Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

PG's Rarely Asked Questions

paulgraham.com

Post reply on HN
Print view View original

31–40 of 221 posts

Re: PG's Rarely Asked Questions

#31

Post by tspiteri »

edanm wrote: post #6
muhfuhkuh wrote: post #4

" I want to start a startup, but I don't know how to program. How long will it take to learn? I would guess a smart person can learn to hack sufficiently well in 6 months to a year." Hmm... interesting take considering the source, especially when contrasted with the general mentality that programming and software development is the finest of all trades and takes a near preternatural mastery only found elsewhere in cl…

Note: he means "good enough", not good. Which brings to mind patio11's constant reminders, that running a successful software business has very very little to do with programming well. Which means that pg thinks that 6 months is the level you need to be at to be able to start a startup. This seems intuitively right to me, especially these days.

Not exactly the same thing, but this does remind me of a quote from I'm Feeling Lucky: The Confessions of Google Employee Number 59:

"I didn't trust Larry and Sergey as coders," said Craig Silverstein. "I had to deal with their legacy code from the Stanford days and it had a lot of problems. They're research coders: more interested in writing code that works than code that's maintainable." According to Jeff, one of the quirks in the early Stanford version of Google was that when something unusual happened the program would print out an error message without any explanation. The message read simply, "Whoa, horsey!"

Re: PG's Rarely Asked Questions

#32

Post by wallawe »

NY_Entrepreneur wrote: post #30
larrys wrote: post #18

Earlier quoted context omitted.

"how to make an A in Women 101-102!" My initial reaction to this was to agree with the statement. And I still sort of do. But I think women (or men) can be a big distraction if you are trying to create something. That being said the correct woman or men can help you achieve as well if the relationship is stable. The statement "behind every great man" there is a ... is basically true from my experience.

"Behind every great man is a woman"? Maybe! In the case of FDR, Ike, Patton, and McArthur, I suspect it was more their mothers than their wives! Few men have such helpful mothers! Then there is your "if the relationship is stable". Well, yes, and 'effective'! So, and especially for the huge fraction of men who don't have mothers as helpful as FDR's, how to get such a 'relationship'????!!!! May I suggest Women 101-102…

I got a good laugh out of this. Thanks

Re: PG's Rarely Asked Questions

#33

Post by bergie »

edanm wrote: post #12
_delirium wrote: post #10

Earlier quoted context omitted.

In fact the usual advice is that you shouldn't be doing something technically virtuosic in a startup. If your startup's critical path includes writing something as technically impressive as the Glasgow Haskell Compiler, you might have trouble... Though there are a few people around here who've built software businesses around things that are also nontrivial technical advances; cperciva comes to mind.

I know that's a sentiment that's thrown around here, but I think it's simplistic. I live in Israel, where a very common path to an Exit is to build some sophisticated technology, then sell it after a few years to a larger company. So it is a very possible path you can take. It has its own advantages and disadvantages over the much more visible, much more HN-friendly, "Consumer Internet Startup".

The exit and valuation models for these two approaches are also different. In the former, some company buys you because they want to use your technology in their own products, in the latter, some company buys you to get your customers/users.

Re: PG's Rarely Asked Questions

#34

Post by dforeman »

muhfuhkuh wrote: post #4

" I want to start a startup, but I don't know how to program. How long will it take to learn? I would guess a smart person can learn to hack sufficiently well in 6 months to a year." Hmm... interesting take considering the source, especially when contrasted with the general mentality that programming and software development is the finest of all trades and takes a near preternatural mastery only found elsewhere in cl…

I think that question was included ironically- I.e., he's only been asked a handful of times. His sample size reflects the dearth of "learning how to program to build a viable webapp" stories and comments here. I don't even know if noob questions are accepted here- I've never seen one. I don't post them because I feel I should be finding answers on google rather than degrading the quality of this board.

