Every time an engineer joins Google, a startup dies | Hacker News Reader

Every time an engineer joins Google, a startup dies

cdixon.org

142 pointsmoses1400149 comments

"Whenever I see a brilliant kid decide to join Goldman Sachs, McKinsey, or Google, I think to myself: a startup just died, and as a result our world is a little less wealthy, innovative, and interesting."

Or, a future founder just began the process of gaining experience and perspective.

This is a bit of a trap. It's easier to take big risks if you have little to lose. Once you're established at McKinsey, Goldman Sachs or Google you have a comfortable standard of living and a job that gives you both challenges and social status. Letting go of all that to be the guy slaving away at a crazy idea is often harder than trying the crazy idea because you have little else.

Also when you're busy at work time can just fly by. It's easy to look up from working and find a few years have gone past.

"This is a bit of a trap."

Yes, it can be. The truth is, some of those brilliant kids going into BigCorp are lost startups, some are future startups, and most won't be startups regardless, because most people aren't founders.

"Also when you're busy at work time can just fly by. It's easy to look up from working and find a few years have gone past."

Roger that.

The truth is, some of those brilliant kids going into BigCorp are lost startups, some are future startups, and most won't be startups regardless, because most people aren't founders.

Exactly. I will be starting at Google in June, but I am not a "startup lost". I have thought about it, and eventually came to the conclusion that I am not startup material for a number of reasons. This may change in the future after I have worked for Google, and if so, I can always join a startup then.

You probably know this already, but it if you dream of eventually doing a start up make sure to keep your expenses under control now. Your expenses tend to grow along your paycheck. It can be really hard to downsize if you decide to go the start up route.

I agree with this. Even if your paycheque is $5k a month, practice living on half of that or less—maybe a lot less. It won't hurt to save the rest anyway.

Agree, agree, 100x agree!

You do not need a BMW, buy a corolla. You do not need the loft, live in the cheapest decent unit you can find. Don't get into the showy lifestyle trap.

If you live in Europe (or New York), you don't even need a car.

Or San Francisco, or Tokyo...

and for everything else, there's ZipCar (I just sold my car :)

One of the rules of thumb I heard somewhere and try to do (though it's pretty damn hard) is to save 1/3 of your gross salary.

I don't have a rule of thumb, but my actual number is usually more like 2/3 of gross saved.

It helps to have the car paid for and not have a mortgage or apartment lease though.

That seems impossible for me - taxes and other deductions are over a third of my gross salary.

Ditto, my taxes are about a third of my gross, but I do save about half of my net, making for about a 1/3 savings rate overall.

Honestly, I haven't hut to cut on many things to get here, so I'm pretty happy about it. One key thing has enabled this:

No car. None, no desire for one. I pay extra to live centrally in the city near good public transit, but the extra I've paid (compared to living in a burb) is a tiny fraction of the cost of car ownership.

By my estimates that alone gets me an extra 10% savings rate at the very least. I have all the spare cash in the world to travel, buy cool toys without ever digging into the rainy day/startup fund.

Being childless certainly helps, too.

Yes, I mostly save "about" a third. The hard part is getting to above a third consistently. I agree with the car part, but with the proviso that it doesn't work so well when suddenly you end up with a job in some industrial park that doesn't have good public transit.

True, but you can always wait to buy a car until you need it, and the savings will* still outdo the cost of a rental to tide you over while shopping for one.

*Estimating here. Looking at Hertz.com, you can rent for ~$250/week, which for me equals about 4 months in minimal car insurance.

You're right. I guess I don't factor taxes in at the time, since I pay them out of savings.

2/3 of net, however is very possible. Works out to something like 50% of gross, maybe a little less.

I save very close or more than 2/3 of my current salary.

I own my car, I bought it in cash. Where I live I need it (I already tried the whole Vespa thing, I lasted one winter and was exhausted).

I don't have a mortgage. I sold my house.

I rent. I don't want the baggage a house brings (time, maintenance, stuck in one place). My rent is ~ half of my old mortgage which was not that much and in a much more enjoyable community (behind campus and about 2 min walk from the video store, coffee shops, and grocery store).

I only buy what I truly believe I will hang on to and be happy or content with.

Why do I do this? I don't know, maybe to retire early. Maybe do the whole start up thing someday.

However, I think it's mostly because it's comfortable for me. I don't feel comfortable in one place and like the ability to just throw everything in my car and go. I don't like things that hold me back or slow me down.

Edit: this is after taxes and whatever else they take off. I'm going by what my paycheck says.

For me it is the startup thing. 1 year of saving 1/3 of your gross (1/2 your net) is one year of runway when you quit.

This is/was more popular in eastern Asia, particularly in China when Confucian culture was in its heyday. I remember growing up a Chinese family bought a house in our neighborhood with cash. He was a garbage man, they had two kids, and his wife didn't work.

I've managed it since I was 25. Here's what I found.

* If you're making below or well below the median salary for your state, good luck getting past 10% of your net pay. 30% of net is much easier when you're at or just above the median. 33% of gross requires you to have made those "long-term purchases" already, such as furniture and a vehicle.

* On cars, try to buy the first "long term" one in cash. Yes, this takes time to accomplish, but the reality is once you can do it the first time, you'll be able to easily continue doing it for the rest of your life. Buy what you truly want, otherwise you'll find excuses to replace it earlier. 18-35 year olds in America would be much happier if they had bought a Miata or a Mustang or what they had actually wanted instead of their Corolla or Taurus they felt was a "more responsible" choice.

