A few years ago, on a Saturday in May, the lovely Mrs. Wednesday Wisdom and I wanted to go out to dinner in Boston. Unfortunately, for one of the most successful and important States in the Nation, Massachusetts has a dearth of good food, an argument that is supported by the fact that it scored its first Michelin star restaurant only in early 2026. There are a few decent restaurants, but reservations are typically mandatory because the entire BCBG of Boston and surrounding areas descend on them. On this particular Saturday however, there were no reservations to be had anywhere. The reason? Taylor Swift was in town and her fans had crowded out every restaurant and every wine bar.
I am not a particular fan of Ms. Swift or her music, but I will give her this: She is everywhere, all the time. Even if, like me, you have never heard one of her songs from start to finish, she is hard to miss.
Compulsory Youtube video: Prog-metal drum god Mike Portnoy drumming to a Taylor Swift song that he is hearing for the first time. Key quote from the video: “It’s all verses, it doesn’t change keys. I don’t even know what to write.”
Taylor’s face is of course plastered over the entire world’s magazines and newspapers all the time. One day I was watching an American football game and there she was, in a sky box, with the camera zooming to capture her lovely appearance whenever there was nothing happening on the field, which in American football is of course permanently the case. Even if you never listen to Taylor Swift, you know she exists and you probably know that the Eras tour was the highest grossing tour of all time. That tour was so big in fact, that its concerts in Stockholm had a measurable impact on Sweden’s inflation rate.
I am thinking a lot about Taylor Swift these days when I am considering how to make something happen in an organization full of engineers, all of which are on their own mission with their own OKRs and their own deadlines. If you work in infrastructure there is always a need to make some changes that impact all the engineers in the company. How to make sure that all these engineers, who could usually not care less about your goals, comply?
Here is a current example: I am working on infrastructure security these days and it would be a significant value add to my life if every service in the company would automatically deploy whenever there are changes. Unfortunately, this is very difficult to achieve safely, as it requires a robust deployment infrastructure with great integration tests, automatic canary validation, and careful design and implementation of the service, so that it cannot only roll forward but can also (automatically) roll back if the new version is not as amazing as we need it to be.
All of this of course requires the help from the engineering teams who develop, and are ultimately in charge of, the services we are talking about. But how to get them to do this work?
To start with, this requires executive backing. Without explicit support from senior managers in the company you will not get anything done because engineers would be foolish to spend significant amounts of time on something that the people ultimately in charge of their compensation and continued tenure at the company are not supportive of.
But, executive backing is not enough. Most executives have no clue about what is going on down there where the rubber meets the road and they always “support” many more initiatives than there is time in the day to work on. Because of this, individual engineering teams still have a large say in determining what they are actually going to work on. The company more often than not offers them a smorgasbord of different things to work on and engineers, with the help of their direct managers, pick and choose. How to make sure that they work on your important thing?
You will typically have no problem getting people to work on your program if the work is inherently interesting or sexy. A few years ago, I led a drive to convert a large system we were working on from Python to Rust and there was no shortage of interest to work on this because Rust is cool. In fact, I explicitly pitched the rewrite as an opportunity to attract internal talent to our teams because of the widespread interest in working with this new and hot programming language.
But what if the work you want people to do is not sexy, but just important? The good news here is that you can still get people to work on this because there are, fortunately, more than enough people willing to work on unsexy but important things because they have an intrinsic motivation to make things better. There is a challenge though: Lots of things are important; how to make sure that people are working on your important thing?
This is where Taylor Swift comes into play. Your program of work, like Taylor, needs to be everywhere, all the time.
Many engineers are suffering from the misconception that some ideas are so good that you just need to explain them to people once and then they jump on board of your merry bandwagon and you are off to the races. Nothing is further from the truth… If you bring something important to the table, it will get filed with all of the other important things that need to be done and if you don’t follow up, nothing will be done because people, like goldfish, mostly have a ten second memory for all the things that they need to work on.
So instead, be like Taylor, be everywhere, all the time. When people come to an all-hands, they need to see your program. You need to create tables and dashboards, and then distribute these to people. You need to have badges, prizes, and swag to reward people who complete their part of your program. You need to have status updates, newsletters, a Slack channel, you name it. One message is no message, so you need to make sure that you create continuous communication events where people are informed that your program is still a thing and that we are not there yet.
Which brings me to the second part of this article, which is the value of time.
Young people always underestimate the value of time.
Time has many wonderful qualities, one of which is that it allows effort to compound. This is shown quite prominently by compound interest, in itself just a simple geometric progression, but with surprising results as time turns a small yield into staggering quantities. It is for that reason that Einstein (probably mistakenly) is believed to have said: “Compound interest is the eighth wonder of the world. He who understands it, earns it... he who doesn’t... pays it.”
Whenever you take on a big activity and evaluate where you are after a few weeks, it is common to feel underwhelmed with the progress you made. If absolutely no progress was made, you might want to reconsider your methods, but if, what is more likely, at least some progress was made, my advice is to keep hacking at it. Just as $100 in a savings account might have only grown to a measly $105 after one year, the real progress happens when you stick with the program and let the small increases of value you make compound into, eventually, staggering results.
There are few, if any, overnight successes. My own employer, OpenAI, started in 2015, but it was not until the launch of ChatGPT in 2022 that it became an “overnight” success. Same for your program. You need to be like Taylor Swift, everywhere all the time and keep at it for weeks, months, or years (depending on the mission) to get anything done.
One thing that continued investment over time gives you is that it deepens the conviction of the people who need to work with you that this is something that is really important and that needs to be done.
Here, I would like to retell a story from an earlier Wednesday Wisdom:
In the second century BCE, the Roman republic was vying for dominance of the Mediterranean with Carthage. A few wars had been fought in which Rome mostly prevailed, reducing Carthage to a small but still wealthy territory in what is now northeastern Tunisia. The Roman senator Cato the Elder was of the opinion that Carthage should be definitively beaten, to which end he steadfastly and at every opportunity mentioned “Ceterum censeo Carthaginem esse delendam”, which means: “Furthermore, I think that Carthage must be destroyed”. In fact, he ended all of his senate speeches with this line, even when the speech was about something else altogether. Finally, in 146 BCE, Carthage was razed by the Romans and its entire remaining population was sold into slavery.
This is a good example of what I am talking about. You could mention it once as something important that needs to be done and people might even agree with you. But then other matters take priority and people forget about it. But Cato kept at it and continued his program until everyone was convinced and he got the results he wanted. I have been to the spot in (what is now) Tunisia where Carthage once stood and let me tell you, there is literally nothing to see.
Of course this requires significant perseverance and stamina. Toiling through the program while you are not seeing any results yet can be disheartening and it might even lead you to lose your executive backing, but there is nothing to be done about it other than continue to work on it smartly and steadily. This is one of the things you can learn from Microsoft. When Windows 1.0. was released it wasn’t great and consequently it was not very successful. But, they persevered and created Windows 2.0. Still not great. Then they created Windows 2.1 (a.k.a. Windows 286 and Windows 386), which used some of the “advanced” features of more modern CPUs (like memory protection and virtual memory). Still not awesome, but you might have been excused for thinking that this was potentially going somewhere.
In the meantime, the hardware had caught up and when Windows 3.0 and eventually Windows 3.1 came out it was not entirely terrible and could run on a lot of modern PCs. Then Windows for Workgroups was released and took the world by storm because of its effortless networking features. And now Windows, like Taylor Swift, is (still) everywhere, all the time.
