News for hackers. Stuff that matters.
Re:How not to die by ujjwalg on Monday August 17, 2009 @06:17PM (#769248)
I am finishing up my PhD and I have a very direct analogy to what PG mentions.

There are multiple research groups with different success rates and what I have noticed is groups with extremes are not doing that well. By extremes I mean professors who try to micromanage and get a daily update and professors who do not care at all. In my group, we have a biweekly meeting for everyone to present what they did in 2 weeks time which makes all of us work hard but on our own terms. I think incorporating similar things in education/research will be very helpful in the long run.

Good article PG!!

Re:How not to die by mrshoe on Monday August 17, 2009 @05:58PM (#769205)
I'd read that before, but I think I needed it today.

Maybe someone should set up a cron job to submit that talk to HN every 3 months.

Re:How not to die by jpwagner on Monday August 17, 2009 @06:08PM (#769228)
Or to randomly choose one of PG's essays every month.

This particular one was last posted almost two years ago.

Re:How not to die by mrshoe on Monday August 17, 2009 @06:20PM (#769260)
Startup idea: present the user with 8 adjectives and ask them to pick the one which best describes their mood, then give them an optimal article to read as they start their day.

We'd all be more successful if we could capitalize on our good moods and not let our bad moods drag us down for long.

Re:How not to die by Zev on Monday August 17, 2009 @06:51PM (#769328)
I don't know if that would be a good startup idea, but it would make for a nice side project to create on a weekend.
Re:How not to die by wensing on Tuesday August 18, 2009 @12:14AM (#769727)
Which is exactly how many startups are born.
Re:How not to die by kirubakaran on Monday August 17, 2009 @06:36PM (#769300)
I was doing something like this that I called "the ugh2yay project" (a webapp). I abandoned it after some time as I thought there is no real need for it as people will just talk to their friends to feel better. Do you think it will be useful? It had features like ability to save stuff when you are happy (say a hilarious lolcat) to be shown to you when you are not so happy etc...
Re:How not to die by TrevorJ on Tuesday August 18, 2009 @10:00AM (#770441)
I really like that. Why don't you do that, but have it parse somebody's status updates for the keywords?
Re:How not to die by andreyf on Monday August 17, 2009 @09:44PM (#769564)
Two years? Has it really been that long? I'm old :( and in need of moving on to new memes - any suggestions guys?
Re:How not to die by edw519 on Monday August 17, 2009 @06:18PM (#769252)
I have cut & pasted all of pg's essays onto my Palm pilot. Along with my bridge program and a bunch of other great content, I never have an excuse not to learn something.

Whenever I sense that I'm getting stagnant, I always go back and reread old faithful, "Good and Bad Procrastination".

http://www.paulgraham.com/procrastination.html

(When someone invents an ebook reader that plays bridge, I'll make the switch.)

Re:How not to die by iamwil on Monday August 17, 2009 @07:02PM (#769342)
There's a link to HN library at the footer. You can browse that at your own leisure.
Re:How not to die by ErrantX on Monday August 17, 2009 @05:51PM (#769185)
I've started 4 companies. 3 died (but not badly) one petered out and I sold the core tech for a pretty decent figure (so I call it a semi-success even though it was the least likely to succeed)

I dont know: I came to HN after I began hacking and doing startups (I never even called them startups till I got here) and my assumption was that 99% failed anyway as a part of the process of learning.

I figure you have to either be super lucky (and make a Twitter, Facebook or Google hit) or you have to fail hard a few times before eventually succeeding. I chatted to some London VC's the other week and they said they trusted founders from previously failed startups a lot more than newbies - simply because they had learned many of the hard lessons and wouldn't then make them at the crucial moment their company could be worth something 6 figures or more.

Re:How not to die by jacquesm on Monday August 17, 2009 @06:12PM (#769237)
I don't know anybody that 'scored' that didn't fail a couple of times first. The ones that hit a home-run the first time out are the ones that you hear about all the time but they're the exceptions.

It's not a rule, but hard work definitely pays off - most of the time.

Re:How not to die by chrischen on Monday August 17, 2009 @06:10PM (#769232)
> I dont know: I came to HN after I began hacking and doing startups (I never even called them startups till I got here) and my assumption was that 99% failed anyway as a part of the process of learning.

Same here I've been doing startups since 13, didn't even know about the dot-com boom bust at the time.

Re:How not to die by jpwagner on Monday August 17, 2009 @06:06PM (#769222)
VC's want the choice of whether or not to invest in you.

So naturally they trust your attributes alot more than others' when they are talking to you.

