4 ms·

Is part of the problem not reading the mythical man month, and trying to hire your way out of problems that don't work that way. I honestly don't get why there

Is part of the problem not reading the mythical man month, and trying to hire your way out of problems that don't work that way. I honestly don't get why there would be more than around 25-30 software developers working for Uber, generally speaking. I mean other resources around them, sure... Many companies will be stuck needing things done, hire teams and teams of new developers and things will only fall farther behind as a result. Instead of more designers, ux, management, leadership, review, marketing ahead of development so a clear experience is defined for a given project, they try to divide into more developer teams each bottle-necked more than if there were just fewer dev teams. Meanwhile, so much money is spent trying to get the smallest uplift in efficiencies when the problems tend to be systematic. --- I mean to exclude certain things like self-driving cars that should be a separate sub-company. Also, a bit off... since a team of 4-5 developer per device category or platform interest. api, web-ui, ios-ui, android-ui, 2x data-services, driver-devices, restaurant-devices, support-apps, 2x admin-apps/reporting (still talking well under 100).

wolco 6y ago
Are they still working on self driving cars? That will push the headcount up. For each developer you have project managers, product owners, database admins, techs, ba/technical ba, qa. Not to mention probably a big legal team. Nevermind marketing / ui/ux guys or guys doing reporting or people internal tools.
- tracker1 6y ago
  imo, self-driving cars should probably be a separate company owned by the parent... and likely is, and doesn't even apply to the thousands of developers alone they are stafing. And for each developer there is probably 1:1 for other staff... in my experience for a team of 4 devs, you usually get 1/2 time for a PM, 1/2 time for an owner, a QA, 1/3-1/4 dba and a BA and maybe 1/4 time of a designer. The primary application could probably do very well with 3-5 dedicated teams to portions of the application with less general conflict and bottlenecking... The support staff would need to scale to meet the larger demands though. Given their funding, even top-tier engineers at $250-300k (Northern California) base would still be less expensive than hundreds or thousands of developers. Putting the self-driving car research into a separate company. The site is not THAT complex, and they are, or should be leveraging a number of existing tools. They probably have a byzantine codebase that takes days to make what should be a 15 minute change at this point because of the shear numbers.

wolco 6y ago

Are they still working on self driving cars? That will push the headcount up. For each developer you have project managers, product owners, database admins, techs, ba/technical ba, qa. Not to mention probably a big legal team. Nevermind marketing / ui/ux guys or guys doing reporting or people internal tools.

tracker1 6y ago
imo, self-driving cars should probably be a separate company owned by the parent... and likely is, and doesn't even apply to the thousands of developers alone they are stafing. And for each developer there is probably 1:1 for other staff... in my experience for a team of 4 devs, you usually get 1/2 time for a PM, 1/2 time for an owner, a QA, 1/3-1/4 dba and a BA and maybe 1/4 time of a designer. The primary application could probably do very well with 3-5 dedicated teams to portions of the application with less general conflict and bottlenecking... The support staff would need to scale to meet the larger demands though. Given their funding, even top-tier engineers at $250-300k (Northern California) base would still be less expensive than hundreds or thousands of developers. Putting the self-driving car research into a separate company. The site is not THAT complex, and they are, or should be leveraging a number of existing tools. They probably have a byzantine codebase that takes days to make what should be a 15 minute change at this point because of the shear numbers.