* Learn to cook. Last night I had a steak, homemade biscuit, asparagus, couscous, drink and a homemade cookie for just under $3. Packaged foods and mixes are never cheap, and the cost of prepared foods is staggering.

* If you don't love it and see yourself using it 10 years from now, why are you buying it? The pricier the item, the longer you better see yourself using it A huge sink in consumer culture is buying "disposable items", particularly furniture, gadgets, and even cars in some cases. Buy once, take good care of it, and avoid having to buy it again for some time, if at all. Furniture especially - go used to save money if the piece is still in production (typical for modern classics), because odds are even if it's already 40 years old it'll still handily outlast you.

* Live as close to work as possible. Mass transit is cheaper than anything (I can go all over Boston, as much as I want, for $59/mo). The higher cost of living is more than offset by the extra time to cook, relax, and decompress.

* Reassess your budget, in full, every year. Prune possessions and figure out why it was a waste. Repeat until your friends refer to you as that really cheap person that somehow has really nice things.

On cars, try to buy the first "long term" one in cash.

You will be absolutely amazed what it does to your finances if you don't have a car payment. Too many people treat them like they're laws of nature. Google for [Drive Free, Retire Rich], which aside from poor math due to overly optimistic assumptions of investment returns is an amazing video.

I don't think the maths was wrong, just the assumptions were possibly optimistic. A $1500 car doesn't really depreciate much (unless it dies), but a $12000 car does. And then there's the 12% investment return figure.

(Still - the overall point is very valid)

I agree, and will add that kids are expensive, especially if you want good education. "Ramen profitable" is a lot harder to reach when your life gets to that point.

> agree, and will add that kids are expensive, especially if you want good education.

These factors aren't linked if you're willing to put in the effort.

When growing up we lived in one of the worst parts of the state regarding school districts. Parents began to teach us how to read, write, and do basic arithmetic before we ever were sent to public schools. I think I entered first grade reading and writing at an 8th grade level and knowing long division. Took until about 8 years old for me to grasp basic algebra. They held back after that point, otherwise we would've had a mind-numbing experience until at least high school. However the school board was made aware that these first graders could be head of the 6th grade class, so we were allowed to freedom to learn other things (and sit in a few other classes once a week - mostly English, math and history).

Oh, right. College. That was our responsibility. They would help with loan payments a bit, but they certainly couldn't afford to even put us through a state school. We were expected to get scholarships. If we didn't, well, it was that much more we'd have to borrow.

Honestly, I can't blame them. We were certainly a lower-class family. They couldn't afford to move near a nice school district, private schools, or help with private colleges, so they decided to get us WAY ahead of the curve (even though I have memories of crying because at 3 years old I was having obvious difficulties with multiplication. ultimately we were incentivized with spare change). Even then, the nice school districts aren't all that great. What my folks did will still allow any child to get far ahead of any child in the best private schools in your area.

In the end we ended up with extremely large scholarships (that I didn't bother with because I'd still have to take out the equivalent of a home mortgage) and teenage years with an enormous amount of free time. Cost to my parents who only had a high school degree (mom) and a GED (dad)? Some of their free time every night.

Cool story, it is good to look back on your childhood and be proud of the decisions your parents made.

As a parent though, you have to remember that even basic play has a very high value to children. My little 2.75 yr old has dinosaurs in the bath and can count one, two, three, four, lots. He knows the lifecycle of the butterfly.

But I'd much rather that he did what some might consider "useless play" and mastered his social skills and self confidence than any curriculum ideas such as maths or reading. Those can wait until he is ready, though he is a very engaged and switched on little boy.

I think that forcing your children to do certain things that you expect from a 10yr old now is just ruining their childhood and can reduce their delight in learning new things.

So be careful whose advice you take, even mine.

Actually, childhood was a blast. The first 8 to 10 years of schoolwork being completely trivialized meant I could be over at some friend's house every day, futzing around in the backyard, or playing with action figures and legos. All the time.

I have few memories of before I was 4 years old.

* Rolling around some cookie monster thing. * Being sick of flash cards. * Spraying my grandfather with the hose, and thinking it was hilarious. * The dog knocking me over to take the entire bag of snausages. * Playing random games we made up in the backyard with the kids next door.

How much do you vividly remember from when before you were 5 years old? I'm guessing not a whole lot.

Interesting thoughts about memory. My early memories aren't quite that good.

I never found school particularly difficult, and I didn't learn how to read or do maths until school. There is a good chance you wouldn't have found it difficult either - even without the flash cards. Perhaps you are just smart.

It sounds like your childhood was still fairly moderate, it is not like flash cards is the only thing that you did. My main motivation for commenting at all is that I don't want any other parents to get weird, extreme ideas.

Anyway, I think I better gracefully bow out of this discussion before I step on any nerves.

I've got kids myself, and I blogged recently about how having a family is forcing me to keep my fulltime job longer than I would like to:

Most startups seem to embrace 60-80 hour weeks, you keep reading about them and begin to believe that this is normal. After 12 hour marathon coding sessions ‘typical entrepreneurs’ walk 10 metres from desk to bed and collapse in their shared accommodations. Living in such lean conditions makes those crucial first months of business far cheaper. This work ethic and minimial living costs maximises the runway before the seed money runs out.

