If you want a quick summary:
At the highest level, I want to build a research environment where each person (or two) owns their own projects. I value high agency and autonomy: when I give you something to do, either in a form of a task, an experiment, building a codebase, literature review, etc., I trust your ability to work independently to complete that task. By doing this day by day, you will gradually learn to do research independently: the ability to build on existing knowledge, connect patterns, and create something new from them. In some way, doing research is very similar to the creative process of making arts. The inspiration rarely comes out of nowhere. You need to have decent background knowledge to create meaningful things.
However, sometimes you will get stuck. This is completely normal, and here is my second piece of advice: a short feedback loop works well. Remember: if an idea fails, it should fail fast. More communication leads to faster iteration of an idea. Ideas are cheap, execution is more expensive, so figuring out which ones survive is harder. Don’t be afraid to open up and discuss your current findings, even if you think it is stupid. More than often, it is not.
To me, the ability to do good research also lies in the ability of critical thinking. Therefore, I love to work with people who can challenge me — a lot. Think of it this way: I hate being in a one-way relationship with a student that resembles a manager giving instructions to an employee. I generally dislike authority and hierarchy, so please do the same to me and just ignore them. You are also encouraged to do that with your friends in the lab, just do not mistake criticism with personal conflict: you are more than welcome to challenge the idea, but not the person.
Finally, at BAIL, when we work, we work, and when we stop, we stop. My personal experience tells me that blending relaxing time with working time does not do well. Think of mixing oil with water: they will not mix, however hard you try, and you will waste extra effort separating them later.