There's another common thought I hear voiced often: That you have to come up with an idea that no one else has done before - a green field idea. That's wrong. Competition in a space is often a sign that there's something valuable there. Competition is energizing. A lot of the ideas I'm thinking about exploring already have established players or startups - it's just that I think I can execute much better than them.
Yep. There's a quote from Bob Parsons (of GoDaddy fame) that goes something like (paraphrased roughly) "Don't be afraid to enter a crowded market, just be better than everybody else."
This is kind of our mindset at Fogbeam Labs, where - depending on how some details shake out - we'll be competing with everybody from Jive Software, Yammer, Alfresco and Broadvision (among others), to Salesforce, IBM, Cisco and Microsoft.
But that's fine. We have an angle to pursue that we think will let us make some inroads against those guys, and we're spoiling for a good fight. Heck, as far as I'm concerned, those guys should be sleeping with one eye open. :-)
I suspect Peter Thiel would likely say something along the lines of pursue the thing you know you're right about and which everyone else insists you're wrong...
Which boils down to either extreme failure or extreme success and little in-between...
This is something I don't agree with Peter Thiel on - at least for your first startup.
Elon Musk couldn't have started SpaceX and Tesla right out of school, he had to build up skill, credentials, and connections first.
For your first startup, make something awesome but try to fly people to Mars. A small acquisition or a business with nice revenue might be the better goal.
I think you meant "but don't try to fly people to Mars".
I'm mostly leaf reading, but it seems to me you are recommending a lean startup / lifestyle business as a first startup. A less ambitious endeavour than say, Google or Amazon, with the primary goal of acquiring skills, small but safer revenue, a stepping stone to the next, bigger, enterprise. Very inspiring, and something young engineers should think about.
BTW, I restate my request from my unsolicited email (hope you don't mind): please share some startup ideas, even if execution is everything, some of us younger, less experienced engineers find brainstorming and finding ideas harder, and would really appreciate some inspiration.
Google started as an umpteenth search company on cheap machines and offered to sell out for one million, and Amazon started just by selling books from a website with buggy tech and workers kneeling on the floor to package the books. They both started at an easily accessible entry point too. But they each had a feasible path from humble to world-changing in a way that Elon maybe didn't going from online media and payments through a step function into disrupting transportation energy and yanking humanity into a spacefaring civilization. Even with, he was learning to get better at the game as he went along, and laying the foundation for SpaceX at the level of asking the rich questions, making the connections, and digging hard for elusive answers while he was still at x.com/PayPal. You probably always have to go through the foundation of exploring for the right questions to explore, to emerge with a meaningful advantage.
You have to take Peter Thiel as an ala carte buffet:
In my own life, had I chosen the startup over grad school, I would easily be worth $100M+ right now. I had a ground floor opportunity I idiotically passed on because I was young, stupid, and too idealistic. So I see his point.
That said, I shudder at the thought of hiring a programmer who struggled with calculus and nothing assures me better on that axis than straight As on math and computer science courses above and beyond requirements for their major.
Individuals vary, and one of the best coders I know got a 2-year associates degree out of community college, but this is a really nice foot in the door for me when I don't know the person already.
I hope you don't feel too bad about passing on the startup opportunity. Hindsight is 20/20, and no startup is guaranteed to earn $100M for its founders. Maybe it's regrettable for not getting the experience you'd obtain win/lose, but the money part is a lottery.
This is something I don't agree with Peter Thiel on - at least for your first startup.
Elon Musk couldn't have started SpaceX and Tesla right out of school, he had to build up skill, credentials, and connections first.
I agree with your point, but "your first startup" is not necessarily "right out of school." In my own case, I'm working on my first startup, after about 15 years in the IT world. And it's been those years of IT experience that has led to me (and to some extent, my co-founders, both of who are younger than I) to develop those beliefs where we feel like we believe something that everybody else doesn't.
For your first startup, make something awesome but try to fly people to Mars. A small acquisition or a business with nice revenue might be the better goal.
