process.

Geoff Ralston：

There used to be talk of doing something that wasn't scalable, and now that we're trying to do that with thousands of applications, it's going to be much harder. Therefore, we need to build scalable software systems, which is what we have been trying to do.

Paul Graham：

Ultimately, when you know what software to build, you build it. While it's fun to spend a lot of time telling how YC got started, I don't think that's the most important thing.

Speaker A：

As a co-founder, how do you deal with different levels of commitment? What to do if you have to fire a co-founder

Paul Graham：

The way to deal with different levels of commitment is to ask yourself the question: would I rather have 30% of someone's commitment, or 100% of someone else's commitment? Personally, I'd much rather have 10% of Robert Morris's brain than nearly everyone else's.

So it was a relatively simple decision.

One way to deal with this is to make sure that someone has a higher percentage of stock than someone else. Usually, the proportion of shares is equal, but there may be situations where it is not exactly equal, for example, if one person holds 51% of the shares and another holds 49% of the shares, the former can control the company.

Speaker B：

How many people do I need to launch a private beta of a product release to?

Paul Graham：

Some people might think that releasing a private beta is good enough without releasing the information to the public. For situations like this, I'd personally lean toward just doing a private test with a handful of people, as that's closer to a final full release than sitting in a coffee shop thinking about your thoughts.

Speaker C：

What do you think of the current fundraising and project structure? Are some projects over-funded, or basically under-funded to achieve project plans?

Paul Graham：

I don't know anything about cryptocurrencies, but I've heard that there are tons of cryptocurrencies in circulation. So it's usually a good idea for a startup to try to get funding if there's money flowing. Other than that, I don't know much. But one of the things we see a lot at YCombinator is that raising too much money isn't always the right thing to do.

Even if you don't have to raise a lot of money, there is a chance that this will happen because the money sitting there has some sort of gravitational effect that makes you spend it quickly. This can lead you to some stupid decisions. For example, there are companies that have raised $1 billion in ICOs, but I'm not sure they'll ever actually be able to become a legitimate company. Or you might get a bunch of LPs, raise $1 billion, and then you hire more people than you actually need. It's usually not fair, and that's what differentiates ICOs from other approaches.

Of course, it's unwise to have a billion dollars and hardly any ideas. It's like some Magic Jump company raised $2.3 billion and seemed to have nothing.

Speaker D：

What's your best tip for a great user interview?

Paul Graham：

The best technique is to understand not only what they think, but why they might be wrong. During the conversation, ask them what is missing in their life, and then start making hypotheses, asking what it would be like if they could do this or that. In this way, a deeper understanding of their true intentions and needs can be gained. Finally, the authors recommend discussing this issue with others.

Speaker E：

You said launch early, right? How do you justify the financial risk and turmoil of launching a bad product?

Paul Graham：

There are risks in delaying a launch, but there are risks in launching earlier. Therefore, there must be a rule of thumb or method for when to release. This method describes the characteristics of the minimum viable product, but it is a useful measure. Once your product reaches a certain level of utility, as long as one person is happy with your product and able to do something they couldn't do before, then you can launch the launch. If you launch something that doesn't generate interest and response, then you're not ready to launch. According to Paul Butchert, if 10 people really like your product, that's a good start. You only need these 10 people to be passionate, and you can start publishing. The attitude of others doesn't matter.

Speaker F：

Can you talk about how you can tell the difference between building what they want and what the client needs? Most of the time, needs can be short-sighted, and real needs can be what people want, but they have some unintended consequences. So how do you go about establishing what customers really need and what they really want?

Paul Graham：

Ideally, we want to build something that people will choose to use, buy, or become our users, but in reality sometimes people buy what they don't really need, and that's the difference between a need and a want.

For example, people want healthy food, but they often buy unhealthy food. As entrepreneurs, we need to choose between meeting people's needs and desires. We hope that the products we make can not only meet the needs, but also be attractive. However, we cannot be too greedy and only cater to market demand at the expense of our customers' health. At the same time, we cannot impose our own understanding on customers, thinking that what they need is what we think.

Therefore, we need to clarify whether the products we make really meet the needs and wishes of customers, rather than just pursuing market needs. We must carefully balance the two to successfully build a popular product.

Speaker G：

How to make a good founder?

Paul Graham：

As a founder, it's normal to see multiple different solutions. When choosing, you should choose the one that you think will be used the fastest. This requires a combination of positive and negative.

The positive side involves determination and a willingness to try new things, but there can be a negative element as well. For example, if you've been at a large company for 20 years, it's very difficult to be a good founder unless visas and other reasons force you to start a business. Because if you had the potential to be a good founder, you wouldn't be in a big company for so long.

In fact, we've noticed companies where alumni often become less successful founders. This is because it is impossible to be a good founder in these companies for a long time. Therefore, when making a choice, we should look for the best solution and avoid being overly negatively influenced.

Speaker H：

Is there a pricing strategy when launching the product, and how will it be priced during launch?

Paul Graham：

If you know your business, you usually know how much to charge, so you can just use your guesswork to set the price. If you want to be more scientific, try talking to some potential users and get their thoughts on the price. Typically, you'll have some docile clients who are actual users, but who are also your friends or something like that, or family, who you can turn to for advice.

Of course, there may be some costs to doing so. However, you can change the price at any time afterwards. If you want to lower the price, no one will complain. If you want to increase the price, it will only affect a small percentage of your existing users, and if your user size is growing exponentially, then there is no need to worry. So, don't get too anxious about it. It is recommended to choose a price that attracts customers at the beginning, because these customers can help you learn a lot, and it is okay to change the price later. Definitely don't charge high prices, because you need customers first, not high prices. Customers educate you and give you feedback, so the first customers not only give you money, but also help you grow.

Speaker I：

How to find angel investors and deal with them?

Paul Graham：

As a founder, fundraising is something that needs serious consideration, but is often overrated. It's best not to spend all your time fundraising because it will distract you. If you don't know how to find angel investors, you may need to attend some entrepreneurial events, although these angel investors may not be the best choice. In fact, companies like Y Combinator help startups attract angel investors.

What if you can't find angel investors? It's like asking a pilot how to walk on the ground. In fact, angel investors will usually come to you, so go out and find them. It would be best if you could find people working at startups and get them to introduce you to their investors.

Speaker J：

In Silicon Valley, it's a trend that companies start hiring people without college degrees. Is YC thinking about getting funding for high school students?

Paul Graham：

Whether to consider sponsoring high school students, I'm not sure because I don't think on behalf of YC anymore. But I don't think it's a good idea to do so. While some high school students may succeed in starting a business, they shouldn't. Just because you can start a successful startup doesn't mean you should. If you start a successful company, it means the days of freedom and fantasy in your life are over. I think in high school or college, you should be figuring out what your options are, not just picking and going with it. Usually those who stop and start a business while in school, like Harvard dropouts, are a tiny minority. Most people still need to dabble, they need to have a fallback, not to be a startup founder.

We'll fund people who put in 100% in college, even in high school. Otherwise, we will not fund them. While the trend is to go straight into work after graduation, between the ages of 18 and 22 you have a lot to do growing up as a human being, a humane human being. At this age, it's great to mix and do a whole bunch of different things. But launching a startup is like catching a dragon's tail, if it works. So be careful and make the right decision at any point in your life.