4 ms·

> Doing this at a FAANG would require pulling in 20+ full timers for a year before shipping our first line of code. Why are these large companies so slow? Are

> Doing this at a FAANG would require pulling in 20+ full timers for a year before shipping our first line of code. Why are these large companies so slow? Are the 10+ yoe engineers, scientists with deep domain expertise, experienced people managers, etc. helping or hurting this?

thwayunion 4y ago
> Why are these large companies so slow? A combination of risk aversion, regulatory burden, and process. The corporate machine can't afford to trust the judgement of individual contributors or even their managers or even their directors or sometimes even their VPs. Credit where it's due: some of that is because of real risks that large corps do need to worry about but scrappy startups don't need to worry about. A lot of it though -- especially the burdens that come from standardized processes and distributed responsibility -- boils down to a combination of (1) the cost of treating employees like cogs and (2) some amount of empire building. > Are the 10+ yoe engineers, scientists with deep domain expertise, experienced people managers, etc. helping or hurting this? It can go either way, but the question is kind of ill-formed. The decisions that slow stuff down happen at the VP+ level. If you're being slowed down by an IC, it's usually not the IC's fault (you'd rather not even have to interact with them in the first place, but someone VP+ demanded that you must). That said, generally the valuable people with tons of experience are the ones building/fixing stuff. The "process implementation cogs" are typically more commodity-like labor.
whatever1 4y ago
Because for FAANG to invest resources in a program, they need to make sure that it will return hundreds of millions of dollars. Don’t forget that every year they need to show to the shareholders that they grow with double digits, and for FAANG this means billions of dollars in extra revenue. So a lot of time is spent evaluating a program in stead of just doing the program.
- thwayunion 4y ago
  This as well. Things that make for fantastic lifestyle businesses -- a few people doing mid-seven revenue with minimal compute expenses -- just isn't worth it for a FAANG spending an order of magnitude more than that per hour just on one guy's compensation package.

thwayunion 4y ago

> Why are these large companies so slow? A combination of risk aversion, regulatory burden, and process. The corporate machine can't afford to trust the judgement of individual contributors or even their managers or even their directors or sometimes even their VPs. Credit where it's due: some of that is because of real risks that large corps do need to worry about but scrappy startups don't need to worry about. A lot of it though -- especially the burdens that come from standardized processes and distributed responsibility -- boils down to a combination of (1) the cost of treating employees like cogs and (2) some amount of empire building. > Are the 10+ yoe engineers, scientists with deep domain expertise, experienced people managers, etc. helping or hurting this? It can go either way, but the question is kind of ill-formed. The decisions that slow stuff down happen at the VP+ level. If you're being slowed down by an IC, it's usually not the IC's fault (you'd rather not even have to interact with them in the first place, but someone VP+ demanded that you must). That said, generally the valuable people with tons of experience are the ones building/fixing stuff. The "process implementation cogs" are typically more commodity-like labor.

whatever1 4y ago

Because for FAANG to invest resources in a program, they need to make sure that it will return hundreds of millions of dollars. Don’t forget that every year they need to show to the shareholders that they grow with double digits, and for FAANG this means billions of dollars in extra revenue. So a lot of time is spent evaluating a program in stead of just doing the program.

thwayunion 4y ago
This as well. Things that make for fantastic lifestyle businesses -- a few people doing mid-seven revenue with minimal compute expenses -- just isn't worth it for a FAANG spending an order of magnitude more than that per hour just on one guy's compensation package.