Yeah, we're thinking IPO, and eventually acquiring Microsoft. If you're going to dream, dream big, I always say. :-)
I suspect Peter Thiel would likely say something along the lines of pursue the thing you know you're right about and which everyone else insists you're wrong...
I didn't even realize it was Peter Thiel who said that, but I embrace that mindset as well. We have a document in our repo called "What we believe" that is where we catalog the things that we think about the world, that we suspect most everybody else would disagree with us on. Looking at that periodically is a valuable tool to help us focus on how we're going to be different.
I am a little tired of the "ideas are a dime a dozen" mantra. It implies they are all equally "cheap". Yes, anyone can dream up some silly idea and spout off all day long about it and accomplish absolutely zilch. My dad was an armchair politician who had a "brilliant solution" to every social ill imaginable and never got off his duff to implement any of his ideas. So he never really got real world feedback concerning which ideas had merit and which were him just blowing smoke up everyone's ass.
But that does not mean all ideas are equally cheap and worthless. They aren't. Good ideas are like the DNA of a product or service. Separating the wheat from the chaff is the purpose of executing and iterating. And don't confuse ideas with how people initially attempt to express what's rolling around in their head. People often don't know how to really effectively communicate such things. Iteration is often just an exercise in trying to figure out how to properly express the underlying idea. That is what pivoting is often about.
Good ideas are not a dime a dozen. They aren't all that common. They are testable, the way Einstein's Theory of Relatively is provably a more accurate concept than Newtonian gravity.
I think what people typically call ideas should actually be called something else. I like the term "seeds" - plant it and it will grow into an actual idea. Execution is the water with which it grows.
A big difficulty I think is that some people have a tendency towards idea reductionism. The idea for the iPod was a great one, but reducing that idea to its bare essentials of "I want to create another mp3 player" leaves a lot to be desired. But there are also some ideas that are so elegant and straightforward to implement that they can't be diminished by over simplification and they are valuable all on their own (hesco barriers being a perfect example). And there are some ideas that are tricky to express correctly and to execute but are still incredibly valuable. Ideas aren't everything, but they're a lot more than nothing as well.
With genuinely new ideas, communication can be extremely hard. Existing language is rooted in existing context. If the idea is paradigm changing, the language to express it doesn't really exist. Attempts to talk about it can be very frustrating. The listener may hear something completely different from what the speaker was trying to convey.
I run into this all the time when trying to talk about how I got myself well. It is a proven concept in some sense as it worked for both me and my sons. But that doesn't automatically give me the ability to effectively convey it to other people. I continue to struggle with that piece of it.
"My dad was an armchair politician who had a "brilliant solution" to every social ill imaginable and never got off his duff to implement any of his ideas. So he never really got real world feedback concerning which ideas had merit and which were him just blowing smoke up everyone's ass."
Part of what I do outside work is help out with housing cooperatives in the UK. We put ordinary not very rich people in charge of their homes. They become the landlord. Targets include < 4% void rent in any year, < 28 days property vacant &c.
Real decision making and budget responsibility (landlord liable for repairs) works wonders. You can see people grow taller, and many now have supervisory jobs...
I and my two sons are currently homeless. One of our big complaints about many of the programs intended to help us is the assumption we are incompetent, lack agency, should be "grateful" to a degree that amounts to agreeing to be their bitch, etc. Kudos on being involved with something better than that model.
UK: housing cooperatives are incorporated as Housing Associations. Many in the Scandinavian countries as well. I thought the values of the coop movement would translate to US quite well (independence, responsibility &c) assuming you are in US. Coops are fully mutual, you can't buy only rent, but there are self build schemes for participant purchase as well.
I do envy the rental sector in US. UK has issues with artificially high prices (my opinion) preventing a free market rental sector from emerging.
Last I heard, coop housing has serious barriers in the U.S. I think those barriers are largely rooted in the housing finance infrastrucure born post WWII.