Paul Graham is one of my favourite bloggers, and his essay entitled The Other Road Ahead he paints a clear picture of how lean a start-up can be, stating "You can literally launch your product as three guys sitting in the living room of an apartment, and a server collocated at an ISP. We did."

No matter how much I’d love to take the lean ramen noodle approach it’s simply not practical for me. I’m a father of two, a husband of one, and an employee. I’ve got a beautiful family with young children, a modest house with a mortgage, and a full-time job to pay the bills. In short, I have a standard existence that I imagine many of my readers share. Living off noodles in a cheap apartment with my co-founders is not the only way that start-ups are built. I’ve taken an alternative route that involves just as much hard graft (harder maybe?) and allows me to remain employed full-time. The food sure is better!

That is exactly right.

I wish to start up a company. Or devote more time to the companies that I already started. But mostly because I don't feel I fit any more in the corporate bureaucracy and politics that drains my energy every day. Yet it is hard to trade the security of a good paying job for the risk of starting up a company.

Then the only other option is to go hell for leather on your start-up in your free time, even if that means you don't have a life. If it doesn't get you that fired up, perhaps you need to brainstorm a better opportunity.

I can't leave my cushy job yet either.

Plus if it is a goodcrazy idea, chances are that it might be built as a google application.

Ah, but if you have managed to make yourself ridiculously employable by proving yourself at a few high-profile companies, then you suddenly have "little to lose" again.

One of the reasons I feel like I can safely Entrepreneur myself down to a few grand in the bank before looking for a contract to top up the finances is that I know I'm pretty much a single Twitter post away from a good job offer.

Had I not first established a bit of a reputation before taking off on the startup track, I wouldn't have that safety net, and might be a bit more quick to grab any contract job that came along.

That is a pretty good safety net right there.

Though in some ways it is great to "burn the boats", because having your back to the wall and being forced to make it work (even if that involves a week of cold calling) can be the difference between success and giving up.

Exactly. And having a "name brand" on your CV helps more than having been to a top college. 3 years at McK or GS and you will never have trouble getting anyone to take you seriously in business. It doesn't excuse you from anything important, but it does mean you don't need to waste time establishing your credentials to everyone you speak to.

Same with startups. One moment you're in college starting a startup, the next you're 3 years older still doing the same thing. Startups are a more difficult path because they don't always have forward momentum because of the high risk.

I'd love to see more people to pursue this stream (Join Google, learn, and become a founder in the future).

There's a lot to learn from Google (especially in terms of code review, writing good, clean, well documented code).

Let's rephrase that: "Every time I see a brilliant kid going to a chaotic startup that encourage shortcuts and brilliant hacks, I see our industry dying slowly."

PS: I understand that not all startups are like that, but what are the odds/ratios?

Not necessarily. After all, Unix was built by brillant kids who glorified shortcuts and brilliant hacks.

I understand this is an ironic thing for me to say, given my alias, but is UNIX (along with C and C-descended languages) the best we could do?

For the PDP-7/11? Probably.

Certainly agreed, it's amazing what they've accomplished on such limited machines, just it shouldn't be held as a pinnacle of engineering or computer science (else, why would the people who built it be at working on plan9 and go?)

The best we can do is usually not the goal. Good enough is the goal.

People look at great products and say that sentiment isn't true, but you always have to release. Take the first iPhone. When it came to feature checklists, it was embarrassed by every competitor out there. It nailed the interface though, and that was Apple's definition of "good enough". The shipped, took the industry by storm, and have since added most of those checklist features that everybody thought they needed in the first place.

Every feature you add makes the menu more complex.

There comes a point where removing a feature is itself a feature.

Were they kids though? I thought they already grew beards and some might have a tiny bit of grey hair here and there...

There may be beautiful hacks in UNIX but the building principals (UNIX Philosophies) lives on and I don't think kids these days that go straight into startup even know these things.

The other thing is that UNIX wasn't built a start-up, it was built at Bell Labs. Google could be argued t be a modern bell-labs.

Most modern start-ups avoid, at all costs, any systems programming (or doing anything at all beyond using a scripting language to read text from a database and display it on a screen).

Note, however, I said most. There are startups solving serious scientific and engineering challenges. It's just "start-up" does not equate "lasting technological/scientific contribution".

"Google could be argued t be a modern bell-labs."

Particularly given the ridiculous number of ex-Bell-Labs employees at Google. One of the SVPs of Engineering at Google was formerly the VP in charge of the organization that birthed C, C++, UNIX, etc.

I think the key here is to avoid doing any work at all that is not your value proposition. If scripting languages are good enough and will benefit you, use them.

We use Python, Twisted, Django, Matplotlib, JQuery, Movable Type, Postgresql all for this reason. Even our firmware is built as much as possible on a 3rd party neworking library and high level hardware building blocks.

That's great for you (if it's money you're after), but it does make the start-up fairly uninteresting to someone who's interested in systems level development (and not just glueing together a product based on work others have done).

Generally, top scientists/engineers are not motivated by money, which means given an offer from a typical ("glue a product together") start-up and Google they'd choose Google, unless they're looking for the "start-up experience".

If you want top technical talent (who will always be able to find fulfilling and challenging work at the whatever the current Google-like company is, e.g. SGI in early 90s, Netscape in mid 90s, etc), you have to give them top technical challenges (which is how these companies have themselves been able to compete for talent when they were start-ups).

Yes, that is my frustration also. I would rather be doing low level optimisation work, but I haven't yet found someone who wants to pay me for it.

