By Jesse Dunietz
Every researcher is also part writer. It’s a label that may be unfamiliar or even unwelcome to many graduate students, professors, and industry scientists. But between grants, papers, and reports to higher-ups, writing is undeniably a huge part of research.
Yet somehow, even with all that practice, the thought of writing for a mass-market magazine or news site can seem like a leap into a world so foreign that it’s unapproachable. The apparent chasm between us and a broader audience is further widened by the mathematical intensiveness of our work.
I didn’t start out being particularly comfortable with the prospect of mass media publishing. In fact, it took me the better part of a year to go from pondering the possibility of writing something to actually submitting a piece. I knew I enjoyed writing, and I felt a longstanding urge to share my love of all things science and computation. Even so, the notion of publishing anything public-facing felt alien, and I didn’t have a clue about the logistics of approaching an outlet.
What finally moved me was as simple as having to put pen to paper. I attended ComSciCon, a student-organized science communication workshop for graduate students where all attendees were expected to draft a piece. Once I ended up with even a partial draft, letting my work molder rather than polishing and submitting it somewhere felt like a shame. A few months after the workshop, my first piece was finally published online — all it had taken was that initial prod and a pointer to an editor. I was hooked.
Emboldened, I started writing for a student blog at CMU, even working my way up to submitting articles to online publications (turns out, the logistics pretty much consist of an email saying, “Hey, want me to write this piece for you?”). But I always hewed closely to my own expertise, writing about topics on which I could speak with personal authority. Finding sources, conducting interviews, and ramping up quickly on unfamiliar topics—in other words, actual reporting—still felt out of reach.
This time, the crucial nudge came from the AAAS Mass Media Fellowship program. Each summer, the fellowship embeds about two dozen STEM Ph.D. students and postdoctoral researchers in major newsrooms like the Los Angeles Times, the Washington Post, and Wired. Thanks to the generous sponsorship of SIAM, I was placed at Scientific American for 10 weeks this past summer. And while I was thrilled to be attempting science writing more seriously, I started the summer still cowering at the prospect of reporting.
Once again, sitting down and just getting started showed me how manageable the task really was. As soon as I jumped into the cycle of scanning upcoming abstracts and pounding out pieces, I had to begin contacting scientists all over the world. Within a few days, interviewing sources went from a major, anxiety-laden production to a casual routine repeated as often as five times a day. The world of the science writer again proved less foreign than I’d feared.
For example, one of my first stories described a new light-based computer chip that holds great promise for artificial intelligence (AI) (see Figure 1). Had I not already been familiar with the centrality of deep learning techniques to modern AI and of matrix multiplications to deep learning, I doubt I would have recognized the study’s potential impact. My mathematical knowledge also proved essential to explaining the technology itself, the crux of which is an optical component that performs an operation isomorphic to a 2×2 matrix multiplication. To describe the mechanics in detail, though, I would have had to explain how matrix multiplications implement weighted sums, how wave interference corresponds to matrix multiplications, and how changes in a light beam’s timing can alter interference patterns. After wrestling through several drafts with my editors, I ultimately managed to craft a high-level analogy involving sluices and pumps. Built on intuitions—developed over years of research—for what matrices really do, the analogy conveyed the key processes without getting bogged down in detail at each step.
The optical computing story wasn’t an exception, either. I had similar experiences covering a cryptographic protocol for genomic data privacy, techniques for tracing cryptocurrency crimes, and a new method for detecting neutrinos. Time and again, my mathematical knowledge enabled me to wade deep into the science and emerge with a clear, concrete explanation.
If you value writing and are hankering to contribute your knowledge to the wider world, I’d urge you not to let the unfamiliarity of science writing give you pause. Just take the plunge — the pragmatic hurdles are smaller than they seem, and your mathematical grounding makes you an especially valuable translator of complex topics. If you need help getting started, the science writing community is incredibly friendly and supportive. Of course, it helps to have direct supervision and feedback from editors, as I did this summer, but most people (including me!) will happily field questions about the publishing process from new writers. There are also fantastic public forums, like The Open Notebook, where established writers share advice, experiences, and thoughts.
Practically speaking, if you are a student or postdoctoral researcher, I can’t recommend the Mass Media Fellowship strongly enough (applications are now open; the deadline is January 15th). If you are past the postdoctoral stage, I’d still encourage you to look for other opportunities to write for audiences beyond academia, be it recurring freelance pieces, the occasional op-ed, or just a personal blog. There’s a real appetite for high-quality information about STEM, and that’s one thing to which we researchers can certainly cater. So embrace your inner writer and make your voice heard!