3 ms·

Hey, excuse me for using you as an example but there’s a lot of this rough idea in this comment section and I’m rather confused by it. I don’t mean to single yo

Hey, excuse me for using you as an example but there’s a lot of this rough idea in this comment section and I’m rather confused by it. I don’t mean to single you out. What would be wrong with that? In a search for low hanging fruit, the planes that can’t make it to the tarmac would be a group of their own. Of the ones that do, next to others some stand out as being noticeably different. Etc. Then patterns may emerge based on that, iterate iterate. Seems to depend on where you’ve set the bar. If you were an office space Bob, depending on the bar placement, you may just be interviewing to see if someone loses their cool when asked “what is it you do here?”. If you were a cop you might bring in every suspect just to see if someone stands out. So what am I missing? Is it the scale of the operation?

Joeri 4y ago
In every field there are ways that things are done. To review a fleet of airplanes I assume “the way it is done” would be based on trawling through inspection reports, not looking at the actual planes. Elon Musk has a complete disregard for “the way it is done” to the point of not even trying to do things that way as a sane default to iterate from, and that makes him the sort of person who would line up the airplanes on the tarmac instead of going through inspection reports. That attitude is why people who are deep into conventional wisdom dismiss him as an idiot. That attitude is also why both his successes and his failures are spectacular. When you deviate from the beaten path and run as fast as you can in that new direction sometimes you reach a far better vantage point and sometimes you just get totally lost in the jungle. With Elon you never know which way things will go.
- recyclelater 4y ago
  Russia has inspection reports on planes and tanks that don’t run. They say they are ready for use. I’m with the GP - sometimes when you have a goal that you need to reach in a very short period of time, you do things that don’t scale and don’t follow standard best practices. I see zero issues with lining up all the planes on the tarmac and walking a pilot mechanic pair through each of them for a fifteen minute review per plane. Do they find every problem? Do they misdiagnose? Of course. Are they able to roughly say “this feels like the industry norm, I don’t see anything you wouldn’t see at another airline” or “I am surprised at the rough shape of many of the cockpits. Yes they fly but there might be other issues. Two planes stood out more than the others I’d start there for a deeper inspection.” Def not unreasonable.
  - Waterluvian 4y ago
    This is a perfect caricature of the “how hard can it be?” tech startup mindset. Well done.
    - recyclelater 4y ago
      Definitely not one to trivialize things and pretend tech can solve everything, or that programmers are so smart. But I am not one to think everything has to be perfect and by the book. Lots of room for rough pragmatism in the world.
jeroenhd 4y ago
If there is a reasonable suspicion that the airplanes are not of considerable quality beyond rumours online and there is an honest intent to fix the airplanes rather than dispose of the ones that are deemed unnecessary, such a plan could certainly be executed over several weeks or months. In this case, there is a major scramble to test every employee as quickly as possible by outsiders who have no idea how the company operates, what their standards are like, or what the problem space is about. Barely self driving cars controlled by touch screens are almost entirely disconnected from a social media network other than that both employ programmers. Everything from the data processing location to the user interaction model is vastly different. The standards of quality of different components vastly differ; a UI bug in Twitter is far from catastrophic unlike a UI bug in your car, but a backend bug in Twitter can sink the business whereas a backend bug with Tesla can be annoying at worst. I'm sure Tesla's developers are competent but the environment they operate in is completely different from the environment the people they review operate in. At worst, the entire review produces no usable results because the Tesla developers don't have time to study the code base and validate any of the claims the devs make. At worst, the rushed reports are used to bypass layoff laws by providing a "reason" to fire people. Either way, the end result is that the entire process is no more than pointless busywork for Tesla's people and has no real benefit to the Twitter code base. Knowing Elon's intentions to fire half of Twitter but his subsequent retraction (probably advice from his lawyers because of layoff laws) I think it's foolish not to be sceptical of this entire event. Code review can be a great tool to improve a business when it's done well; however, the rate at which things are progressing now indicates to me that this isn't done well.

Joeri 4y ago

In every field there are ways that things are done. To review a fleet of airplanes I assume “the way it is done” would be based on trawling through inspection reports, not looking at the actual planes. Elon Musk has a complete disregard for “the way it is done” to the point of not even trying to do things that way as a sane default to iterate from, and that makes him the sort of person who would line up the airplanes on the tarmac instead of going through inspection reports. That attitude is why people who are deep into conventional wisdom dismiss him as an idiot. That attitude is also why both his successes and his failures are spectacular. When you deviate from the beaten path and run as fast as you can in that new direction sometimes you reach a far better vantage point and sometimes you just get totally lost in the jungle. With Elon you never know which way things will go.

recyclelater 4y ago
Russia has inspection reports on planes and tanks that don’t run. They say they are ready for use. I’m with the GP - sometimes when you have a goal that you need to reach in a very short period of time, you do things that don’t scale and don’t follow standard best practices. I see zero issues with lining up all the planes on the tarmac and walking a pilot mechanic pair through each of them for a fifteen minute review per plane. Do they find every problem? Do they misdiagnose? Of course. Are they able to roughly say “this feels like the industry norm, I don’t see anything you wouldn’t see at another airline” or “I am surprised at the rough shape of many of the cockpits. Yes they fly but there might be other issues. Two planes stood out more than the others I’d start there for a deeper inspection.” Def not unreasonable.
- Waterluvian 4y ago
  This is a perfect caricature of the “how hard can it be?” tech startup mindset. Well done.
  - recyclelater 4y ago
    Definitely not one to trivialize things and pretend tech can solve everything, or that programmers are so smart. But I am not one to think everything has to be perfect and by the book. Lots of room for rough pragmatism in the world.

jeroenhd 4y ago

If there is a reasonable suspicion that the airplanes are not of considerable quality beyond rumours online and there is an honest intent to fix the airplanes rather than dispose of the ones that are deemed unnecessary, such a plan could certainly be executed over several weeks or months. In this case, there is a major scramble to test every employee as quickly as possible by outsiders who have no idea how the company operates, what their standards are like, or what the problem space is about. Barely self driving cars controlled by touch screens are almost entirely disconnected from a social media network other than that both employ programmers. Everything from the data processing location to the user interaction model is vastly different. The standards of quality of different components vastly differ; a UI bug in Twitter is far from catastrophic unlike a UI bug in your car, but a backend bug in Twitter can sink the business whereas a backend bug with Tesla can be annoying at worst. I'm sure Tesla's developers are competent but the environment they operate in is completely different from the environment the people they review operate in. At worst, the entire review produces no usable results because the Tesla developers don't have time to study the code base and validate any of the claims the devs make. At worst, the rushed reports are used to bypass layoff laws by providing a "reason" to fire people. Either way, the end result is that the entire process is no more than pointless busywork for Tesla's people and has no real benefit to the Twitter code base. Knowing Elon's intentions to fire half of Twitter but his subsequent retraction (probably advice from his lawyers because of layoff laws) I think it's foolish not to be sceptical of this entire event. Code review can be a great tool to improve a business when it's done well; however, the rate at which things are progressing now indicates to me that this isn't done well.