The "boys" came home, often had high savings and no kids, "Rosie the Rivetter" was encouraged to quit her job so a man could have it, many women did so quite happily so they could have kids and the fulltime mom thing. Prior to the war, we were deep in The Great Depression and many people moved every thirteen months to take advantage of one free month's rent if you could pay rent in full and on time for twelve months straight. People were doubled up in small homes, etc.
Levittowns were born to meet the sudden demand for single family homes. The nation created finance mechanisms to facilitate the birth of these instant suburbs. We are still living with the legacy. It can be extremely difficult to arrange financing for anything other than a single family home.
Actually ideas are very much like armchair political solutions. Often easy to state, and often very hard to do. There is even an axiom that goes "For every difficult problem there is a solution that is both easy to do and wrong." or something like that.
The reason ideas are so cheap is that they are only the reflection of an action, not the action itself. Often folks take what they 'want' and then ponder a way to 'get' that and poof! idea. Like "I 'want' to work on my car's engine computer, all I need to do is take the engine out and mount it on an engine stand and hook it up to my laptop!" Simple idea, remove the engine and instrument it, but actually doing that is a lot of work. (doable, just a lot of work).
But this can also be a useful tool. Ideas that aren't a lot of work to implement but seem to have high value are usually useless. And if you can't tell why they are useless then it is sometimes worthwhile to pursue them enough to understand where the real work is.
But take something that is pretty much pure software. Let's take electronic publishing briefly. You have no doubt read a paperback book, you've no doubt read one (or looked at one) on an e-reader. All words, no pictures, just chapter breaks, perhaps the occasional font change to represent internal vs external dialog. Why can't I cut the spine off my old paperback, feed it into my Fuji Scansnap1500, and have it drop a .epub file that works on a kindle/nook/ipad/whatever ? The scansnap will give me a nice PDF of that book, even searchable. Caliber will convert it into a horrendous epub file. Why isn't this a no-brainer? I'd pay $20 for that software, no problem.
Idea, no execution. Marginal return on investment (while I would buy it if it worked well, there is no assured market here) There are bodged together ad-hoc solutions out there. Clearly there is some demand but as Gabe called out, you need passion. You need to want to build this because you want it to exist. And that, that passion, is not there.
I tell armchair politicians, hey run for city council. Solve problems. If you are good at it people will give you bigger and bigger problems to solve. If you aren't you'll both have an appreciation for what politicians are up against and a compassion for what they do get done. Not nearly as fun though. At a Tom Campbell said at a fund raiser once, its hard to appreciate the job from the outside.
No, ideas are a dime a dozen. You seem to think of ideas as "thought + action" but then it's not an idea anymore; it's a real product with an idea behind it. Your example of your dad is a perfect example of why ideas really are a dime a dozen. That mantra isn't about saying that ideas themselves are worthless, it's about the idea that ideas alone with no action are worthless. I think the mantra stands as correct still.
Sorry, I don't agree. There are some very important game changing ideas. The concept of zero is historically important. The concept of gravity is historically important. Those were not companies, they were only ideas. Yet they changed the world. So I don't think it is that simple.
That "ideas are a dime a dozen" mantra upset a fair amount of people, especially combined with my assertion that I could easily brainstorm up good ideas in no time. I changed it (accurately asterisked) to "ideas are easy", which more accurately reflects what I think.
Competition in a space is often a sign that there's something valuable there.
The company that this really brings to mind is DuckDuckGo. They're competing for the search market with Google and Microsoft--two of the biggest competitors you could possibly face. But search is huge and if you can claim even 0.1% of that market, you've got more than enough users to justify your existence.
Google is fat and happy in search and ripe for the death by a thousand paper cuts IMO. DuckDuckGo is the first cut, Siri is the second, CloudSearch is the third. There will be many more.
Siri? Doesn't Siri essentially aggregate Google's results directly? Also I'm not sure what user statistics on the service are but I don't know anyone who regularly uses Siri for anything other than to show people when asked about it.
My perception may be shaped by past experiences at various LargeCo's, but when I hear the phrase:
"I'm leaving to work on my own thing, but don't have an idea yet."
