Thread by @Michael_J_Black

In the LLM-science discussion, I see a common misconception that science is a thing you do and that writing about it is separate and can be automated. I’ve written over 300 scientific papers and can assure you that science writing can’t be separated from science doing. Why? 1/18

Anyone who has taught knows the following is true. You think you understand something until you go to teach it. Explaining something to others reveals gaps in your understanding that you didn’t know you had. Well, writing a scientific paper is a form of teaching. 2/18

Your paper is teaching your reader about your hypothesis, problem, method, the prior work in the field, your results, and what it all means for future work. When you write up your work and find it challenging, this is typically because you don’t yet fully understand it. 3/18

The writing reveals what you don’t know. Years ago, Michal Irani gave me good advice. She said you can write the introduction to your paper long before the science is done and that this helps structure your thinking. 4/18

Of course, you have to rewrite it once you know the outcome of your work but she’s right. You can tell the story before the ending is known. I am constantly training my students how to structure a story because doing so leads to good science. 5/18

What’s the problem in the world? Why isn’t it solved? What's your key insight that lets you solve it when others couldn’t? Why is it novel? What’s your solution? Who’s your audience? What do they care about? What do they know? 6/18

None of these questions are about grammar. None are about English proficiency. They're about “thinking proficiency”. They're about understanding your contribution and why it’s important. They help structure an argument that's logical. 7/18

Use Grammarly, please. It's not cheating unless you’re taking a grammar test. By trying to write your story, you find the holes in your story. If you do this early, you have time to go back and let this drive your science. 8/18

Because people often have trouble getting started writing, I tell them to make a talk instead. The idea of telling a story to a room full of people makes it easier to get the structure right. If you have a really great talk, turning it into a great paper is relatively easy. 9/18

I have my students write a “shitty first draft”. All I care about is structure, logic, and story. We often sketch the figures and plots that we imagine on a whiteboard, take a photo, and put them in the paper. We build the whole structure before the work is done. 10/18

I guarantee that this leads to better science and better papers than rushing to write up something after you finish all the experiments. The early writing process leads to you realizing that you are missing experiments. 11/18

Most students would love for me to write their paper for them. They know I can do it faster. They also know that I’ll rewrite a lot of what they’ve written. So why should they do it? Because, it will help them become good scientists. 12/18

A good scientist has to be a good communicator. We don’t do science as a private activity. Imagine I structured all the papers and wrote them all and the students just coded and ran experiments. Or imagine an LLM did this. 13/18

One of the hardest things I ask them to do is to come up with an “elevator pitch”. This requires distilling something huge and complex into a sentence. To do so, you really have to understand the core of your contribution. 14/18

What about related work? Nobody reads that right? That could be automated, right? I think it's one of the most important parts of a paper for both the author and the field. 15/18

I am often grappling with a sprawling literature and my job is to organize it in some way that provides insight. This may let others see the field in a new light and can lead them to new discoveries. 16/18

I know that I’m fortunate. I’m a native English speaker, raised by literate parents, and the mechanics of writing come naturally. But I don’t care about grammar. I care about ideas, logic, and story. It’s your argument that matters. 17/18

Summary: science thinking, writing, and doing are inseparable. Focus on story. Write early. Write a shitty first draft. And do yourself a favor: write it yourself. I promise that writing about your science will improve your science. 18/18