Re:How not to die by Estragon on Monday August 17, 2009 @07:50PM (#769405)
So, how are octopart and wufoo doing? (They both seem to be around, still. Have they been in the news, lately?)
Re:How not to die by pg on Monday August 17, 2009 @08:37PM (#769464)
They're both doing well, actually. Both are profitable. Wufoo remarkably so, but they've had longer.
Re:How not to die by prakash on Tuesday August 18, 2009 @06:08AM (#770129)
Quick question: Do you make a note of how successful a startup would be when you fund it, and then revisit those notes a year or 2 down the line? How do you make sure you have the right feedback mechanisms in place?
Re:How not to die by Estragon on Tuesday August 18, 2009 @07:11AM (#770190)
wufoo's demo video peters out just where it gets interesting to me, and it confuses me about their value proposition: once you've got the form made, what do you do with it? Do they just give you the html and point you to where the URL for the CGI ought to go?
Re:How not to die by unfoldedorigami on Tuesday August 18, 2009 @12:18PM (#770770)
You're right in that creating a a form in Wufoo is only the beginning. At that point, you can start sharing the link to the form or embedding it on a web page to start collecting data, leads or even payments. You can do all this from our Code Manager: http://wufoo.com/docs/code-manager/

Built into every form is also the ability to subscribe to the incoming data in real time to your email inbox, mobile device via SMS or your RSS Newsreader.

We also have a robust report building feature that can help you visualize and export your data so you can better understand the information you've collected.

You can learn more about all these different steps in our other video tutorials at http://wufoo.com/docs/getting-started/

Re:How not to die by Estragon on Wednesday August 19, 2009 @06:40AM (#772442)
Thanks for the explanation.
Re:How not to die by rms on Tuesday August 18, 2009 @11:08PM (#771959)
Wufoo is the easiest to use backdoor to a database out there.
Re:How not to die by diN0bot on Monday August 17, 2009 @09:26PM (#769533)
i asked the MIT Electronics Research Society http://mitERS.mit.edu if anyone used octopart and what they thought. here are the responses so far (sent to public list not just me personally):

"It's a great idea, it's useful, but it's a work in progress and not yet a replacement for searching distributor websites." -henry

"If you know what you're looking for and Octopart has incorporated it, you'll find it in seconds flat, and it's fantastic. Unfortunately, it seems spotty in terms of catalog completeness, and it is not as comfortable as browsing/guided search on, say, DigiKey or Mouser, so it gave me mixed results." -darwin

Re:How not to die by andres on Tuesday August 18, 2009 @12:15AM (#769729)
thanks for asking mitERS for their input. please let everyone know to email me with suggestions on how we can help them search for parts ([email protected]).
Re:How not to die by cmos on Monday August 17, 2009 @11:12PM (#769672)
I agree and disagree. Octopart is very revealing if you do low volume manufacturing had have been caught in the 'digikey convenience turbine'. Like any search engine, it's useless if you don't have the part. But there are a lot of core parts to a generic circuit board listed, so I end up using it quite a lot.
Re:How not to die by gjm11 on Monday August 17, 2009 @08:35PM (#769459)
I fear there may be less to this than meets the eye.

It reminds me of a foolproof recipe for (human) immortality, which I think I read in a book by Raymond Smullyan. It goes something like this: every day you say "I will say this sentence again tomorrow". If you always keep that promise, then you are assured of never dying.

The point, of course, is that "if X, then Y" doesn't mean that by doing X you can ensure Y, even if on the face of it X is something under your control.

Similarly: yes, indeed, if your company continues interacting with PG and other Y-Combinator-funded companies, then it will not die -- because if it dies, it will no longer be able to have those interactions.

Re:How not to die by pg on Monday August 17, 2009 @09:49PM (#769570)
Similarly: yes, indeed, if your company continues interacting with PG and other Y-Combinator-funded companies, then it will not die -- because if it dies, it will no longer be able to have those interactions.

But that's false. People don't lose the ability to email us once their company dies.

Re:How not to die by gjm11 on Tuesday August 18, 2009 @11:49AM (#770681)
The companies no longer have that ability, even though the individuals do.

Don't get me wrong; I think "keep doing something and stay in touch" is an excellent piece of advice for someone doing something difficult where failure is strongly correlated with losing motivation.

But I wonder whether you're overestimating the causal role of demotivation here. Toy model: a startup has some kind of "health score" which varies according to exogenous events; "motivation" is some function of health and d(health)/dt, which is small when the startup is unhealthy and getting healthier; the startup dies if its health goes negative. In this model, motivation has no causal role, it's a pure epiphenomenon; but I bet you'll find that startups that die typically have very poor motivation shortly before their demise, and that poor motivation is a good short-term predictor of failure.