In the next 6 to 18 months, I am going to have to start swapping out the big building blocks with parts meant to scale. At that point there should be plenty of room for the systems engineers to join me.

That's fine then. It's better to use high level components to build a product that's correct (i.e. does what customers want it to do well) and then make it fast then build a fast, but incorrect product.

Be careful, however, with treating scale as equivalent of performance. You have to design for scalability, not optimize for it: given the same algorithm, sorting a list of items in python might take 50 times as long as doing it in C; yet using an n^2 vs. n log(n) algorithm a 1 000 000 item list would take 50000 times as long to sort.

Please excuse the tone of this post. I have spent a lot of time thinking about potential performance issues, and rationalising my choices.

I'm a ACM ICPC (programming contest) finalist, and have competed in the Google Codejam, so I know about big-oh.

Anyway, who implements a sorting algorithm in pure Python? I use the builtin sort() function, which is implemented in C (making callbacks into python for comparisons for some types). It uses TimSort - an approach designed to minimise the number of comparisons. http://svn.python.org/projects/python/trunk/Objects/listsort...

I thought you were going to point out that I should ensure that my system is modular (stateless, partitioned, etc) and I can spread the load across several machines - I can.

You are right that I could make a screaming fast bubble sort and still end up with major issues. Fortunately I already have had that epiphany while solving ACM practice problems.

I appreciate your advice, there are far too many who haven't yet understood it.

Sorry if I sounded condescending. I didn't know the specifics of your system (is it mostly on the server side, embedded, etc...?) so I couldn't give you any specific advise.

The sort example was merely a metaphor for design for scalability vs. optimization for performance. Perhaps better examples would have been dividing the server side portion of your application into asynchronously invoked services is design for scalability. Rewriting some of these services in C or OCaml, switching from JSON to Protocol Buffers, switching from an HTTP server and a layer 7 load balancer to a custom non-blocking server and ZooKeeper (for cluster membership) are optimizations for performance/stability/cost.

There's a small caveat though. It may have been a clearly good choice to go to Google in early to mid 2000s, but since then there's already an incredible amount of legacy code written supporting business critical functions at Google. This isn't a criticism of Google at all, but it is a limitation on the new things you could do at Google or any other company with legacy systems. Working at Google doesn't free you from the responsibility (not just to yourself, but to the people using your work), as an engineer, to learn on your own what you aren't learning at your workplace.

That being said, I have incredible respect for Google for being a few engineering/science driven Internet companies out there, which respects great hackers and beautiful, quality code. Most Internet companies (including many startups, with few great exceptions) are solving social, rather than scientific or technical problems and treat developers the same way Google treats servers (as a commodity, where the ability to replace them quickly -- and make do if they're unavailable -- is more important than their quality).

Speaking about legacy code, I read somewhere (probably via twitter) that new hires were given Michael Feather's book: "Working Effectively with Legacy Code".

I have a lot respect to companies that acknowledge these kind of issues and try to fix them.

I know there are many companies that just tell the engineers to quit whining/complaining about the current system and keep patching issues (fire and extinguish).

That's very interesting! I'll have to check that book out (just given some of the open source projects I've dabbled in), anyone else have an opinion on it?

They have me a copy of that book during my orientation at Google, along with a series of other books on the core languages used within Google. Obviously, it's up to every engineer to put that to good use, but the company is making an effort to have people do things the right way.

May I know the titles of the other books?

Don't go into Google to learn to code. Go into Google to learn to lead. That's what you will need the most as a founder.

Exactly, most truly successful entrepreneurs I know had a few (or a lot of) years in a larger company before starting their company. This gave them not only vital experiences, but also a lot of insight into the problem which their company tries to solve.

Including Steve Jobs (although I'm not sure on the number of years for him)

I absolutely disagree. With YCombinator and the dozen similar programs out there, the traditional barriers to starting companies, namely "not knowing what you don't know" are being removed. For a web start-up especially, the only cost associated with the startup for a group of young founders is time.

The goal at Google, Goldman Sachs, and McKinsey is to recruit top talent, and give them enough money and benefits that they are completely content and never have a reason to leave. Some still will, but most are probably destined for a career as a middle manager.

Maybe this is true at Google, but you need a fair amount of dedication to stick around at Goldman or McKinsey. Nobody becomes a career middle manager there. Well, maybe some back-office jobs at GS.

According to friends at Goldman, it's quite easy to be a career middle manager at Goldman in IT. There are many people there for 20+ years, the only job they've ever had.

I've been told that McKinsey has an "up or out" culture (get promoted or leave).

1. Get recruited by Google

2. Work at Google for a few years, start a 20% project

3. Leave Google, become a founder

4. Get funding, hire more employees

5. Launch a successful product, build a big userbase, get favorable media coverage

6. Get acquired by Google

Isn't the 20% product owned by Google already?

Yes, but once you're "the guy/gal who wrote $SUCCESSFUL_GOOGLE_PRODUCT", you have a lot of credibility, experience, etc. as resources to make your own product that you do own.

The point of the 20% products for Google is to give the company startup-style flexibility. It works the other way, too - if you take a product from whiteboard doodling all the way to an official Google product, well, a startup is like that, except 100% of the time, on your own bank account, and with no admins/middle managers to take care of the business side for you.

Yes, but given the 20,000 employees at Google, how many $SUCCESSFUL_GOOGLE_PRODUCTs have you seen per employee?