The decision to learn enough programming to build what I want stemmed from the realization that understanding how my webapp will work at the technical level significantly refines my vision for the product- bringing it much closer to reality. Whereas before I was talking to recommendation engine people because I wanted "recommendation features like Hunch," now I'm working in PHP/SQL and getting exactly what I want. My current strategy is to talk to the CTO's and creators of similar webapps to understand their technology and language choices (e.g. GAE or MAMP, SQL or Solr), go as far as I can on my own, and then work with a tutor for an hour or two to get momentum and correct my architecture mistakes.

Re: PG's Rarely Asked Questions

#35

Post by aangjie »

tspiteri wrote: post #31
edanm wrote: post #6

Earlier quoted context omitted.

Note: he means "good enough", not good. Which brings to mind patio11's constant reminders, that running a successful software business has very very little to do with programming well. Which means that pg thinks that 6 months is the level you need to be at to be able to start a startup. This seems intuitively right to me, especially these days.

Not exactly the same thing, but this does remind me of a quote from I'm Feeling Lucky: The Confessions of Google Employee Number 59 : "I didn't trust Larry and Sergey as coders," said Craig Silverstein. "I had to deal with their legacy code from the Stanford days and it had a lot of problems. They're research coders: more interested in writing code that works than code that's maintainable." According to Jeff, one of…

7 years ago i would have called myself similar.Though i was just graduating and could have hardly called it research. But 3 years working in the industry with Big blue and i see i have got OCD with maintainability. Heck, i wouldn't even have noticed, if i hadn't quit and went to do a master's. But 4 years since, and i still haven't quite undone the OCD, mitigated yes, but not back to the get shit done passion....I keep working on it...so, i will qualify that 6 months to start a startup, with something about the type of product you aim to build. - May be something along the lines of the impact you expect to make with it.

Re: PG's Rarely Asked Questions

#36

Post by jackfoxy »

What should I read to learn more about history?

I used to consider my knowledge of history better than at least 95% of the population, but while reading The Decline and Fall of the Roman Empire I realized how sketchy my view of history really was. So at the ripe old age of 35 I set off on a course of study centered on two series of books, The Story of Civilization, by Will and Ariel Durant, and Timeframe, a Time-Life series focusing on a timeframe in human history and what was going on in all parts of the world inhabited by man: lots of pictures and of course superficial, but it painted in a lot of gaps I othewise would have never covered. The Timeframe series starts much earlier than the Durant's, but once both series were in sync I would read the books in both series for an epoch, as well as at least two other books, either written in the era or about the era, drawing mostly from science, culture, and biography. For instance I read all the books of Euclid, Newton's Optiks and Principia (I slogged through the Motte transaltion before the first modern English translation became available), The Wealth of Nations, Shelby Foote's 3-volume history of the Civil War, and The Origen of Species. (It's real easy for me to spot folks who spoot-off about Wealth or Origen who have not actually read the books.) My program culminated with Tragedy and Hope, which being such an inflammatory work, I did not trust to read without the full background of history. The process was like watching Western Civilization unfolding.

Now for the unintended consequences: I became a bore at cocktail parties. I wanted to talk about the ideas in the fascinating book I was reading. I used to love arguing politics. Even with my prior knowledge it was hard enough finding opponents who would engage in rational discourse, now it is impossible. It's been so long my debating skills have totally gone down the tubes. The sad thing is I believe my problem is really society's. Political correctness (among other problems) in academia, has produced a generation of intellectually crippled intellectuals; and the entertainment industry, including the 24-hour news cycle as entertainment, has just stupefied people. I fear for democracy and republican government.

