I don’t know if it’s true, but apparently Elon did a LOC count. Implying he counted lines of code contributed by employees over the last year, and fired the bottom half. I am not sure if I agree upon this process, and obviously there are developers doing a massively important job having zero LOC – But for an organisation who’s losing 4 million dollars daily, where the owner comes to work the first day, already hated by the majority of his employees, due to a public smear campaign – It actually makes sense.

20 years ago publicly available research showed us that the average software developer could produce between 325 and 750 lines of code per month. This was the industry average, like it or not. Aista Magic Cloud can produce 1,000,000 lines of code per second. That’s a ratio of 1 to 234,000,000 – Implying our software system is 234 million times faster than a human being. Ignoring all other parameters, this implies that low-code can at least in theory perform on a single machine equally good as a quarter of a billion human software developers exclusively using LOC counts as your yard stick.

Tesla as a company and Aista as a company therefor have similar goals. Tesla is focusing on automation too, only in a different segment. I saw an interview with Elon Musk the other day where he basically argued that it was not clear if the economy even have an upper boundary if Tesla could create 100% perfectly self driven cars. I obviously agree, having dedicated my life towards software development automation ideas.

Swallowing the joke

Doing LOC counts have been considered a joke in the industry for 4 decades now. Bill Gates sometime back in the 80s said; “IBM used to measure our performance by LOC count. That is the equivalent of measuring an airplane’s quality by its weight”. Bill Gates is not wrong, but he’s not right either – As usual the truth is somewhere in between. When you’re going to measure a software developer’s productivity, LOC count is one of the few metrics you have. It is brutally honest, and to some extent extremely accurate. Of course, you’d have to analyse the actual code produced to make sure it’s high quality and DRY in nature, and that it has good architecture.

If you know you’re being measure by LOC, you can easily manufacture millions of lines of useless code by simply doing copy paste. However, the odds are at Elon’s favour here, since nobody knew they were going to be measured from LOC, so nobody cared about it, thinking “nobody would be that insane”. Well, I guess Elon just proved them wrong ...

For a man (me) having argued we need to bring back LOC count for years, the process Elon did feels a bit like a “victory” I have to admit

It’s of course sad when people lose their jobs, but seriously, Twitter was burning 4 million dollars every day. It was on its way to self destruct. Sometimes you have to get rid of dead flesh to save the remaining healthy flesh. If you don’t agree with me here, feel free to apply for an asylum in the last communist republic on Earth; North Korea!

Heisenberg’s LOC count uncertainty principle

The problem with LOC based performance evaluations is that if you know you’re being measured by LOC count, you can easily game the system. If you have no idea you’re to be measured by LOC, you won’t bother to even try gaming the system. This resulted in that as Elon entered Twitter, he could use it as a valuable tool, at least to some extent, to measure who was working and who was free riding. Simply because nobody at Twitter believed anybody would ever be that insane and measure their LOC count. I think I’ll refer to this as ...

The Heisenberg uncertainty principle of LOC count based performance evaluations

A year ago GitHub nulled my commits due to me changing my email address. At the time, I had 7,500 commits towards GitHub. The guy that was number 2 on the same list for the island of Cyprus had 3,500 commits. This implies that I am roughly twice as productive as the second most productive software developer amongst 2 million people, exclusively using commits as your stick, assuming GitHub is the world.

That’s a lot of assumptions of course. Most developers in Cyprus probably don’t even push towards GitHub at all. Some developers pushes monster commits once per week. Other developers are working on really hard problems, such as Kubernetes infrastructure architecture, resulting in very few lines of code, but still being crucial to the success of their employer. I want to emphasise that LOC is not the only important metric, there’s a lot of exceptions, and a lot of brilliant developers have zero LOC count. However, continuing waving that it is not important at all like Bill Gates used to do is as of now definitely in the past!

Say whatever you wish about Elon, but i’s impossible to argue that he is not a genius

If you want to learn something here, it would be to use LOC count, but you can only use it once, and you can't tell anybody you're using it before you do.