5 ms·

Sadly, the cultures are not compatible. By their nature, start-ups are risk takers. You can experience a few bumps along the road to a successful launch. In ma

Sadly, the cultures are not compatible. By their nature, start-ups are risk takers. You can experience a few bumps along the road to a successful launch. In many government/large organizations, failure of any kind is not an option. The reason for the large contracts and bidding processes is political. The idea is two-fold. In the event of a failure, you have fingers to point and contractual obligations to use for re-mediation. There is also a political shell-game in place such that the decision makers in the organization can use to re-define failure by extending deadlines (at higher cost), changing requirements (at higher cost) and managing expectations (my favourite!). The money is not being spent on hardware or software. It is being spent on thousands of person hours in consulting, lawyer fees, salaried staff, hotels and meals etc.

harel 15y ago
But they do fail, often and to the tune of billions. How can an internet based project rack up a 100 million budget. How can any IT project rack up 20 BILLION pounds: http://www.telegraph.co.uk/news/uknews/1542486/20bn-NHS-computer-system-doomed-to-fail.html http://www.telegraph.co.uk/news/uknews/1542486/20bn-NHS-comp... http://cityroom.blogs.nytimes.com/2011/06/29/bloomberg-seeks-600-million-refund-from-citytime-company/ http://cityroom.blogs.nytimes.com/2011/06/29/bloomberg-seeks... I'm not saying to run the project like a startup and I'm not saying to have startup type people implement it. Just that those bids, which are often a ten thousand page manuscript nobody actually reads is passed by professionals from the startup world so that they can mark up the bullshit that those companies use to generate such bills.. You can build Google, Facebook, Microsoft and Twitter, to 1:1 scale for 20 Billion pounds...
- jerf 15y ago
  When purephase said "they can't fail", (s)he was not saying that it is impossible for the process to produce an outcome that an outsider would consider failure. It is impossible for the process to produce a result that an insider would consider failure, and all necessary damage to the budget, original spec document, and yea verily the very laws of time and space itself will be done to make sure this result occurs. Plain reality will be denied, if necessary, such as confident declarations that yes, it's OK if entering someone's name, birthdate, and address takes 40 keystrokes and half an hour, yup, successful outcome! BTW, this does precisely bupkis to solve the problem. Having competent people around makes it harder to pretend that everything went great and the job is done, so the competent people would simple be ejected at the earliest available opportunity and everybody would know it. If you really want to fix this problem you have to go a lot deeper at striking at the incentives. Lack of competence is a second-order effect of much, much deeper pathologies. (Actually I'm honestly uncertain of a way to structure incentives in such a way that this outcome won't occur in a government project. And I mean, actually won't occur, not a handwaving fairyland theory about how it might not occur, but an actual process that will produce real engineering results in the real world. The incentive mismatch in government goes all the way to the top. This is a non-trivial part of the reason behind my libertarianism.)

harel 15y ago

But they do fail, often and to the tune of billions. How can an internet based project rack up a 100 million budget. How can any IT project rack up 20 BILLION pounds: http://www.telegraph.co.uk/news/uknews/1542486/20bn-NHS-computer-system-doomed-to-fail.html http://www.telegraph.co.uk/news/uknews/1542486/20bn-NHS-comp... http://cityroom.blogs.nytimes.com/2011/06/29/bloomberg-seeks-600-million-refund-from-citytime-company/ http://cityroom.blogs.nytimes.com/2011/06/29/bloomberg-seeks... I'm not saying to run the project like a startup and I'm not saying to have startup type people implement it. Just that those bids, which are often a ten thousand page manuscript nobody actually reads is passed by professionals from the startup world so that they can mark up the bullshit that those companies use to generate such bills.. You can build Google, Facebook, Microsoft and Twitter, to 1:1 scale for 20 Billion pounds...

jerf 15y ago
When purephase said "they can't fail", (s)he was not saying that it is impossible for the process to produce an outcome that an outsider would consider failure. It is impossible for the process to produce a result that an insider would consider failure, and all necessary damage to the budget, original spec document, and yea verily the very laws of time and space itself will be done to make sure this result occurs. Plain reality will be denied, if necessary, such as confident declarations that yes, it's OK if entering someone's name, birthdate, and address takes 40 keystrokes and half an hour, yup, successful outcome! BTW, this does precisely bupkis to solve the problem. Having competent people around makes it harder to pretend that everything went great and the job is done, so the competent people would simple be ejected at the earliest available opportunity and everybody would know it. If you really want to fix this problem you have to go a lot deeper at striking at the incentives. Lack of competence is a second-order effect of much, much deeper pathologies. (Actually I'm honestly uncertain of a way to structure incentives in such a way that this outcome won't occur in a government project. And I mean, actually won't occur, not a handwaving fairyland theory about how it might not occur, but an actual process that will produce real engineering results in the real world. The incentive mismatch in government goes all the way to the top. This is a non-trivial part of the reason behind my libertarianism.)