Re: PG's Rarely Asked Questions

#37

Post by chernevik »

ofca wrote: post #7
muhfuhkuh wrote: post #4

" I want to start a startup, but I don't know how to program. How long will it take to learn? I would guess a smart person can learn to hack sufficiently well in 6 months to a year." Hmm... interesting take considering the source, especially when contrasted with the general mentality that programming and software development is the finest of all trades and takes a near preternatural mastery only found elsewhere in cl…

I am also fueled by his answer here. I guess that learning parameter depends on your teacher - the better the teacher, the better and faster you learn. I just started learning HTML, CSS, jQ, PHP recently , and I found myself advancing at surprisingly high pace (now watching jQ new_boston tutorials on youtube). Main problem for me is finding good material to learn from, since I only learn from the internet. You have a…

It has taken me considerably longer than a year. But I might not be smart. And I've tried to do it on my own, rather than taking a job (see prior sentence).

Caveat: I'm not a developer, so my comments about learning how to program should be taken with considerable skepticism. But I have some opinions based my effort to learn.

For me, programming syntax is less a problem for me than poorly organized code, poor grasp of problems and poorly thought out solutions. Design is a subtle and important topic that does not lend itself to the shorter coverage typical of internet material. Book-length immersion in the concepts is only a beginning to an understanding. So I think your "internet only" restriction is a severe constraint.

But it isn't as if there are many design books for beginners. These generally target people who've mastered some languages and know the design-problem space. "Pragmatic Programmer" was a big help.

I have found test-driven development an enormous help. It feels cumbersome, but you always know what problem you're trying to solve. "Test Driven JavaScript Development", Johansen, is an excellent overview of the approach. I'm also trying to work through Beck's "Test Driven Development By Example" doing the exercises in Ruby, but it's a side project that hasn't gotten much attention.

I repeat: I'm not a developer. Hopefully this is thought-provoking, but I'm no authority.

Re: PG's Rarely Asked Questions

#38

Post by enko »

edanm wrote: post #6
muhfuhkuh wrote: post #4

" I want to start a startup, but I don't know how to program. How long will it take to learn? I would guess a smart person can learn to hack sufficiently well in 6 months to a year." Hmm... interesting take considering the source, especially when contrasted with the general mentality that programming and software development is the finest of all trades and takes a near preternatural mastery only found elsewhere in cl…

Note: he means "good enough", not good. Which brings to mind patio11's constant reminders, that running a successful software business has very very little to do with programming well. Which means that pg thinks that 6 months is the level you need to be at to be able to start a startup. This seems intuitively right to me, especially these days.

> Which brings to mind patio11's constant reminders, that running a successful software business has very very little to do with programming well

That may be appropriate for patio11. Not to demean the guy but let's be clear, he runs an online bingo service. It's my understanding that he makes much more money from consulting gigs leveraged from this modest endeavour.

Rest assured that if your ambition is to write the next Dropbox or, hell, any YC company, at least 50% of your founding team will need to be very good programmers indeed.

Re: PG's Rarely Asked Questions

#39

Post by chernevik »

siglesias wrote: post #20

Re: philosophy, I think understanding the difference between representing the world in language and representing it in logic is critically important to programmers. It has implications for natural language processing as well as artificial intelligence. I personally recommend to anyone endeavoring to understand Wittgenstein's transition from Tractatus Logico Philosophicus to Philosophical Investigations . In fact, Goo…

I disagree with pg about the use of philosophy. If nothing else, Plato should disabuse anyone of the notion that what's obvious to them isn't really that solid. Aristotle at least demonstrates your point about just how hard it is to think systematically about this stuff.

The humility gained from those encounters would do a great deal to open people's minds that "there's more on heaven and earth than are dreamt of in your philosophy". Maybe people would less adamant, and pay more attention to why they hold their opinions. If everyone did that, we'd have far more productive discussions.

Re: PG's Rarely Asked Questions

#40

Post by dgreensp »

>I never had to manage anyone in our startup, even though I was the president. The other hackers were my peers, and would have given me the raspberry if I'd tried to "manage" them. We operated by consensus. And the rest of the company reported to our experienced COO, who was also more of a peer.

Operating by consensus and not valuing the role of "manager" only goes so far. It may work when you're a few people living together, but I think it ultimately leads to cultures like Google's, where every decision requires a room full of engineers to agree.

Post reply on HN