Introducing SciTech

TL;DR

I’m proposing that people start using the term “SciTech” as follows: “SciTech is a category of software products that are used primarily by natural scientists and engineers.” There should be a SciTech community – conferences, job boards, blogs, etc – and software people should be excited to build it as a way to help contribute to fundamental advances in science and engineering.

Tech means software

There’s a naming convention in software product development that’s been around since I’ve been in the workforce. It goes like this: you take some real-world industry, think about software products that could improve that industry, and then slap the term “tech” at the end of the name of the industry, and now you have a ready-made, Silicon Valley-friendly buzzword for an exciting software product area. Examples:

EdTech = Education + software

FinTech = Finance + software

GovTech = Government + software = GovTech

AgTech = Agriculture + software

AdTech = Advertising + software

[Exception: BioTech doesn’t fit this pattern as it tends to mean all manner of technics related to biology, not just limited to software].

I spent about five years in and out of the “EdTech” world. There were EdTech meetup groups, conferences, podcasts, websites, and job boards. There were college students who were interested in looking for jobs in EdTech – at one point I was one of them. And there were, of course, EdTech investors and startup accelerators. You could get a Master’s degree in it.

In other words, EdTech was a useful organizing concept, not just a marketing term on a PowerPoint presentation (Though it definitely was also that). You could meaningfully talk about the “EdTech community” or “EdTech ecosystem.”

What about SciTech?

Following the formula above, you might expect to see this formulation to have shown up in the wild:

SciTech = Science + software

But it hasn’t. SciTech is just not a term you see around much. When you do, it doesn’t mean anything in particular. Its Wikipedia page disambiguates to Serbian science magazine, a high school in Pennsylvania, and an aeronautics conference, among other random references. Maybe the most common definition you see in the wild (for example, on the SciTech Reddit community) is what you could call the Trivia Night Definition:

SciTech = Anything related to science, nature, and/or technology, definitely not limited to software

So what?

But who really cares that SciTech doesn’t mean what it could theoretically mean?

The first problem is that, because the term never took off like FinTech, there are no SciTech conferences, no job boards, no podcasts, no college clubs, no meetups.

There is no SciTech community, and there should be.

The world desperately needs scientific and engineering solutions to the solve complex global problems of our time: severe climate and ecosytem imbalances, superbugs and infectious diseases, renewable agriculture. We need to invent cleaner and more sustainable ways to live, to work, to get around, to interact with the natural world. We need to make sure that scientists aren’t duplicating work unnecessarily and are collaborating effectively. You get the idea.

Every developer we can steer towards these great problems – and away from zero-sum activity like ad-tech or casual gaming – is a win for the world. Maybe this is too postmodern an argument, but I think that just having a good keyword gives people something to search for, something to point to when you’re a student or developer or entrepreneur or funder with a software background and you’re thinking, “I’d like to do this but not that.”

“I’m a business major and I’m interested in photography and SciTech.”

The Great Vision is that if we get more people to say something like that, that’ll mean more meetups, more Reddit activity, more business plan ideas, more funding, more progress for science.

But can non-scientists be helpful to science?

One of the replies I got regularly when I was looking in the SciTech job market was that there simply wasn’t a lot of need for non-scientists to serve as tool-builders for scientists. Like this comment from my Reddit post asking for opportunities where a software engineer could be useful to scientists:

Get a high paying job and donate excess income to your research field of choice. Most R&D is constrained by funding and not man power.

The attitude here – and it wasn’t the only time I got it – was that there are already enough scientists out in the world who can’t their work funded, so what the world needs is to fund them, not let unqualified software people meddle in their affairs.

I don’t agree. There’s obviously a place for fantastic software tools for research – and tools that don’t require having ultra-specific expertise in a domain.

Here’s my current listing of SciTech software product areas, with one example of an existing product that serves it:

Knowledge management. There’s more science out in the world than anyone can possibly fit in their head. Figuring out what to cram in one’s brain, what to save externally, how to share it efficiently, and how to organize it for future recall and to optimize for novel insights, and to make it computer-readable are all very hard problems that scientists face every day. Software has everything to do with all of this. (Example: Roam Research)

Bioinformatics. This is already a pretty established field, but I’ll just add that COVID-19 has underscored how essential computer models to understanding biological processes. (Example: Bedford Lab)

Experiment simulation environments. More powerful computers are enabling more high-fidelity simulation environments that allow scientific experimentation to happen to happen orders of magnitude more quickly. Similarly, for mechanical engineering, AI can help engineers “explore design space” by trying and simulating new product designs before they’re manufactured. (Example: DeepMind AlphaFold)

Data and methods collaboration. It’s true that perverse academic incentives make it often absurdly prohibitive for researchers to share data and methods. But at the same time, software tools should be out there to make that process of collaboration as easy as possible. (Example: protocols.io)

Research productivity. This is the general category of automating out as much of the tedious part of research as possible. (Example: Jupyter)

Imaging and visualization. Software has everything to do with MRI and advanced imaging and visualization. (Example: Randbee)

All of these projects have software engineers on staff who do not have a Ph.D.

In summary

SciTech may be a smallish discipline. The SciTech customer base – scientists and engineers – is always going to be smaller than, say, the theoretical customer base for FinTech, which you could argue is anybody who makes financial transactions.

And also, I’m not suggesting that sprinkling better software onto it is going to magically fix the state of science R&D. There are bigger issues – like the academic funding and publishing ecosystem – that get in the way of great science research.

But still, SciTech deserves its place in the pantheon of Techs. Let’s make it a thing. Let’s do the conferences, and the job boards, and the podcasts. Let’s make the world of science less inaccessible to programmers without doctorate degrees.

If we can use it as a way to steer just one precious young mind away from AdTech, I’ll take that as a win.

I’ve claimed a fresh new SciTech_ Reddit page. Come join.