3 ms·

I also worked at SpaceX - they are not efficient. They refuse to buy software or use cloud solutions, their existing software systems have immense bloat, and th

by Revys 2y ago

I also worked at SpaceX - they are not efficient. They refuse to buy software or use cloud solutions, their existing software systems have immense bloat, and their infrastructure is held together by 1 overworked IT guy, and if he left half the factories at the company would shut down overnight.

talldatethrow 2y ago
That sounds pretty damn efficient, although risky.
- Revys 2y ago
  The real-world outcomes from what I described is that it takes teams of dozens of software developers being hired to either reinvent the wheel or spend all of their time managing tech debt rather than solving actual problems.
hammock 2y ago
Serious question, what is costlier, for a critical high-risk long-gestation operation like space travel - paying 1 IT guy to stick around, or having to go through an entire transformation/ migration/ contract a custom build when your third party cloud intrastructure inevitably goes under, or deprecates, or fails to meet your needs, etc? Granted, there is probably a better third way
- booi 2y ago
  depends on the bus factor...
- makeitdouble 2y ago
  The former is costlier. It's litteraly a gamble that your one IT guy won't: - just leave - get sick (including burn out) - move to another country - get hit by a bus It just doesn't make sense in the long term. PS: bot even counting that the negociating power of that single guy is incredible.
  - ryandvm 2y ago
    I used to worry about "bus factor", but after seeing countless "indespensible" people leave abruptly and the companies recover just fine, I don't really think it's an issue. It generally sets objectives back a few weeks and that's about it.
    - talldatethrow 2y ago
      I knew an entire custom manufacturing company with 40 employees run with one main engineer, who worked 50-60 hours to complete his work so the employees in the shop could run the machines and do assembly. He died one day suddenly at 50 years old. He had been the sole engineer for 10 years. A week later I was training his replacement on how to use the PM tool I provide for his company. Same office, same chair.
    - makeitdouble 2y ago
      It then comes down to whether you can afford to be off for weeks. Another aspect of this is the hiring of a new guy: I've seen it times and times, where the company doesn't have anyone to evaluate candidates, so they end up peddling more time and money to even get to the point they can start interviewing. Of course it all comes down to the balance between how much you save keeping a single guy on staff, vs how much it costs when shit hits the fan.
- readthenotes1 2y ago
  I gave the danger of the bus number one lecture to a bunch of managers once, and had one pushback pretty hard by saying that what he was working on was so complex that no one else could possibly learn it in time to be useful for the next release (every release...). They bought me lunch a few months later to apologize after having the unique person die and leave them with no backup. -- I have a relative who just retired and was replaced by five people who started a year ago. It is extraordinarily difficult to replace decades of knowledge and wisdom.
gigatexal 2y ago
My buddy works at Tesla here at the gigafactory in Berlin and says the same thing. It’s all on-premise run using open stuff. Which isn’t horrible but the culture seems to be we will build it ourselves.

by Revys 2y ago

talldatethrow 2y ago

That sounds pretty damn efficient, although risky.

Revys 2y ago
The real-world outcomes from what I described is that it takes teams of dozens of software developers being hired to either reinvent the wheel or spend all of their time managing tech debt rather than solving actual problems.

hammock 2y ago

Serious question, what is costlier, for a critical high-risk long-gestation operation like space travel - paying 1 IT guy to stick around, or having to go through an entire transformation/ migration/ contract a custom build when your third party cloud intrastructure inevitably goes under, or deprecates, or fails to meet your needs, etc? Granted, there is probably a better third way

booi 2y ago
depends on the bus factor...
makeitdouble 2y ago
The former is costlier. It's litteraly a gamble that your one IT guy won't: - just leave - get sick (including burn out) - move to another country - get hit by a bus It just doesn't make sense in the long term. PS: bot even counting that the negociating power of that single guy is incredible.
- ryandvm 2y ago
  I used to worry about "bus factor", but after seeing countless "indespensible" people leave abruptly and the companies recover just fine, I don't really think it's an issue. It generally sets objectives back a few weeks and that's about it.
  - talldatethrow 2y ago
    I knew an entire custom manufacturing company with 40 employees run with one main engineer, who worked 50-60 hours to complete his work so the employees in the shop could run the machines and do assembly. He died one day suddenly at 50 years old. He had been the sole engineer for 10 years. A week later I was training his replacement on how to use the PM tool I provide for his company. Same office, same chair.
  - makeitdouble 2y ago
    It then comes down to whether you can afford to be off for weeks. Another aspect of this is the hiring of a new guy: I've seen it times and times, where the company doesn't have anyone to evaluate candidates, so they end up peddling more time and money to even get to the point they can start interviewing. Of course it all comes down to the balance between how much you save keeping a single guy on staff, vs how much it costs when shit hits the fan.
readthenotes1 2y ago
I gave the danger of the bus number one lecture to a bunch of managers once, and had one pushback pretty hard by saying that what he was working on was so complex that no one else could possibly learn it in time to be useful for the next release (every release...). They bought me lunch a few months later to apologize after having the unique person die and leave them with no backup. -- I have a relative who just retired and was replaced by five people who started a year ago. It is extraordinarily difficult to replace decades of knowledge and wisdom.

gigatexal 2y ago

My buddy works at Tesla here at the gigafactory in Berlin and says the same thing. It’s all on-premise run using open stuff. Which isn’t horrible but the culture seems to be we will build it ourselves.