"Within two sentences, Jobs was on him again. Tim put up his next slide, about the new plant, but again Jobs came at him with a flurry of half-insolent questions. Where are you building a plant? Why are you building a plant? Why are you manufacturing the machine yourselves?
Partly, explained Tim, because giving our code to someone else would be a great risk. Not a good reason, in Jobs's view, because the code could easily be reverse-engineered. No it couldn't, said Tim. Could, said Jobs. He added that Tim should be spending money and management time on other things, especially since there was no way he could convince any world-class manufacturing and procurement people to move to New Hampshire, for God's sake, his tone implying that only slow-witted rubes could bear such a place." - Paul Buchheit
I remember reading that when it first surfaced, but still a good story. It'd be interesting to take a Steve Jobs clone and point him at a few things... - Louis Gray
"If you launch early, you can start earlier on the process of acquiring users. Don't launch with a crappy product -- launch as soon as what you have is better than what is out there. But don't wait for a perfect product -- launch as early as you can, get user feedback, and keep improving the product." - Paul Buchheit
Look at the lower graphs. Who says mainstream media is dead. If anything, mainstream media is the enabler of new media. - Omar Ismail
it isn't dead, but it is dying. it isn't the influence that is the measure of its lifespan, but their ledgers. - Mark "Rizzn" Hopkins
Looking at those graphs, my conclusion is actually press doesn't matter, it creates a curiosity pop that quickly goes to zero. What matters seems to be organic growth. Somewhere around July '07, users started sticking, the baseline slope became nonzero, and traffic took off. I guess press pops are good for giving you feedback early on, though. - ⓞnor
Love David's openess. Really interesting graphs - Immad Akhund
where is the friendfeed's various logos to choose from for pasting on blogs? - huixing
slope became nonzero? you mean right after the Time article? Also, why are you ignoring the massive step increase that's precisely synchronized with the Newsweek article? - j1m
It seems hard to believe that the Time article created a long-term positive trend (when lots of previous press had not). The Newsweek step function, maybe could be something. I guess it boils down to what they say: make a product people will actually stick to, and then get some press to start the exponential accumulation rolling. - ⓞnor
The newsweek change is pretty big and crisp, I'm not sure you can explain it away. Poisson model, one process/user, probability of a return-to-site event determined by a stickiness function? - j1m
I'll buy that, plus some probability of word-of-mouth spreading. This is why you really want that "where did you hear about us" survey question. - ⓞnor
给我们的启示
无论是程序还是公司,架构很重要,就是如何把人员和资源搭成梯子,文化上有让别人更伟大的导向,让一个刚刚进公司的人,可以迅速的做到比他进入其他公司的同龄人获得更多的支持,这才是一个公司的结构上的成功。
好的公司,尤其是非常成功,并且长久成功的公司,一定是在基础机构上面投资最多的公司。 - Ivan Lee
"For those working on the iPhone, the next three months would be the most stressful of their careers. Screaming matches broke out routinely in the hallways. Engineers, frazzled from all-night coding sessions, quit, only to rejoin days later after catching up on their sleep. A product manager slammed the door to her office so hard that the handle bent and locked her in; it took colleagues more than an hour and some well-placed whacks with an aluminum bat to free her." - Bret Taylor
Time to combine these two postings! (Or copy/paste your comment on mine). It's like elementary school... First! - Louis Gray