A novelist and artist named Stephen Marche made a point in The Guardian this week that I have not been able to stop thinking about. Marche has been using AI longer than most of us have been arguing about it -- his first algorithmically generated story ran in Wired in 2017, and he published the first AI-generated novel reviewed by the New York Times in 2023. His work is also on display in a museum in Sweden. 

His argument is pretty simple: AI is genuinely pretty solid at producing a really average output. The stuff that fills inboxes and clogs up editorial queues–that AI slop I talk so much about here. The stuff that, honestly, most of us spent years learning to create on deadline without thinking twice about it. So when he argues that  AI is a cliché machine, he is not speculating or fear-mongering. He has actually spent years trying to make real art with it, and he knows exactly where it falls apart.

What Marche is really saying is that mastery of the banal used to be a marketable skill. It was enough to prove you could write. You could produce clean copy fast, hit a word count, and match a house style, and that demonstrated professional competence. That skill has essentially no market value anymore, because a language model can do it in seconds, for free, at scale.

That is a genuinely uncomfortable thing to sit with if you have built a career on being reliable and fast. But I think it is also clarifying, in a way that is actually useful for senior journalists like us right now.

The journalists most exposed in this environment never built the part of the craft that cannot be automated: the source relationship that took three years of showing up before it paid off, the beat knowledge that tells you immediately when a number does not add up, the institutional memory that lets you contextualize a story in a way no algorithm can replicate, and the specific point of view that makes an editor open your pitch instead of the 40 others sitting in their inbox. There’s still real value in being quick; it’s just that now you have to add more value to your work. 

Marche uses chess as his analogy, and it is a good one. The current world chess champion, Gukesh Dommaraju, trained for years without AI before he ever touched an engine. His coach wanted him to develop his own instincts, his own creativity, his own feel for the game first. Only then did he use the machine as a tool. That order matters: A journalist who goes straight to AI before developing real reporting instincts is building on a very unsteady foundation. 

The other thing Marche points out is that language is more powerful now, more than ever. The people who built the architecture that underpins generative AI work believed, (against the prevailing research consensus at the time), that language was the key to abstraction. 

The ability to use language with precision and intention, to construct an argument, to ask the right question, to write a pitch that sounds like a specific human being with a specific reason to tell a specific story -- that is the whole game right now, and that’s what will make you stand out as a journalist in the world of AI-alop. The better you are at using language to write, think, AND prompt, the better journalist you can become.

While I am not a chess player myself (did you know you can learn to play Chess on Duolingo?), I am learning the, and one of the key things I’ve found is that having a coach matters. I’m by no means a Grand Master, but I’m getting better at it and have found it really helps with my strategic thinking and storytelling. Seriously.

To that end, my other half has built an app that provides individualized coaching to help you become fundamentally better at Chess. I’ve been using it to analyze my games on Chess.com.

He’s looking for a few beta testers (you’ll get the full app for free by signing up via this Google Doc link) to try it out and give feedback before releasing it into the wild. Once you sign up, you’ll get a link to download the app. Right now, it's only available to Android users, but he plans to open it to Apple users once he completes this round of testing.

The app is called Chessalyze, and it is a new Android app that imports your games from Chess.com and Lichess, then gives you move-by-move Stockfish analysis with accuracy ratings, move classifications (brilliant, blunder, and everything in between), and an interactive evaluation graph, plus player insights including strengths and weaknesses, all running locally on your device. You can rerun old games to learn where you went wrong, and you own all the data.

It is a robust week for jobs with lots of great opportunities both in freelance and full-time work.

This week’s freelance work includes opportunities with

This week’s Full Time work includes opportunities with

Dishonorables are BUMPING this week, as are Passion Projects. 

All jobs pay at least $1 per word and up to a huge $255,000 per year, and all the jobs on this list have been listed in the last seven days.