Should I prioritize stability or new features?

This is a classic dilemma that almost every internet team grapples with daily. It's like asking, "Is speed or safety more important for a car?" The answer is definitely "both are important," but your focus will certainly vary at different stages.

Let me give you an analogy to help you understand.

Stage One: From Zero to One, Survive First

You've just started a business, and your product is like a small raft. Your goal is to quickly paddle to the other side to see if anyone there needs your raft.

At this point, the core is "new features." You need to quickly build the core functions of your raft (e.g., it can carry people, it can float). You need to iterate quickly. If a design doesn't work, move on to the next immediately. If you spend half a year polishing the planks, only to find out that no one needs to cross the river in that area, wouldn't that be a waste of effort?
What about stability? "Just needs to hold together." As long as your users can barely get on board and experience your core value, they can tolerate minor issues or occasional wobbles. The biggest risk at this stage isn't "the boat isn't stable enough," but "no one gets on board at all."
Conclusion: In the earliest stage, boldly allocate resources towards new features to quickly validate your ideas. The goal for stability is "usable," not "perfect."

Stage Two: From One to Ten, Time to Expand

Your raft is a success! Many people are lining up to use your raft to cross the river, and some are even willing to pay.

At this point, the balance starts to tip towards "stability." You have more and more users, and they rely on you to cross the river every day. If your raft constantly has problems, leaking today and breaking an oar tomorrow, users will quickly lose patience and look for someone else's boat. That's how reputations get ruined.
Do you still need new features? Of course. This is like upgrading your raft to a small ferry. You might want to add a canopy or sell snacks and drinks. These are your growth points. But the prerequisite is that you must ensure this ferry can safely and punctually deliver people to the other side.
Conclusion: At this stage, stability and new features must be equally important, or you might even consider a "70/30" or "80/20" split (70%-80% of resources for new features, 20%-30% specifically for technical debt, bug fixes, and optimization). Stability itself has become one of your most important "features." Without stability, there is no growth.

Stage Three: From Ten to One Hundred, Becoming Infrastructure

Your ferry has become a cross-river bridge, with countless people and vehicles passing over it every day, even becoming part of the city's transportation system.

At this point, "stability" is everything. The reliability of the bridge must be 99.99% or even higher. Any minor malfunction could lead to massive chaos and losses.
New features will be introduced very cautiously. For example, if you want to add an observation deck to the bridge, it would require countless rounds of deliberation and testing to ensure it poses no risk to the bridge's structure, and then it would be opened in phases and on a small scale.
Conclusion: In the mature stage, stability trumps everything. All new features must serve the stability and safety of the entire system.

Here are some simple judgment methods for you:

Listen to your users: Are users complaining, "This thing keeps crashing, it's unusable," or are they lamenting, "I wish it had XX feature"? The users' pain points are what you should address first.
Look at the "pain index": Does the team suffer more from "lack of new features" or "system instability"? Whichever hurts more, fix that first.
Don't make it a multiple-choice question, make it an allocation problem: Never think about "only doing A and not B." A healthy approach is to consistently allocate a portion (e.g., 15%-20%) of your effort, regardless of the stage, specifically to address legacy issues and optimize the system. This is like regularly cleaning your house; it prevents trash from piling up until the whole house becomes uninhabitable.

In a nutshell: New features help you attract users, while stability helps you retain them. The art of entrepreneurship lies in dynamically adjusting the balance between the two.