My mind translates that into:
"I have an extremely clear idea of what I want to do, I also am extremely aware that any ideas that could possibly be thought of on company time may be claimed as the intellectual property of my employer."
This may or may not be the case of Gabor and Google as I can't speak to their intellectual property arrangements, but there is nothing quite like a very public statement that you are intellectual property unencumbered to start things off.
That would always be the safe response. There isn't a whole lot to be gained for the employee by telling the company why they are leaving. Some folks like to 'get off their chest' all their grievances in an exit interview but even that isn't advisable. The fact that you're leaving those 'losers' behind is great now, but later you might need to work with them again and well, lets just say that debt accumulates.
On that note, do yourself a favor, write the resignation letter (on your own computer, turn the internet off) that you really want to write. Add all your gripes, hatreds, and problems with the place. Tell the true story of your leaving and why your soul was at risk if you stayed any longer.
now print it out and delete the electronic copy
Now you can turn back on the internet (since you now don't have to worry about accidentally sending something) and write a one paragraph resignation saying what your last day is and how you love everyone. Print and hand that in.
Every time in the future when you wonder why you left, read the print out of your rant.
While I've never given the boat-burning scorched-earth exit interview, I have given professional-yet-candid opinions on the occasion that I cared enough to do so and had someone who cared enough to listen.
That is theft, pure and simple, same as if you incur a billion dollar thought on your own and fail to immediately declare it to the IRS and pay proper taxes on it. Not.
(This is satire. Obviously you don't owe taxes just by having a valuable thought, and by extension my point was that obviously a thought alone can't actually belong to an employer if you haven't even acted on it. Of course an employer can still bully you to hand over what you produce outside the company, but that's simply not right, whatever your contract might say. No one is a slave or can sign a meaningful contract to that effect, in my personal opinion. Judges may differ.)
Not always the case. Some of us simply need a break to clear our heads. I have absolutely no idea what I'll do next, and I like it that way. I know it will come to me in time.
So it sort of happened once 10 years ago and it involved actual that was probably worked on during the work day and on the company's equipment. Whereas here we've been referring simply to possibly having and idea ready to work in soon after deserting the company. I'm not sold.
Having done this myself recently, I'd say it's more like:
"I have an extremely clear idea of what I really want to do, but I haven't even built a prototype yet, so what I'm working on 2 weeks from now might well be one of the dozen other completely different ideas I've got. And I don't really want to be forever remembered as The Guy Who Left To Start Dumb Project XYZ."
There's value to oneself in declaring publicly that you are mentally unencumbered.
But there's negative value in telling people what you're really up to. I first heard it from Derek Sivers who said that Repeated psychology tests have proven that telling someone your goal makes it less likely to happen [1]. And it actually makes sense...
So it might be better to engage in some little public self-embarassment than go around telling what you're up to...
interesting, thanks. Hemingway said the same about writing: he'd work until he knew where he would pick up the next day. Then he could stop. In between, he'd not want to talk about what he'd worked on or planned to next sessions as that would get in the way of actually doing it.
I've had success writing code this way. If there's a wall, talking about it helps; if there's a flow, shutting up and letting your subconscious at it helps more.
That is not the case. I have a couple of spaces I'm interested in, and will start prototyping in the next few weeks. But I don't have a concrete and fixed idea.
This blog post in not a CYA tactic but was triggered by the comment I read on Hacker News, and my belief that leaving to do one specific idea is in most cases the wrong way to do it.
Well I'm glad my comment got you thinking. You really answered my question with the last line of your post: "it's just that I think I can execute much better than them." In a competitive space where so many products are doing the same thing execution becomes everything. You obviously have the experience and I wish you the best in whatever idea you pursue.
One thing I do have to mention is that I bring a different perspective when I made my comment. I'm an electrical engineering student (graduating in two weeks) with only work experience from JPL and SpaceX. If I take up a job at any of those places after graduation, I would have to think more than twice before I ever contemplate leaving to pursue an idea. Then again my generation is full of risk takers, and I think I'm one of them.