Of course, this is only a ridiculous toy model; in reality motivation clearly does have causal effects of its own, but it's also clearly affected by outside causes that themselves have an effect on the success of a startup. An observed correlation between demotivation and death is only weak evidence that demotivation is the underlying problem more than it's a symptom of having no customers, no money, technical problems you can't solve, etc.

(Doubtless you have more evidence than that correlation, having seen lots of startups succeed and fail.)

Re:How not to die by keyist on Monday August 17, 2009 @07:56PM (#769415)
Do not go gentle into rewrite / Rage, rage against the dying of the site.

/me ducks

Re:How not to die by zaidf on Monday August 17, 2009 @06:23PM (#769273)
This ranks right up there with Steve Job's Stanford speech.

It was even more powerful when pg actually delivered it. I wish a video from that night's dinner existed.

Re:How not to die by iamwil on Monday August 17, 2009 @07:05PM (#769347)
The part I found most useful is where he quotes Paul Buchheit, and about pleasing that first small set of users. He doesn't tell you where to find them, but at least once you've found them, you can hang on to that through thick and thin to protect your morale (given that morale is what motivates founders not to quit)
Re:How not to die by iamwil on Monday August 17, 2009 @05:58PM (#769202)
the previous discussion when it was last submitted.

http://news.ycombinator.com/item?id=48294

Re:How not to die by swombat on Monday August 17, 2009 @06:07PM (#769223)
I thought it was relevant (and far enough away time-wise) for a repost, given the news from SnapTalent today.
Re:How not to die by zackattack on Monday August 17, 2009 @08:05PM (#769425)
Re:How not to die by AlphaEvolve on Tuesday August 18, 2009 @12:53AM (#769765)
This post really came at the right time for me. Such a moral boost. I created and is currently operating a startup in orange county, ca. I am also in the middle of a divorce. So to sum it up: Though times. Personal advise: If you believe your idea can provide value to clients, end-users, or to the community as a whole: Never ever give up. It will eventually pay off.
Re:How not to die by paul9290 on Monday August 17, 2009 @06:20PM (#769259)
Sorry I think the part about starting up other things is ill-advised.

TWITTER WAS A SIDE PROJECT!

Re:How not to die by nostrademons on Monday August 17, 2009 @06:51PM (#769327)
Side-projects are always confusing because the answer to "Should you do a side project?" is always "Yes - if that side project will eventually become your main project." Unfortunately, it's almost never possible to tell that a priori. So it always comes out as "You should do side projects, except in the cases where you shouldn't", which isn't terribly helpful.
Re:How not to die by aditya on Monday August 17, 2009 @06:23PM (#769274)
When they knew Odeo didn't have enough traction: http://www.stubbleblog.com/index.php/2009/06/the-real-lesson...
Re:How not to die by steveplace on Tuesday August 18, 2009 @09:10AM (#770343)
In addition, once ev realized the side project was better than his current startup, he shut down the original, gave back all the money to his investors, and raised for twitter.
Re:How not to die by martincmartin on Monday August 17, 2009 @07:01PM (#769340)
Am I the only one that thought PG was going to come out as an extropian?
Re:How not to die by nopassrecover on Monday August 17, 2009 @08:08PM (#769427)
I thought it was customary to include the year at the end of old/reposts.
Re:How not to die by helveticaman on Monday August 17, 2009 @09:11PM (#769507)
What if the economy is a shitstorm? I'm not talking about right now, either; I'm talking about an Argentina-style, complete meltdown. What then? Can it become implausible to do well?
Re:How not to die by pg on Tuesday August 18, 2009 @01:05AM (#769775)
Obviously there's a point at which it could, yes. If you lose civil order, for example, you lose most of the economy. That's what really destroyed the economy of the Roman Empire. There were of course deeper underlying causes, but the proximate cause was plundering and brigandage.
Re:How not to die by elai on Tuesday August 18, 2009 @01:11AM (#769787)
Well argentina isn't the only economy in the world ;)
Re:How not to die by wastedbrains on Tuesday August 18, 2009 @10:33AM (#770517)
I read this whenever we are currently on a down swing on the roller coaster that is a startup.
Re:How not to die by sound2man on Monday August 17, 2009 @05:47PM (#769179)
Work hard and never give up. Got it.
Re:How not to die by aditya on Monday August 17, 2009 @05:49PM (#769182)
or, work hard and know when to give up. much harder.