I'd guess ex-Googler startups tend to draw more from their founders' general experience at Google than from specific products, but I've heard of rebooting the idea of a 20% project outside with a broader scope.

Interesting timing :) Looks like this is what happened with Aardvak. I'm just speculating, but the formula seems to hold true in this instance.

http://news.ycombinator.com/item?id=1118214

Not to mention that startups can't do everything a company the size of Google can (such as start their own ISP providing gigabit service).

There is a place in our society for all shapes and sizes of companies.

I think the fact that this comment keeps getting up-voted is an indicator of the high ratio of Google employees to startup founders on HN this morning.

Funny you should mention Goldman Sachs. The startup I work for is founded by a former Goldman guy, and I'm looking out my apartment window at a startup founded by a bunch of Google guys (BloomReach).

Not to mention that if you're successful at Goldman Sachs, you can quit and spend the rest of your life doing startups and still have enough left over for a golden pyramid when you die.

Don't forget Microsoft.

Looking back over the last decade, it seems like Microsoft is a lot closer to IBM than to Google.

People don't go there to learn how to innovate. they go there to work a full-time job for the rest of their lives.

And gaining contacts and cash. Which in turn help him later to build a startup. So I agree, it's not all bad.

There are a bunch of advantages to the non-startup route:

- Weekends are yours.

- Guaranteed good pay. After ~4 years at MSFT, I was making nicely north of $200k gross. As a developer. In 2004 dollars --- it's probably more now.

- There are technical people more experienced than you, interested in mentoring your growth. I used to joke as a hiring manager that you could tell quickly in an interview how long someone had been in the startup pit because they had learned no new concepts (only new technology) since they started.

Why would anyone technical who doesn't explicitly want the startup experience do it these days? It's not like the mad IPO cash-out and low corporate pay days of the late 90s / early 00s.

I think this is the flip-side of the startup coin that many people forget about. The quote from the article that stuck out most for met was "But most of all: nothing is more motivating and inspiring than the sense of ownership and self-direction only a startup can provide."

Having decided to walk away from the startup I helped to found, I couldn't disagree more. The three things mentioned by larsberg are three things I'd grown to sorely miss. In particular, no-one is around to teach me test-driven design, and when I'm already putting in 120% during the day, I'm deadened to the desire to self-learn through books and scavenging on the internet.

Did you join MSFT straight out of college? If so, that's amazing money for only 4 years experience... or at least it seems that way to me

I agree with you. If you are really a technical person, being the founder is not the most enviable thing to be. You can learn much more being in a successful big company. However, if you see technology as a way to make money, founding a business is the best way to get ahead. It all depends on what you really value.

> - There are technical people more experienced than you, interested in mentoring your growth. I used to joke as a hiring manager that you could tell quickly in an interview how long someone had been in the startup pit because they had learned no new concepts (only new technology) since they started.

It's hard to over-state that. When joining company (start-up or not), I always make sure I am would not be the smartest person in the company (having made that mistake once and deeply regretted it).

Care to explain?

there are probably 100 Steve Jobs born every year. The vast majority just never have a chance or give a thought to starting a revolutionary new company.

A Steve Jobs that doesn't give a thought to starting a company is not a Steve Jobs. That's all he thought about from a young age. Just because our "top talent" often goes to big firms doesn't mean our top founders are going to big firms. A startup founder is more than just top talent; it's a very rare breed.

I think the reason so many VCs think the VC business needs to shrink is not that there' some limit on the number of startups, but because of a historical accident that caused the VC business to bloat suddenly. The bubble of the late 90s caused a great increase in the number of VC firms, and many of the new arrivals were incompetent. When VCs say the VC business is going to shrink, it's largely a tactful way of saying that those firms are going to get GCed.

Actually, I pretty strongly think that there is a limit on the number of startups, if by "startup" you mean "small company in which people create wealth largely through computer programming".

According to the BLS, there are 1.3 million "software engineers" and programmers in the United States (cite: http://www.bls.gov/oco/ocos303.htm). This number will go up over the next fifty years, but not by that much; projected growth is 2% annually. One could see this number increasing more rapidly if programmers are in so much demand that their salaries (either in the form of cash or startup equity) go up very fast, but we don't see that happening; programmers nowadays don't get paid that much more, in either dollars or equity, than programmers five or ten years ago.

Now, how many of these programmers have the skills needed to start a successful company? Let's be optimistic and say 10%, which is probably above your estimate ("The top 5% of programmers probably write 99% of the good software", cite: http://www.paulgraham.com/wealth.html).

Now, out of those, how many have the personality and determination needed to start a successful company? In the overall population this number is probably less than 5%, but the traits needed to become a good programmer (high energy, determination, ability to think well) probably correlate significantly with the traits needed to start a successful company, so I'll be optimistic again and say 25%.

Now, what's the viable window for programmers to start companies? At a guess, I'd say probably around age 22-37, so 15 years. I'm going to be optimistic again and assume that all working programmers are within this age range.

Now, how many founder-class programmers do you need to start a successful company? This number could probably range from one to five, but I'm going to be optimistic again and say one.

So, dividing it out, the number of successful startups every year is bounded above by 1,300,000 * 0.1 * 0.25 / 15 = ~2,200, and this is almost certainly an overestimate.

You've excluded successful startups founded by non programmers.

I explicitly defined "startup" to mean companies that need programming skill (as this seems to be the way the word is used nowadays), and there are very few of those that have zero programmers as founders.

