ChatGPT did not write this post

Like everyone else, I’ve been watching the rise of “generative AI” with both interest and trepidation. (“Generative AI” is software that creates “new” text (ChatGPT) or images (DALL-E) from a user prompt – I’ll explain the quotes on “new” later.) Now, I know only a smattering about how generative AI works, so don’t expect technical insights here. But I’ve noticed an interesting gap between what I think these systems are doing and how people are reacting to them.

My interest in generative AI, especially text generators, is easily explained and probably obvious. Since I was in high school I’ve watched software get very slowly better at imitating the kind of writing humans do with great effort, and the kind of conversational interaction that humans do without a second thought.* The latest round is, superficially, really impressive: it can chatter pleasantly about nothing much, write a poem, program in R,** write an essay about Canadian history, explain linkage disequilibrium, and more. Or at least, it often looks like it can.

Advertisement

My goodness that’s a lot of squirrels

This is my 500^th post on Scientist Sees Squirrel*, and my goodness, that’s quite the logorrheic accomplishment.

When I started this blog, back in January of 2015, I really didn’t know where it was going. (Like many of my major life decisions, starting a blog wasn’t thoroughly thought out.) It’s astonishing to me that I’m still at it, nearly 7 years later; and that I’ve written 500 posts and thus, roughly, 550,000 words. That’s just a little bit less verbiage than War and Peace, and a little more than Harry Potter and the Order of the Phoenix (the doorstopperiest of the series) and Moby Dick (which, by the way, carries an important lesson for scientific writers) combined.

Like Dynamic Ecology, I know I’ll eventually stop spelling ‘banana’; but not yet.

Farewell (sort of) to Dynamic Ecology

Yesterday, the wonderful bloggers at Dynamic Ecology announced that they were hanging up their collective hat. Well, mostly; Dynamic Ecology will no longer have regular posts, but all its content will remain available and new posts may appear from time to time. (I hope!)

I want to take a moment to acknowledge just how good Dynamic Ecology was, and for how long. In part, that’s about content. There are so many posts on Dynamic Ecology that one could (one should!) keep going back to.

Blogging, writing practice, and self-discovery

I’ve been writing Scientist Sees Squirrel for almost 6½ years now – something on the order of 450 posts. With blogging being (supposedly) a dying form, and with a non-trivial amount of effort involved, you might wonder why I persist. There are lots of reasons, actually, but today I’m going to mention two: writing practice, and self-discovery.*

First, writing practice. As scientists, we write a shocking amount; in fact, being a writer is as much, maybe more, a part of our jobs as stats or teaching or experimental design. It’s not just papers – I write grant proposals, reports, administrative documents, and as you may have noticed, I’ve also written two books. So it might seem odd that I spend some of my time doing more writing.

2020 was weird for blogging, too

Warning: navel-gazing.

Did anyone else notice that 2020 was a really weird year?

OK, yes, you probably noticed. Lunatic wannabe despots trying to subvert elections; overwhelmed professors desperately struggling to move entire curricula online on a moment’s notice; idiots insisting that a scrap of cloth covering their mouth and nose is a fundamental infringement on their freedom. It was that kind of a year – thank goodness there’s now light at the end of the tunnel.

But you don’t want to read about that serious stuff, not this week, and not when you’d rather be enjoying that glimpse of the light.  So instead: 2020 was weird for blogging, too.  I mean, what on earth do you people want?

The blog post that dooms the universe

Warning: silly.

Got your attention, did I?

You know what got mine?  Noticing, a while ago, the apparently inexorable growth of interest in what I thought was a fairly dull* post, Friends Don’t Let Friends Use “cf.”, first published here in June 2016.  That post got a bunch of views when I first posted it, which isn’t unexpected.  Then it was largely ignored for a year or so, which isn’t unexpected either.  Then something odd happened: exponential growth.

That’s what’s shown in the graph above: month-by-month readership statistics for Friends Don’t Let Friends Use “cf.”.  It’s a lovely curve, isn’t it?  Let’s ignore the first year (which is dominated by novelty; every post gets a spike when first published).  Let’s make a semilog plot of the remainder, because that seems right for a curve like that.  And let’s fit a line to that semilog plot, because we’re scientists and we like to do that kind of thing.

Gosh, that’s a lot of squirrels – thoughts on 5 years’ worth

Warning: navel-gazing.

Scientist Sees Squirrel is five years old today.  That’s not very old for a human, a whale, or an oak tree, but it feels like something of an accomplishment for a blog.  So, no new post this week; instead, a few reflections on the squirrels along the way.

Metaphorical squirrels, that is.

Blogs are dying; long live blogs

Happy Boxing Day!  Which is also the Feast of Stephen, and although that’s obviously named for a much earlier Stephen, I do approve of feasts.

Anyway.

You hear a lot about how blogs are dying.  You’ve heard that for many years, actually, and to some extent it’s probably true: I gather that there are fewer “big” blogs than there were a decade ago, and the ones that are left worry about declining readership.  Among other things, some of the discourse that happened on blogs now happens, with obviously reduced quality, on Twitter* and other shorter-form social media.

You will not be surprised to hear me argue that there is still tremendous value in blogs – both in writing them and in reading them.

Science, social media, and my fear of corporate puppetry

Image: Marionette, © Thomas Quine CC BY 2.0 via flickr.com

When we do science, we presumably want that science to have both impact and reach.  By “impact”, I mean more than citation counts: I mean that what we’ve done adds to human knowledge and changes how we think about, and interact with, our world.  By “reach”, I mean that the impact happens broadly: not just with the six other people in the world who do research on the same questions and systems I do, but with scientists more broadly, with journalists, with policymakers, and with the general public.

Do I want my science (and my science commentary here at Scientist Sees Squirrel) to have impact and reach?  Of course I do.  It would be rather peculiar to publish science, and write a blog, and hope that nobody ever heard about it or was influenced by it.  So yes, I want my science, and my commentary, to have impact and reach.  But I’m also afraid of that impact and reach.  And while that seems very strange, even to me, I think it’s not uncommon and it distorts our scientific message.  Let me explain.

Internet trolls and integration by parts

Image: Don’t feed the trolls, © Sam Fentress CC BY-SA 3.0; plus integration by parts.

I’m usually pleased when people read my blog posts and ask questions about them.  Usually – but not when they’re trolls.

I ran into a very-likely-troll a couple of weeks ago.  I’d written a post I called Charles Darwin’s Other Mistake, about Darwin’s disdain for the use of authorities with Latin names.  It’s more interesting than it sounds – really – and it was picked up by Real Clear Science and (partly as a result) attracted quite a bit of readership.  And one reader left a question that had a distinctly suspicious odour to it.