Yes! Thanks a lot for the comment amirmansour. It certainly was the source of much controversy on here :-)
One interesting change from last time I started a company (in the nuclear winter of 08/09) is that angel and seed funding is much easier to come by. That leads to a situation where there are lots of companies in almost every interesting space. Thus being able to out-execute them becomes even more relevant.
For web and mobile products, the act of building or prototyping is where you do most of the learning. I believe that you can't come up with great ideas in a vacuum. You build some mocks, you write some code, and you discover a bunch of things that you could improve on.
Please tell me this sort of thinking isn't typical of Google's product organization.
You do learn as you execute, but this just moves you towards a local maximum - a local maximum that might not be that exciting, if you've badly misunderstood the people you're building for.
You do the most learning prior to starting, when you listen to the people you plan to build for and get to know them intimately - their problems, aspirations, values, and ethics.
Getting to really know the worldview of your targets is how you come up with the direction that leads toward the global maximum, and it's all done before you write a line of code.
Good point, but taken to the extreme, if you listen to people too much they'll tell you they want a faster horse.
You have to combine the aspirational thinking of "let's build a self-driving car" with the day-to-day grind of actually building it. Without the first, you end up in the local maximum you describe, without the latter you land in the "talk is cheap" box.
My post is about the fact that you can in fact leave your job and go through a phase of prototyping one product after another, and deliberately testing each idea for not just how great it sounds but how buildable it is.
I'm no expert in entrepreneurship, but have to believe this is true for the most part. Before leaving, most people probably have a vision of something to start on -- but without (1) building the real product, (2) dealing with the details, and (3) interacting with customers, its only a vague vision. Along the way, things will change, people pivot, founders get new ideas, etc.
Here's another way to think about it: there is probably a good reason behind YC accepting people without an idea.
Ideas are dime a dozen. Give me some smart friends to brainstorm with and we'll come up with several great ideas today that will be $100M+ businesses in a few years.
Those are some big words to swallow. I doubt he could do that but I'd love to be proven wrong. If anything, what he wrote comes down to the old adage of "easier said than done".
53 comments
[ 3.1 ms ] story [ 493 ms ] threadYep. There's a quote from Bob Parsons (of GoDaddy fame) that goes something like (paraphrased roughly) "Don't be afraid to enter a crowded market, just be better than everybody else."
This is kind of our mindset at Fogbeam Labs, where - depending on how some details shake out - we'll be competing with everybody from Jive Software, Yammer, Alfresco and Broadvision (among others), to Salesforce, IBM, Cisco and Microsoft.
But that's fine. We have an angle to pursue that we think will let us make some inroads against those guys, and we're spoiling for a good fight. Heck, as far as I'm concerned, those guys should be sleeping with one eye open. :-)
In one case the great new idea is a new product, in the other case the great new idea is how to be better than the competition.
Which boils down to either extreme failure or extreme success and little in-between...
Elon Musk couldn't have started SpaceX and Tesla right out of school, he had to build up skill, credentials, and connections first.
For your first startup, make something awesome but try to fly people to Mars. A small acquisition or a business with nice revenue might be the better goal.
I'm mostly leaf reading, but it seems to me you are recommending a lean startup / lifestyle business as a first startup. A less ambitious endeavour than say, Google or Amazon, with the primary goal of acquiring skills, small but safer revenue, a stepping stone to the next, bigger, enterprise. Very inspiring, and something young engineers should think about.
BTW, I restate my request from my unsolicited email (hope you don't mind): please share some startup ideas, even if execution is everything, some of us younger, less experienced engineers find brainstorming and finding ideas harder, and would really appreciate some inspiration.
In my own life, had I chosen the startup over grad school, I would easily be worth $100M+ right now. I had a ground floor opportunity I idiotically passed on because I was young, stupid, and too idealistic. So I see his point.
That said, I shudder at the thought of hiring a programmer who struggled with calculus and nothing assures me better on that axis than straight As on math and computer science courses above and beyond requirements for their major.
Individuals vary, and one of the best coders I know got a 2-year associates degree out of community college, but this is a really nice foot in the door for me when I don't know the person already.
