Slashdot Mirror
Former Google CIO Suggests 'Do Dumb Things'
itwbennett writes "Speaking at the CA Expo in Sydney, Australia, former Google CIO Douglas Merrill shared some management tips he learned during his tenure at the search giant. At the top of the list: 'Don't be afraid to do dumb things.' Merrill recalls that 'most of the early Google hardware was stolen from trash and as the stuff they stole broke all the time they built a reliable software system. Everyone knew we shouldn't build our own hardware as it was 'dumb', but everyone was wrong. Sometimes being dumb changes the game.' Another pearl of wisdom from Merrill: 'the more project management you do the less likely your project is to succeed.'"
Does anyone here even read the summary together with the article itself and see if it makes sense? He did *not* say "do dumb things". That statement implies that you know its a dumb thing to do and it will not work, yet you do it anyway. In this case you *are* dumb and should be fired. He said "don't be afraid to do dumb things", which has a totally different meaning. It means that you should try approaches that may be non-obvious, but at least you are attempting to solve whatever the problem at hand is in an ingenious way. Sometimes it does not work and you look foolish, but you often get innovative solutions to tough problems.
Project management is an inverted parabola. Too little and you will fail due to lack of direction and budget, too much and you will smother it.
Continuing this line of thought, Merrill said, "Put all your eggs in one basket; Count your chickens before they hatch. Serve some wine before its time, find yourself an itch to scratch."
Bow-ties are cool.
If you're not buying that how about buying reading lessons.
He's saying that getting cheap crappy hardware that failed all the time forced them to write software with a high tolerance for failure. A little between the lines and he's saying that if they didn't go through the hardship of failing hardware they would've never written something that could fail over to other machines. I can imagine that a system that fails over to a new server quickly would also be highly scalable with a little tweaking.
That 'dumb decision' to not invest in reliable hardware indirectly helped them build a highly scalable search system that became their whole companies foundation. Over management and strict adherence to known process would have produced a Google search that would've required a massive rewrite to their (at the time only) product right when they were getting off the ground.
TLDR: If you're big enough to hire a consultant that tells you to, "Think outside the box", you're probably too big to think outside the box.