4 ms·

You're being downvoted, but you're right. I used to work in defense/intel, and then spent a lot of time at a large Federal consulting firm, where I learned jus

by JPKab 5y ago

You're being downvoted, but you're right. I used to work in defense/intel, and then spent a lot of time at a large Federal consulting firm, where I learned just how insanely wasteful these agencies are on software projects. It's not THAT infuriating that the product is so thoroughly _NOT_GREAT_, until you find out that the budget for it exceeds by orders of magnitude anything you've ever seen. The size of the teams on these projects is mind-boggling, and when you find out there's multiple managers for every developer, that's when your head will explode. People on this site tend to think that any attack on the horrific spending efficiency of these agencies is motivated by an antipathy of the government in general, or a view that the government CAN'T do it right. I'm not of that mind at all. My gripe is that the people who are most invested in expanding the government's role in the US for positive goals seem to have zero interest in even discussing the rot within these institutions that has crippled their ability to execute on the funding they already have. I can confidently assert that, if I'm being generous, 3/5s of the current Federal (and yes, contractors too!) work force is what my old chief called "furniture". I have no doubt that many, many private corporations have the exact same problem. (I've seen the inside of the average insurance company, and my god are those places bloated).

ekidd 5y ago
> It's not THAT infuriating that the product is so thoroughly _NOT_GREAT_, until you find out that the budget for it exceeds by orders of magnitude anything you've ever seen. I once interacted with a private government contractor. They did an adequate job on a project. It worked, and it wasn't terrible. But they billed $5 million for a project that could have been built for $50 thousand, or—once you put up with a drawn-out sales process, endless client meetings and the occasional change request—more like $500,000. The simple fact is that many large organizations, public or private, couldn't do a $50,000 software project. It costs more than that to close the sale and negotiate contracts. As Joel Spolsky pointed out, once you need to put salespeople on airplanes, even shrinkwrap software costs $75,000: https://www.joelonsoftware.com/2004/12/15/camels-and-rubber-duckies/ https://www.joelonsoftware.com/2004/12/15/camels-and-rubber-.... > I have no doubt that many, many private corporations have the exact same problem. (I've seen the inside of the average insurance company, and my god are those places bloated). Yeah, that has often been my experience. Sufficiently large companies tend to be deeply inefficient about small things. Partly this is because they're optimizing for different needs than you might think, and partly because $1 million is basically pocket change if it actually solves a problem. But also, once a company grows beyond a certain size, it operates internally like a centralized command-and-control economy. Internal politics matter far more in the short run than any market feedback. If anything, governments can have a slight advantage if they keep things in-house. Government employees are often slightly more idealistic than their private sector counterparts, and they're definitely paid less. So an organization like the US Digital Service can outperform the original private contractors who built healthcare.gov. The USDS was staffed by idealistic FAANG developers working for the common good at below-market salaries. They did a better job than CGI Federal, the original private contractor.

by JPKab 5y ago

ekidd 5y ago

> It's not THAT infuriating that the product is so thoroughly _NOT_GREAT_, until you find out that the budget for it exceeds by orders of magnitude anything you've ever seen. I once interacted with a private government contractor. They did an adequate job on a project. It worked, and it wasn't terrible. But they billed $5 million for a project that could have been built for $50 thousand, or—once you put up with a drawn-out sales process, endless client meetings and the occasional change request—more like $500,000. The simple fact is that many large organizations, public or private, couldn't do a $50,000 software project. It costs more than that to close the sale and negotiate contracts. As Joel Spolsky pointed out, once you need to put salespeople on airplanes, even shrinkwrap software costs $75,000: https://www.joelonsoftware.com/2004/12/15/camels-and-rubber-duckies/ https://www.joelonsoftware.com/2004/12/15/camels-and-rubber-.... > I have no doubt that many, many private corporations have the exact same problem. (I've seen the inside of the average insurance company, and my god are those places bloated). Yeah, that has often been my experience. Sufficiently large companies tend to be deeply inefficient about small things. Partly this is because they're optimizing for different needs than you might think, and partly because $1 million is basically pocket change if it actually solves a problem. But also, once a company grows beyond a certain size, it operates internally like a centralized command-and-control economy. Internal politics matter far more in the short run than any market feedback. If anything, governments can have a slight advantage if they keep things in-house. Government employees are often slightly more idealistic than their private sector counterparts, and they're definitely paid less. So an organization like the US Digital Service can outperform the original private contractors who built healthcare.gov. The USDS was staffed by idealistic FAANG developers working for the common good at below-market salaries. They did a better job than CGI Federal, the original private contractor.