Elon Musk couldn't have started SpaceX and Tesla right out of school, he had to build up skill, credentials, and connections first.
I agree with your point, but "your first startup" is not necessarily "right out of school." In my own case, I'm working on my first startup, after about 15 years in the IT world. And it's been those years of IT experience that has led to me (and to some extent, my co-founders, both of who are younger than I) to develop those beliefs where we feel like we believe something that everybody else doesn't.
For your first startup, make something awesome but try to fly people to Mars. A small acquisition or a business with nice revenue might be the better goal.
Yeah, we're thinking IPO, and eventually acquiring Microsoft. If you're going to dream, dream big, I always say. :-)
I meant: "For your first startup, make something awesome but don't try to fly people to Mars."
I didn't even realize it was Peter Thiel who said that, but I embrace that mindset as well. We have a document in our repo called "What we believe" that is where we catalog the things that we think about the world, that we suspect most everybody else would disagree with us on. Looking at that periodically is a valuable tool to help us focus on how we're going to be different.
But that does not mean all ideas are equally cheap and worthless. They aren't. Good ideas are like the DNA of a product or service. Separating the wheat from the chaff is the purpose of executing and iterating. And don't confuse ideas with how people initially attempt to express what's rolling around in their head. People often don't know how to really effectively communicate such things. Iteration is often just an exercise in trying to figure out how to properly express the underlying idea. That is what pivoting is often about.
Good ideas are not a dime a dozen. They aren't all that common. They are testable, the way Einstein's Theory of Relatively is provably a more accurate concept than Newtonian gravity.
I run into this all the time when trying to talk about how I got myself well. It is a proven concept in some sense as it worked for both me and my sons. But that doesn't automatically give me the ability to effectively convey it to other people. I continue to struggle with that piece of it.
Part of what I do outside work is help out with housing cooperatives in the UK. We put ordinary not very rich people in charge of their homes. They become the landlord. Targets include < 4% void rent in any year, < 28 days property vacant &c.
Real decision making and budget responsibility (landlord liable for repairs) works wonders. You can see people grow taller, and many now have supervisory jobs...
I do envy the rental sector in US. UK has issues with artificially high prices (my opinion) preventing a free market rental sector from emerging.
The "boys" came home, often had high savings and no kids, "Rosie the Rivetter" was encouraged to quit her job so a man could have it, many women did so quite happily so they could have kids and the fulltime mom thing. Prior to the war, we were deep in The Great Depression and many people moved every thirteen months to take advantage of one free month's rent if you could pay rent in full and on time for twelve months straight. People were doubled up in small homes, etc.
Levittowns were born to meet the sudden demand for single family homes. The nation created finance mechanisms to facilitate the birth of these instant suburbs. We are still living with the legacy. It can be extremely difficult to arrange financing for anything other than a single family home.
The reason ideas are so cheap is that they are only the reflection of an action, not the action itself. Often folks take what they 'want' and then ponder a way to 'get' that and poof! idea. Like "I 'want' to work on my car's engine computer, all I need to do is take the engine out and mount it on an engine stand and hook it up to my laptop!" Simple idea, remove the engine and instrument it, but actually doing that is a lot of work. (doable, just a lot of work).
But this can also be a useful tool. Ideas that aren't a lot of work to implement but seem to have high value are usually useless. And if you can't tell why they are useless then it is sometimes worthwhile to pursue them enough to understand where the real work is.
But take something that is pretty much pure software. Let's take electronic publishing briefly. You have no doubt read a paperback book, you've no doubt read one (or looked at one) on an e-reader. All words, no pictures, just chapter breaks, perhaps the occasional font change to represent internal vs external dialog. Why can't I cut the spine off my old paperback, feed it into my Fuji Scansnap1500, and have it drop a .epub file that works on a kindle/nook/ipad/whatever ? The scansnap will give me a nice PDF of that book, even searchable. Caliber will convert it into a horrendous epub file. Why isn't this a no-brainer? I'd pay $20 for that software, no problem.
