Choose Boring Technology

Early in my career, I was a big fan of the latest and greatest technology. I kind of still am to be honest, and love using cutting edge tools.

However, I also remember running into a multitude of problems trying to put into production tools and frameworks that were still being developed or in beta. These issues would cost time, money, and slow down product development.

Then I stumbled upon Dan McKinsey’s post Choose Boring Technology and suddenly I had a way of thinking about technology that resonated with me.

The basic idea is that, as a startup, you only have three innovation tokens to spend. These innovation tokens are where your product is pushing the limits of what is possible. Your differentiators. Without these areas of innovation, you are only making something incrementally better, not 10x better.

For example, if you are building a new live-streaming service, you may choose to innovate on:

• The video capture experience, make it seamless and easy to use
• The content distribution network, make it faster than anything else on the market
• Audience interactions and interactivity, making the most engaging experience imaginable

If this is where you spend your three innovation tokens, that means you have no innovation tokens left for the rest of the system. Your frontend and backend languages/frameworks are going to be tried and true. Your CI/CD pipeline is going to be off-the-shelf. Your infrastructure is going to be simple, basic, easy to debug and easy to reason about. In other words, everything else should use boring technology.

Don’t think about using the latest AI-powered database, or rolling out a sophisticated kubernetes cluster. Focus on innovating where it matters most to your business.

This can be really tough to implement, particularly with folks that have just become senior engineers — those who are still searching for that silver bullet that will make their problems go away. It is only the most senior engineers who have come to realize that everything, everything, is a tradeoff. When things fail, and they will, you don’t want to be the first person in the world who is searching for a solution. You want the search results of the error to instantly point you to the well known work around.

Perhaps counterintuitively, this is where AI really shines. The best AI tools and agents have been trained by years of boring technology. They know the ins and outs, the patterns that work, and will produce great code the first time. I don’t mean to pick on any specific technology, but what do you think Cursor, Copilot and ChatGPT are going to be better at helping you with?

There is a time and place for innovation, and it is in service of the business. And if it is not differentiating to your customers, use boring technology.