26 Comments
User's avatar
Adam C. Siegel's avatar

Preach it, Michael! This is a massive problem that will only continue to increase. The struggle of the 21st century isn’t “does this exist?” but “is this signal or noise?”

Expand full comment
Jeff Giesea's avatar

Great thoughts Michael. We're planning to host an essay contest for the Boyd Institute, so this is useful info and we'd love to beta test your tool if that's of interest. Regarding your big picture thinking: if you can solve the discovery-of-quality problem and build an economic model around it, that would be good. lt will be hard to do both. I think the former problem will be easier to solve. In fact, I'd be curious to see a digest of the top-rated essays on Substack each week or month, according to your tool, just to discover writers and prioritize my reading.

Expand full comment
Michael Dean's avatar

Thanks Jeff, curious to hear more about your essay contest. Let's chat about it. Re: your essay digest idea, I've definitely thought about that. I think it could be valuable replacement for the engagement-based leaderboards. Definitely some thorny issues around doing this for the entirety of Substack (ie: if you hate the concept of quantification, which is, possibly a lot of people, how would you feel if suddenly your whole writing history was ranked based on a standard you don't yet understand?). I like starting with the competition model because it's opt-in. On the sooner side, I could see writers submitting URLs or RSS feeds to be featured on a leaderboard. TBD.

Expand full comment
Laura London's avatar

This seems awesome Micheal! I’m wondering—is your long term goal to use essay architecture to innovate in publishing? You mention a few details about this here and those potential projects sound exciting.

Expand full comment
Michael Dean's avatar

It's possible! For years I've been wanting to put together some kind of collection/anthology. The original goal was just to make it as good as possible and publish it traditionally, but now I'm starting to see new ways to experiment. I still very much believe in self-publishing over anything else, but I guess you could say I'm interested in ways to augment and unite independent writers.

Expand full comment
Will Mannon's avatar

Love this! Cannot wait for the contest, this will make waves on Substack and beyond. Also so many one-liner gems in this particular essay...here's to initial spark of the Golden Age of the Essay

Expand full comment
Dr. Anthony Howard's avatar

Love this idea

Expand full comment
Chris Coffman's avatar

Go Michael—what a great idea! I think it’s time to coin a neologism and replace “Platonic” with “Deantonic”

Expand full comment
Pamela Wang, PhD's avatar

This sounds pretty exciting. It does remind me of webnovel where the society itself is run by AI and there is a ‘system’ that evaluates all goods based off certain metrics quality, originality or even contribution to society. A way to enable meritocracy through clear testing.

It’s pretty important in a world where quality products are sliding towards OEMs leveraging their past reputation (E.g. Singer Heavy Duty sewing machine) and AI is making it so easy to make these low quality passable products (coloring books, fiction by AI where the writers did not even get beta-edits so the prompt was left in)

When is your software going to be available? I am a little confused about how the evaluation is going to work, do we have the ability to see the score change as we make edits, or is it a one time submission?

Expand full comment
Michael Dean's avatar

Interesting. Do you remember the name of that webnovel?

My launch date is 9/15. It takes around 15-25 minutes to analyze your essay, but you'll get feedback on each pattern every 45 seconds or so. Unfortunately it doesn't do live editing yet, since it runs about 1,000+ evals on your draft; it just takes time right now for it to be precise/accurate. But you're able to save your favorite editing suggestions to a checklist, implement them yourself, and then upload again to see the progress.

Expand full comment
Pamela Wang, PhD's avatar

It’s a trope I have seen in a few webnovels. There was a Mecha Design translated webnovel (where the system only judged mecha designs - it was a sci-fi intergalactic novel) and another one judged all goods. It has been too many years since I read it, so I can’t remember the names.

I see. Thanks!

Expand full comment
Charlie Bleecker's avatar

"Trite little Notes that read like greeting cards are getting more attention than entire editions of The Iowa Review."

Hahaha so good! This is awesome.

Expand full comment
Jessica J.J. Lutz's avatar

Very exciting, Michael. And indeed, we so urgently need something to preserve what makes writing human. Can't wait for more news on this initiative!

Expand full comment
Kelly Thompson TNWWY's avatar

I’m so excited for this!!! I’m in.

Expand full comment
American Woman 1984's avatar

So you will fund the prize money by selling the scoring rubric? I’m not criticizing but curious to know if I’ve understood and if that’s how it was designed to work.

Expand full comment
Michael Dean's avatar

Close, but I wouldn't say I'm selling the rubric. I plan to make all 27 patterns definitions public, so that will be like 2,500 words open to everyone (where most contests only go as far as "we look for writing that pushes boundaries"). This project started as an essay feedback app, which is what I'm selling (for $3-9 per essay). So everyone who submits is in the same position (they get feedback, scores, and the full book). But you're right in the sense that I'll be using software revenue to fund future prizes.

Expand full comment
American Woman 1984's avatar

Gotcha ! Cool idea. I like the idea of using algorithmic tools to make society and individual human lives better rather than worse. I like that you are balancing the judging with a human component, as writing is much like photography or music, something can be objectively perfect based on some defined rules of composition and still not touch someone in a meaningful way. Seems like using this framework you could have an open submission yet still award prizes in specific topics without needing a human to decide what goes where, and on the other side, writers don’t have to scan through a million RFS to find what fits their writing style. Hope it takes off! Maybe I’ll send something in when you get it going. Cheers.

Expand full comment
Davey's avatar

ur rocking it

Expand full comment
CansaFis Foote's avatar

…time to dust off the typewriter and unleash “the day that i became a turtle” onto you and the machine…

Expand full comment
Michael Dean's avatar

I had to edit my Experience pattern so that it understood the difference between an authentic human experience and an AI trying to imitate a human experience, but now I need to edit it again so it knows the difference between humans pretending to be turtles, AI pretending to be turtles, and authentic turtles.

Expand full comment
John Carpenter's avatar

This is cool. "AI judge as a service" is an interesting idea in its own right, too!

Expand full comment
Tom Pendergast's avatar

I love it!

Expand full comment
Jack Purdy's avatar

Brilliant model to incentivize great essays - excited to see the end result (and put together a submission!)

Expand full comment