Idea, no execution. Marginal return on investment (while I would buy it if it worked well, there is no assured market here) There are bodged together ad-hoc solutions out there. Clearly there is some demand but as Gabe called out, you need passion. You need to want to build this because you want it to exist. And that, that passion, is not there.
I tell armchair politicians, hey run for city council. Solve problems. If you are good at it people will give you bigger and bigger problems to solve. If you aren't you'll both have an appreciation for what politicians are up against and a compassion for what they do get done. Not nearly as fun though. At a Tom Campbell said at a fund raiser once, its hard to appreciate the job from the outside.
Newton had to write a book to get his ideas out there and accepted.
The company that this really brings to mind is DuckDuckGo. They're competing for the search market with Google and Microsoft--two of the biggest competitors you could possibly face. But search is huge and if you can claim even 0.1% of that market, you've got more than enough users to justify your existence.
"I'm leaving to work on my own thing, but don't have an idea yet."
My mind translates that into:
"I have an extremely clear idea of what I want to do, I also am extremely aware that any ideas that could possibly be thought of on company time may be claimed as the intellectual property of my employer."
This may or may not be the case of Gabor and Google as I can't speak to their intellectual property arrangements, but there is nothing quite like a very public statement that you are intellectual property unencumbered to start things off.
now print it out and delete the electronic copy
Now you can turn back on the internet (since you now don't have to worry about accidentally sending something) and write a one paragraph resignation saying what your last day is and how you love everyone. Print and hand that in.
Every time in the future when you wonder why you left, read the print out of your rant.
Never burn the boats as you might need them!
I've never heard of this actually happening. Does it?
"I have an extremely clear idea of what I really want to do, but I haven't even built a prototype yet, so what I'm working on 2 weeks from now might well be one of the dozen other completely different ideas I've got. And I don't really want to be forever remembered as The Guy Who Left To Start Dumb Project XYZ."
There's value to oneself in declaring publicly that you are mentally unencumbered.
So it might be better to engage in some little public self-embarassment than go around telling what you're up to...
[1] http://www.ted.com/talks/derek_sivers_keep_your_goals_to_you...
I've had success writing code this way. If there's a wall, talking about it helps; if there's a flow, shutting up and letting your subconscious at it helps more.
This blog post in not a CYA tactic but was triggered by the comment I read on Hacker News, and my belief that leaving to do one specific idea is in most cases the wrong way to do it.
One thing I do have to mention is that I bring a different perspective when I made my comment. I'm an electrical engineering student (graduating in two weeks) with only work experience from JPL and SpaceX. If I take up a job at any of those places after graduation, I would have to think more than twice before I ever contemplate leaving to pursue an idea. Then again my generation is full of risk takers, and I think I'm one of them.
One interesting change from last time I started a company (in the nuclear winter of 08/09) is that angel and seed funding is much easier to come by. That leads to a situation where there are lots of companies in almost every interesting space. Thus being able to out-execute them becomes even more relevant.
Please tell me this sort of thinking isn't typical of Google's product organization.
You do learn as you execute, but this just moves you towards a local maximum - a local maximum that might not be that exciting, if you've badly misunderstood the people you're building for.
You do the most learning prior to starting, when you listen to the people you plan to build for and get to know them intimately - their problems, aspirations, values, and ethics.
Getting to really know the worldview of your targets is how you come up with the direction that leads toward the global maximum, and it's all done before you write a line of code.
You have to combine the aspirational thinking of "let's build a self-driving car" with the day-to-day grind of actually building it. Without the first, you end up in the local maximum you describe, without the latter you land in the "talk is cheap" box.
My post is about the fact that you can in fact leave your job and go through a phase of prototyping one product after another, and deliberately testing each idea for not just how great it sounds but how buildable it is.
Here's another way to think about it: there is probably a good reason behind YC accepting people without an idea.
Those are some big words to swallow. I doubt he could do that but I'd love to be proven wrong. If anything, what he wrote comes down to the old adage of "easier said than done".