I followed your calculations, but I am not sure what point are you trying to make. Are you saying that the upper limit of start ups is 2200 and it is limited by the number of highly capable programmers. What about the Market? Is there enough space in the market for 2200 new start ups?

My point is, even though there are an infinite number of applications to write, there aren't an infinite number of people to write them.

That's equivalent to asking, are people's desires limited? The answer appears to be no, in the short term.

It is more asking:

A) Do people have the time/money to use X products? and B) Are there X people who want to use a given product (at a given price/time point)?

I have no doubt that there is a limit on the number of successful startups due to A and B.

The number of programmers is pretty elastic. People can and will learn to program if it starts to seem like that's a valuable skill. E.g. there is a founder in this current YC cycle who came to the US to go to business school, but then realized that if he wanted to start a startup he'd be better off spending the time learning to hack instead.

I'd say the upper bound on the number of startups is the number of people who are determined and smart enough. That's probably between 1 in 100 and 1 in 1000.

"People can and will learn to program if it starts to seem like that's a valuable skill."

That's basically just a restatement of what I said (more people will learn to program if there are greater rewards for doing so). The thing is, there aren't greater rewards for doing so. Startups aren't any more profitable now than they were ten years ago, and programmers at big companies don't make more now than they did ten years ago (inflation-adjusted).

"I'd say the upper bound on the number of startups is the number of people who are determined and smart enough. That's probably between 1 in 100 and 1 in 1000."

Few could argue that, say, investment bankers are quite smart and extremely determined (and very good at sales besides), but there's extremely little overlap between investment bankers and technical startup founders, and that doesn't look like it's going to change. Same for corporate law, management consulting, etc. Plenty of i-bankers and consultants and corporate lawyers go into other fields, including non-technical startup positions (eg. Jessica Livingston). Indeed, one of the primary reasons people go into these industries is for the exit opportunities. But very few seem to learn how to code.

Maybe I should clarify what I mean by "upper bound". "Upper bound" means "how high could the number go if parameters X, Y, and Z are allowed to vary", but neither of us specified what X, Y, and Z were. The hypotheticals I was considering were "suppose the startup community all acted in unison towards increasing the number of startups". For instance, if the startup community exerted enough collective pressure, we could plausibly get Sarbanes-Oxley loosened, or force VCs to become more technically competent, or make it easier for founders to immigrate. But we can't just decide to increase how profitable it is to learn to program or start a startup; these things are driven by consumer demand.

People can and will learn to program if it starts to seem like that's a valuable skill.

But they won't tho', not really. In the dotcom boom there were plenty of such people who learnt HTML and were (well) paid for it for a couple of years. But they didn't or couldn't make the jump to possessing a skillset that kept them in the industry after the meltdown. The guys with C++ or Sybase or whatever under their belts, just kept right on truckin'.

All of the scientists I know know how to hack. A philosophy grad student I know built a neat little philosophy bot. http://webscript.princeton.edu/~amendelo/skeptic/

Hacking is starting to be a new literacy. What's more, the most important part of startups isn't really programming, which is getting easier and easier, but understanding customer's problems. Here a wide range of backgrounds are useful.

I think this number also depends on the bound on markets that can be monetized, when you see the scale that some startups need to achieve to become profitable their is only really room for a finite number of successful startups in any one market space.

One interesting thing is that a very good programmer might look at those numbers and go "how am I going to beat 1,300,000 people to make the next facebook" whereas it is closer to "how am I going to join the 2,200 startups that make it big?"

You are right. However the airline industry is a demonstration that persistent oversupply can sometimes exist indefinitely.

Though admittedly that example is being propped up by the Chapter 11 bankruptcy laws that repeatedly lets companies on the verge of dying get rescued instead. (I'd love to see chapter 11 bankruptcy go away.)

If chapter 11 bankruptcy goes away it will be bad only for employees. Because investors are already wiped out by bankruptcy and upper management can always negotiate sweet deals for when/if it happens.

I'm guessing that you're not very familiar with the history of the airline industry.

The problem is that there are too many competitors, someone fails, go into chapter 11, with the competitive advantages they get from chapter 11 they become healthy, there remain too many competitors and another one fails. At this point most of the major airlines have been in and out of chapter 11 multiple times, and the airline industry still has structural problems of too many airlines. But with every round life has gotten worse for the employees. To such an extent that pilot has gone from a good job to a piece of crap job that people only stay in because they have already sunk the cost of their training and love to fly.

If chapter 11 wasn't there then this cycle would never have caught on. But the artificial boost for companies that are ready to go away is artificially distorting the market to everyone's disadvantage. And it has stayed distorted for a period of decades.

I disagree, I think it's simply that the VC's are incompetent. They don't use data to make their judgements, relying instead on gut feelings. They don't have a systematic procedure/checklist to make decisions, or even a systematic way for entrepreneurs to to see them. They don't even like to hire analysts as is common in the rest of banking. It's simply not a 'prefessional' industry yet.

It's not the fault of individuals in the vc (they have to run fast just to keep up), just the industry hasn't had enough academics/software telling them to how to more reliably run their business.

When I was in college, I went to the career center looking for a job. I had my resume all spiffy and on it was a startup I had been working on in college.

The recruiter asked why I didn't pursue the startup idea and I said, "I need to pay the rent."

