You don't need to know how to code in a backend language. You do need to know enough to make a good hiring decision and to spot when an engineer is choosing what they enjoy over what your product needs.
This video runs through the four languages you'll actually hear proposed — JavaScript (Node.js), Python, Java, and Go, and compares them on the things a founder cares about: how fast you can build, how easily and cheaply you can hire, and whether one engineer can cover both sides of your stack.
The honest headline: which language matters far less than you'd think. For most early products, Node or Python is the sane default — pick Node if your frontend is already JavaScript (one engineer, one hiring pool), and Python if your product leans on data, automation, or AI. Java and Go are usually "later" or specialist choices, so seeing either proposed for a brand-new MVP is a smell test.
Because the language rarely kills a product, the wrong-size choice wastes runway. The real failure isn't picking Python over Go; it's hiring scarce, expensive Go engineers to build something 100 people will use.
So when an engineer proposes a language, ask why, and listen for whether the answer is about your product and your hiring, or about what they personally enjoy. The first is a reason. The second is a preference wearing a reason's clothes.