Disclaimer: This isn’t a Sandhills News, Rant or NC Local product—it's The Curious Scout, Diara J. Townes’ personal newsletter exploring culture, politics, and moral courage through community-centered storytelling.
All the Small Things
December is full of “last dance” energy for millions. Rushing to get gifts for family, submitting assignments, sending out emails, sealing final deals, and the like.
And this vibe is no different in the digital journalism, tech and non-profit spaces.
While I pushed to hit my own deadlines, it was blessedly intertwined with the annual release of the Nieman Lab Predictions for 2026.
This time of year is “eatin’ season,” as my sister calls it, and the Nieman Lab delivered journalists a gauntlet of a full-course industry meal.
So, I decided to use a little bit of AI and some old-fashioned copy/paste techniques to create a navigable spreadsheet for folks to filter and sort all 150 (and more to see) predictions.
But before I jump into what I hope is a bite-sized resource that makes it easier to absorb all the brilliance (and healthy contradictions) from experts who actually know what they're talking about, here’s how I got there.
Giphy
I Write Sins Not Tragedies (And AI Prompts)
Look, I had a Simple Plan: throw together a quick spreadsheet of Nieman Lab's 2026 predictions for my monthly newsletter. Easy-peasy (aw, Bob). My assumption about the number of articles caught me flat-footed, to say the least.
Giphy
A Boulevard of Broken Spreadsheets
I wanted to share out my top 2026 predictions for my final newsletter. But first, I needed to see them all.
Couldn’t be more than what, 30 entries? 40? 50?? Then, I’d list out my top five or so, and call it a day. I have, at best, basic b*tch data skills, so I thought I’d deploy some tech here.
Claude, the advanced AI assistant and large language models developed by Anthropic, rival to OpenAI’s ChatGPT, became my unwitting accomplice in what turned into a multi-day archaeological dig through digital journalism's collective anxiety about 2026.
Chop Suey! (The Breakdown)
I counted 50 predictions on the main continuous scroll homepage for the predictions. So, that’s how Claude set up the table.
We then attempted to scrape the Nieman Lab page, only to discover the site was blocking us with 403 errors. So I inspected the HTML myself and found 212 unique URLs. Claude pulled those but, of course 19 of them were broken links or just led back to the homepage.
Yikes. I deleted those. I had Claude read through the links to pull names and titles. It did what it could, searching other web-based sources to find what it couldn’t access, including personal websites of contributors and LinkedIn posts.
Then, Claude let me know we were still missing another 37 entries. I went back through the URLs in our new editable spreadsheet and manually added in the authors and the titles. I ran Claude again and it found ANOTHER 59 predictions, either wrong as a result of its own AI-brain mistakes, or simply not ran through when I initially sent the prompt with the unique URLs.
By now, I was In Too Deep.
A screenshot of Claude digging through the chaos at 3:45AM on December 30, 2025.
I had to add those back with placeholder fields highlighted in anxiety-inducing yellow.
This is how I discovered that two predictions was attributed to Sarah Marshall when it was actually Robin Kwong's. Fixed that. (How did that even happen?)
I renumbered everything after deleting duplicates. Verified every URL actually went with a “/2025/12/” slug and not some rogue 2024 holdover.
Then came The Great Affiliation Split of 2024™, where I decided one identifying column for contributors wasn't enough—I needed separate "Title/Role" and "Affiliation/Organization" fields.
Peak My Own Worst Enemy behavior.
By the next morning, after flying from North Carolina to New York to spend the end of the year with my sister/former roommate (both in childhood and adulthood) in the Greatest City in the World, Claude and I went through all 212 entries.
More than 75% of these needed manual detective work because organizations weren’t properly captured, or roles were incorrect, or summaries were plain missing. What's My Age Again? No, what's my job again? But alas, this was all my doing.
Felt more like this than not by the time 3am rolled around. Gif by greenday on Giphy
I also directed Claude to build a two-tier theme system because lordt, there were so many “yes-and” issues: 18 broad categories like "Artificial Intelligence,” "Business," and “Trust & Credibility,” then specific sub-themes like "Product" and "Community."
All Killer No Filler, as the kids say.
I’m using Claude to edit my summaries for every single prediction. Do you know how hard it is to summarize "AI will break the hamster wheel of journalism" in one sentence? It's already basically one sentence! But I did, its happening. Don't tell me this ain't a Basket Case situation.
And that's how you start accidentally building a journalism database: one prediction, a couple of panicked facepalms, and one punk rock reference at a time.
So, don't write your girl off yet— we’re only in The Middle of this data journalism project spiral.
What You're Getting: A Work-in-Progress Worth Sharing
Look, I didn't expect this to be such a lift. But I also didn't want to hold back what I had finished just because it wasn't perfect. So here's the deal:
The final product will be a fully searchable, filterable spreadsheet where you can sort by theme, filter by author, search sub-themes, and click directly to any prediction.
But because this is the last day of 2025 and The Curious Scout waits for no one (not even spreadsheet perfectionism), here's what's ready now—plus what's still cooking:
Sort by theme (40+ AI predictions if you're into doom-scrolling)
Filter by author (find all your faves)
Search by sub-theme ("Human Value" if you need hope, "Industry Crisis" if you don't)
Click directly to any prediction without hunting through Nieman Lab's site BUT with a direct link back to read the whole write-up
Actually understand what over 200 experts think is coming if you don’t have the time to read all 212 separate articles
If you find an error or a needed edit, let me know! Drop a comment in the spreadsheet!
Is this excessive? Totally.
Is it useful? I think so.
Did I learn more things about the state of digital journalism and its future that I wasn't spiritually prepared for? Absolutely.
So, you're welcome, I'm sorry, but more importantly, please enjoy because I'm not doing this again 💀
Scroll past jolly ol’ me and my Poodle-mix Butters for my hand-and AI-crafted guide to the predictions!
Diara J. Townes and Butters the Poodle-mix at the Cape Fear Botanical Gardens in Fayetteville, NC on Thursday, December 11, 2025. Photo by Dr. Dé Townes.
Fig. 1: Across 212 entries, 36 were categorized under the Artificial Intelligence theme. Data as of Dec. 31, 2025.
While brilliant minds were forecasting journalism's future (and I was compiling all their insights), I was also on the ground locally reporting for half the year in North Carolina—from citizens fighting fracking to families losing SNAP benefits.
My Top Five stories that I actually reported this year: A mix of top performers and personal favorites
That’s all for now!
You’ll hear from me every month with:
Updates on published stories and insights from current investigations
Commentary essays exploring the intersection of culture, politics, and what it means to show up with courage
Opportunities to share your perspective, your concerns, and ways to get involved
Occasional quick notes on local developments worth watching
Resource shares for those in tech, media and democracy spaces
Journalism isn't just about reporting in a community; it's about listening, asking, and finding answers together.
Thanks for being here. Please share this newsletter with others who are interested in local stories, cultural commentary, and how journalism gets made.