I think that sums it up right there for most. We enter college with nothing, and we leave college with less than nothing (i.e. $100,000 in debt). So we go get a job so we can pay the bills meaning we get $5k a month coming in pretty easily and that's a lot of money. It's definitely enough money for a college kid to pay for rent and be happy for a while and they probably end up being so happy that nothing else is worth the risk.

Why risk $60k a year on a startup that will buy you (if you are lucky) some ramen to eat and some space in your parents' basement? A BMW plus a phat apartment in a high rise is difficult to turn down for something that appears, for all intents and purposes, to have about a 1 in 1000 (at best) shot of success.

That's why we have no entrepreneurs anymore. Life is too easy without them.

"That's why we have no entrepreneurs anymore. Life is too easy without them." Any evidence to support a decline in entrepreneurs? The internet seems to be overflowing with them.

It's probably selection bias. It's easy to find people you want to find on the internet, but look around you in the real world. If you aren't in silicon valley, it's very difficult to find entrepreneurs and it's even harder to find them outside the internet.

I'm in Beijing, I'm in the food industry, and I'm surrounded by entrepreneurs who almost never go online. I turned down the posh jobs out of college and my partner quit McKinsey to found our company. A friend just bought a million dollar factory for his t-shirt and pants company. Entrepreneurship is not just online and the notion that entrepreneurship is in decline feels very wrong to me. I'd like to see data.

You are biased towards software startups. I live in the US and I've lived in India. Look around you as you drive home. There are entrepreneurs everywhere. Small businesses drive around 70% of the economy.

"A BMW plus a phat apartment in a high rise is difficult to turn down for something that appears, for all intents and purposes, to have about a 1 in 1000 (at best) shot of success."

What's a good place to find actual statistics on success rate and financial return for startups?

V.C. is suffering because there are not enough companies that have got to a big enough stage for them to fund.

There needs to be more put into early stage companies. I meet endless companies in the U.K. that can't easily even get $20k to get off the ground. Yes, we've got SeedCamp, but it only funds 5 or 6 companies a year. There are 1000s of good potential companies in the U.K. alone that can't get this first step of funding. And that's the U.K. which is fairly developed for this and there are lot of countries that this could apply to.

Basically, I think we need several Y-combinator style companies in every country in every industry. I think as people see the returns come into Y-combinator there will be explosion of these companies as there a lot of people can spend $2m/year on a range of outside bets.

I know a number of people who after they graduate would love to focus on startups under the right circumstances, but one of a number of constraints they have is that they need health insurance, which isn't at present something that fits easily into a "Boot Strap Budget"

It only gets worse when you have a family to insure as well. Individual health insurance can have lifelong implications and it's a massive, massive amount of additional stress.

I, for one, don't enjoy worrying about a member of my family getting cancer or diabetes or something and dealing with pre-existing conditions clauses for the rest of our lives.

Do not forget about the massive amounts of debt that many students leave college with. Payments start 6 months after graduation.

It actually isn't so bad for CS students at public schools. Some internships (msft, google, facebook, etc.) pay enough to cover a year's worth of tuition and housing.

Well, they could go without health insurance. Frankly, if you're going to be uninsured at some point in your life, such a young age is the best time to do it... it's when you're old and sick that you want to be insured most, not when you're young and healthy.

It not the simple for folks with congenital health conditions or who are can't function due to a medical disorder without some prescription which is easily affordable with insurance, but impossible to afford in a "ramen profitable" san insurance context.

An easy example would be someone with severe adhd. The relevant sort of medication can cost as little as 30 dollars with insurance, but total around $400 dollars otherwise (apparently there can be a nontrivial difference in efficacy of generic versus nongeneric).

Such a person, working at/on a startup, would need to be able to functionally transition between working on a variety of tasks through out the day, ever day, irrespective of how engaging each task is to work on. These are precisely the circumstances where that person's adhd would be most problematic.

And thats a simple example, what about say a person with Crohn's Disease (and need to be able to cover the expenses that happen whenever it becomes problematic), or someone with bipolar disorder who with the right prescription and monitoring at worst has on vs off weeks in terms of productivity.

Yes, these examples are relatively sparse, but nonetheless there are certainly easy to construct examples where the absence of insurance to cover certain regular expenses makes "ramen profitable" untenable

Health insurance is $89/month if you're young. That comes in on the monthly budget right alongside beer and hosting.

Housing and Car payments are the thing that get you.

Could you point to this health insurance? The cheapest I've found was ~250/mo.

If you're in WA, OR, or AZ:

www.lifewise-insurance.com

Balance that with $50/mo at www.qliance.com and you have everything you need.

I have Blue Shield VitaBlue 1900 which costs about $69/month including dental. If you are willing to accept an extremely high deductible ($1900 annually here), you can get relatively cheap health insurance.

Nice, that sounds perfect (though I'm not sure Blue Shield operates in NY...). Know of any good sites that show what insurance plans are available in what states? (or did you find that via some other method?)

I actually just found out from a friend.

There are definitely sites out there that show all rates; just search around for high deductibles. You should have a Blue Cross, Blue Shield, etc. affiliate in NY.

have you checked out the freelancers union in NY? They offer cheaper insurance and other benefits.

Looks like the 10k deductible plan for health alone through Freelancers is still $196/mo... better, but still seems a bit high, considering I've never been hospitalized.

You need to find the plans that aggressively discriminate based on health history. If they aren't hard for the average American to qualify for, you won't get the low rates.

Oh and I don't know your health and finances, but if you are healthy (and presumably hardly risk-averse), no health insurance beats a 10k deductible that costs you $2.4k a year.

