Thoughts on Twitter

I do not work at Twitter, but I have been a professional Software Engineer for about 10 years.

On Infrastructure

For infrastructure to be low-cost and low-maintenance, it needs to have an elegant architecture and robust automation. It needs to have pervasive and organized monitoring. It requires thorough, tested and well-documented operating procedures. If Twitter continues to operate, it’s to the credit of the people who designed and built such a system.
When operating infrastructure, there are some things you need to know. What are the critical pieces, and how do they interact? How are they monitored? When something breaks, who will find out about it? Will they have the necessary information to fix it, or the ability to contact someone who can?
When downsizing a team, you need to make sure to answer all of these questions, and retain as much knowledge as possible. You do this by identifying and retaining the most essential employees, or by giving people adequate time to transfer their knowledge and responsibilities to others. Of course, Musk did none of these things.
With this in mind, if Twitter continues to operate, it’s also to the credit of the remaining employees who are managing this chaotic transition. They will take responsibility over additional pieces of infrastructure that they are not familiar with. They will have to reverse-engineer systems, in cases where the previous owners have left the company. They will have to be on-call constantly as operational knowledge is recovered, and escalation pathways and rotations can be re-established. They will have to scramble do this quickly, while under the looming possibility of the system collapsing.
If drastic layoffs needed to happen, there was a way to do it without incurring nearly as much stress or risk. I really hope the people who bear the brunt of this will get something out of it.

On Product

The ideal role of infrastructure is to quietly and predictably power the site. Most people working at companies like Twitter do not think of infrastructure, but rather work on changing and improving the product.
So what might be the consequences of firing such employees? You may find yourself with a lack of imagination - missing opportunities and possible solutions to problems. You may find yourself constantly putting out fires, since you don’t have anyone to warn you of issues before they become a crisis. You may find yourself stumbling and making avoidable mistakes, as you fail to anticipate outcomes of your decisions, or try things that have been tried before.

On Success

So what would it mean if Twitter doesn’t collapse or suffer major outages? What if in 5 years, Twitter is worth more than 44B?
When betting on the future of a company, there is a considerable amount of luck. Maybe you are skilled and knowledgeable, but a pandemic happens, or Russia invades Ukraine, and your assumptions about the world don’t pan out. On the other hand, maybe you are in the right place at the right time, and a pretty obvious opportunity is handed to you. You need to be able to execute, but maybe it’s not so hard, and plenty of other people in your situation would be able to do the same.
Attributing success is an incredibly difficult problem, and one that will likely never have a clear solution. How would we tell if Twitter (or Tesla, or SpaceX?) succeeds because of Musk, or succeeds in spite of Musk?
Wealthy people can gamble and lose. If they get rich enough, they can gamble again and again, and they have the resources to keep playing, and power over thousands of people to reconcile their mistakes with reality.

On Management

The reason this story has been occupying so much of my attention lately is not really Musk or Twitter, but rather the effect that such events have on people’s understanding of what good management looks like.
I’ve known several people who worked at SpaceX and Tesla, and absolutely hated it. They were drawn in by the mission (or by necessity, as for some specializations those companies are pretty much the only employers). They lasted for a few years, but all of them eventually quit due to burnout.
Twitter presents another datapoint - does this look like a well-managed transition? Lack of communication. Swaths of people fired based on very limited information, then some asked to come back. Teams asked to grind out a new “Twitter blue” feature under incredible pressure, but then the launch was botched, predictable bad things happened, and then the feature was pulled. Offices were announced to be closed, then engineers were asked to report to the offices. Employees antagonized publically by their billionaire CEO.
Advertisers have pulled out, the platform is in chaos, and there are legitimate reasons to worry about whether the site is going to continue operating. And none of this was necessary! Elon is still paying all of these people for the next 3 months!
This is not what good management looks like.

On Culture

I think mainly I am writing this for the other people (engineers and otherwise) who work in tech. It is easy to be taken in by the culture that Elon represents, especially if you are younger, or new to this industry.
Grinding young, motivated people to burnout is not a new idea. Coercing employees (through money, access to healthcare or immigration) is not a new idea. Unfortunately, such methods still get results, and so they keep being re-discovered as “hard truths”. Hopefully this article has convinced you that this is not actually how things work, and that there are better ways of doing things.
Beyond just “what is most effective”, I think you should also consider what kind of world do you want to live in. Should people’s worth be defined by their productivity? If someone is unable or unwilling to be “extremely hardcore”, should they be fired? Disrespected? Should your employer be able to fire you without cause? Should your employer dictate where you live, and take precedence over all other ways you may want to spend your time?
Our culture is the sum of our beliefs and expectations. We all deserve better than this.