The Research Process

A live document collecting learnings about the research process as they come.

I will occasionally update this post with reflections on important questions about the research process, as they come.

How to choose research bets?

The creation of this post was spurred by a surprisingly high number of responses to a question I tweeted about. The responses seem too useful to lose to the broken X search algorithm (which Google search / LLMs may not be able to index well).

The question:

How do PhD students / researchers manage the sinking feeling of having a growing bucketlist of interesting ideas/directions but not enough time to try any of them?

How do you select when multiple seem promising?

At the time of copying this here, there are 48 comments. There was obviously no single answer, which is precisely why its helpful to look at the whole set of comments: https://x.com/ShashwatGoel7/status/1992606250698408155?s=20

I’m grateful for everyone who took out time from pursuing their research ideas to leave advice for my question. For the sake of those who don’t want to click and scroll through 50 comments, I’ll also paste my top 5 favourites (but different people would value different advice!):

by remembering that the hypothesis space is in fact infinite and that there is no way to possibly accomplish all of it; to instead look inwards towards my own desires and goals; to look outwards at the change i want to see in the world. knowing it is hopeless to try doing it all. - Glenn Matlin

You could work with coauthors and do a little bit more that way. But you kind of need your own lab with students to scale much more. You could try to recruit smart undergrads though. I suppose if you don’t care as much about author credit you could post the ideas and any student looking for stuff to do could run with them. Shashwat’s 17th problem. - Neal Parikh

many people suggested writing down, and tips on prioritization:

Write them down in a big list of one-off ideas and cluster them over time. If the same idea keeps coming up in different forms, there’s probably something larger interesting there! - William Merrill

By elimination. For example, you can rank them by expected impact, novelty, feasibility given the time frame and resources, personal obsession (i.e., would you regret not doing it), potential for follow-up publications, and whether the background or technique learned has general applicability for other research directions. LLMs made it possible now to brain storm the ideas/criteria with them, and even use their help to rank them. But you need to use good prompts to get objective results. - Mohammad Alfiky

and finally, the most liked one :)

One option is to just roll a 25 sided die.

Sometimes people get stuck by the tyranny of choice and it can be better to just pick randomly to offload that pressure. - Phillip Isola

---