You can keep even housing and car payments under control if you are willing to A) live in a smaller place (you can easily get to $450/month in the valley if you have roommates) and B) buy a low-mileage, 8+ year old car.

"Whenever I see a brilliant kid decide to join Goldman Sachs, McKinsey, or Google, I think to myself: a startup just died"

Many smart "kids" (gee, thanks dad, glad you know what's best for me!) want to solve interesting problems, not muck around with all that other stuff you have to deal with as an entrepreneur.

Arguably the best place for that though is a startup where your first employee with a large stake? There's no doubt there would be many very interesting positions at Google, but I'd imagine there would also be some very boring positions for developers straight out of college.

First employees have to deal with a lot of gruntwork. Possibly more than the founders, because they're often the ones stuck setting up the production servers, dealing with customer complaints, hacking through that sticky bug, firefighting, training/interviewing new employees, and generally doing all the boring stuff while the founders are out sealing business deals.

As for Google - it's what you make of it. I know some people who've gotten stuck in a narrow corner of the company, but I've personally been exposed to a pretty deep cross-section of the search stack, and worked on some fairly exciting projects. Working for a big company doesn't excuse you from actively managing your career - if something exciting is coming up, volunteer for it! And nobody can stop you from working on other stuff - if there's something interesting, just go hack it up and ask for forgiveness later.

"if there's something interesting, just go hack it up and ask for forgiveness later."

This is probably the best advice given on hacker news today.

"Whenever I see a brilliant kid decide to join Goldman Sachs, McKinsey, or Google, I think to myself: a startup just died"

What does mere dumb brilliance have to do with creating a successful startup? I thought it had more to do with stamina, discipline, luck, persistence, flexibility, hunger, vision etc.

If that brilliant kid is more interested in working for a big company, maybe it just means that they were smart enough to realize that they weren't cut out for the ramen life.

Or it could be that they didn't even know it was an option for them. Going to work for Google is the easy thing to do after college.

Yeah but not all engineers who join Google stay in Google forever.

The ones who eventually leave and start their own company bring with them whatever parts of the Google culture that they felt were valuable: maybe it's the data-driven, engineer focused management style, maybe it's the code review and use of real computer science in the creation of software, maybe it's the free food rule.

They also leave after meeting a lot of smart people and making a lot more connections that could potentially help them in the future.

Most people here seem to think that startups are more net productive / innovative / whatever than large corporations? Even if this is true, that doesn't mean that everyone should join a startup.

Big companies like Goldman, McKinsey, and Google provide value not only to normal people, but also startups. Google builds tools and infrastructure that help power the Internet. There are probably tons of startups that rely on Google products like Gmail and Google Maps, and without smart engineers working at Google, those startups would be worse off.

McKinsey provides consulting services to companies big and small, including startups. Some startup engineers may not have the best knowledge for how to best access a foreign market, and companies like McKinsey, who do this research all the time, provide valuable business advice about how to set prices, enter new markets, structure growing organizations, and all kinds of other business questions that startups may not know how to answer.

Even Goldman Sachs has something to contribute to startups. GS financial advisers (as well those other banks, obviously) are the ones telling large sovereign wealth and pension funds to diversify their allocations into categories like venture capital. This money is critical for lots of startups to grow.

Maybe these companies aren't completely necessary for startups to start and grow, but they do have a lot to contribute to this space.

The company where I work currently, an analytics company, was founded by a McKinsey alumn.

The company is going on 9 years.

Our biggest client last year was McKinsey and when he first started the company (i.e. before biz dev), he leveraged the contacts he'd made at McKinsey heavily.

All in all, I'd say the article argues one extreme, when as usual the truth is somewhere in the middle.

Many ex-googlers I know are now startup founders.

Right, however most are not.

Most non-ex-googler engineers aren't either.

This article is very true.Normally when a person is fresh out of college the responsibilities are less, he might not be married or might not have a housing loan etc, so it is the right time to start a startup. Once a person joins a big company he comes into something known as a 'comfort zone' , its easy to get into that zone by following a defined path or trend i.e to say good college + good marks + some contacts and you land up a good job with good pay.Its difficult to get out of this comfort zone. I personally feel innovation happens mostly in startups.It does happen in big companies but in bits and pieces but in startups it happens all the time because innovation is the only thing that can make a startup a success.

BS. This is true of anybody joining any company.

He didn't say it wasn't.

Is it just me, or does it seem unfair to lump Goldman Sachs, McKinsey, and Google all into the same class? In particular (having worked at none of them), I'd reject putting Google, an engineering company, into the same category as the other two, which largely cater to institutional clients.

They're all subsidiaries of BigCorp. Once you get past a certain size (employees, sales, market cap, whatever) it's unavoidable to become BigCorp, regardless of the differences in your particular company's culture. Once you've become so capitalized that your company's stock is itself a commodity, you're BigCorp.

Thank you Mr. Tautology.

Not I think. He anybody has ideas and/or dreams and believe in himself he would have pursued it anyway. Now if you have some doubts then obviously go for the safer option. And may be as nfnaaron points out may be they are gaining experience.

This seems to be subtly jabbing Aardvark's acquisition, specially since Hunch (cdixon's startup) is still out in the open.

VC just needs to get out of its parents basement and spread out more. Why not open offices at major universities?

It's hard out there for a venture capitalist... Cry me a river. Let engineers choose the life that it best for them.

Reply on news.ycombinator.com