As we dive headlong into a new year, artificial intelligence continues to loom large, like an alien spaceship hunkered down in the sky above the Empire State Building. It is the most intriguing class of technologies I’ve seen since the advent of the web in the late 1990s. I use it, and my work has benefited greatly from AI transcriptions, summaries, research, and editing feedback.

But it’s not omnipotent. I predict that we will increasingly recognize the limits of AI in 2026.

Kids are already using “that’s AI” instead of “I don’t believe you.” Many of those I talk with are annoyed by the flood of obviously AI-generated content in their social media streams. The word of the year was slop. The New York Times’s Sam Kriss wrote a great piece on the ubiquity (and mediocrity) of the new “AI-ese.” Personally, I am starting to feel a real aversion to AI imagery.

And in 2025, I saw way too many examples of copy that clients had created using AI, and which they were asking me to edit or fix.

Subscribe to Rough Drafts

A deeply personal, professional newsletter that grows out of decades of writing, editing, and leadership communications, along with my dedication to the arts of deep listening and mindful living. 

by Dylan ~ About 1,000 subscribers ~ Roughly 2-3 posts/month

I had mixed feelings about this. On the one hand, this AI copy and my clients’ questions proved to me why my services are so valuable. On the other hand, fixing AI copy is a drag. In most cases it’s just easier to start over from scratch.

In 2026, this realization will hit people in many domains, from coding to content. This, I hope, will lead us to focus more on where AI is truly helpful, discarding the overblown claims and breathless hype.

~ ~ ~

This may sound tautological, but if you’re looking to stand out from the sea of AI slop, be wary of relying too much on AI. Used carelessly, it pulls your writing toward the mediocre middle rather than lifting it above the noise.

The problem, as with any new tool, is that people haven’t learned to use AI well yet. In writing, it is obvious when someone relies too much on AI to draft their copy. Those who use it well are in control of the process, and don’t try to use it as a substitute for their own judgement, feelings, dreams, and experiences.

So here’s a tip for all you writers: Do not start a writing project with AI. Never use a chatbot to create an outline or a first draft. If you do, the AI copy will seem too good to you, and you’ll overlook its obvious flaws. You may miss major structural or factual problems. In short, you’ll be going down the wrong road.

Fixing bad AI copy is much harder than using AI to fix your own bad copy.

Instead, write your own outline. Bang out a crummy first draft. Then use AI for feedback and to help fill in gaps.

Try this prompt with your favorite chatbot, alongside your draft:

I’m attaching a draft of a (blog post, email, byline, report) that I’ve written about (brief topic description). I am writing for (target audience), and I’d like them to (conclusions, ideas, or actions you’d like your audience to take from your writing). Please examine this draft as if you were a professional editor and give me specific feedback on what’s working, what could be improved, and what is missing.

Field notes

This is the part of the newsletter where I share a few interesting links. I’m starting my 11th year of delivering content and communications results to tech companies and execs. So, for this issue, I’m going to share some highlights from the past year of work at Tweney Media, in three categories.

Enterprise tech thought leadership:

AI, responsibly: We worked with a wide range of PWC principals and executives to create content including two white papers and a number of blog posts on Responsible AI. One of those projects included planning, interviewing principals, developing the content, and writing this series of eight blog posts. PWC: Delivering Responsible AI.

Agents of chaos: GitLab CISO Josh Lemos and the GitLab comms team worked with Tweney Media to create this byline, published in Dark Reading. We did dozens of other pieces with GitLab in 2025, helping the company hone its messages for developers, security professionals, and operations pros (DevSecOps). GitLab’s Josh Lemos

The $2M outage: We worked with Reidy Communications to support New Relic’s report on the observability industry. Our contribution included data analysis, creating storylines, outlining, and writing both the main forecast and several regional reports. New Relic Observability Forecast

Complex tech stories and media products:

Onsite power: Tweney Media client Bloom Energy makes clean, compact fuel cells that are increasingly in demand for data centers that want to generate their own electrical power. Through a series of sponsored posts, blog posts, and a white paper, we helped Bloom differentiate their complex technology and tell a timely story about how it plays a crucial role in enabling AI. Taking power into their own hands

A 100x opportunity: Mayfield Fund consponsored the September Imagination in Action event at Google and Stanford. Mayfield wanted to maximize its investment and create something more lasting than the two-day event. Tweney Media brought in a team of three veteran tech reporters and worked with a photographer and designer to create this 60-page digital magazine. It made a big splash and helped Mayfield cement its vision and leadership in AI. Future of AI

Coaching, workshops, and more:

First-generation leaders: We continued to support the nonprofit COOP Careers with writing workshops for their fellows, all first-generation college graduates aiming at careers in digital marketing or data analytics. COOP Careers

Workshops and presentations: I helped a PR firm hone the art of writing effective pitches to media, taught an 80-person marketing team powerful frameworks for effective copywriting, and delivered one of the best-attended Forbes workshops of 2025, on the art of thought leadership. I also gave a couple of speeches about my career as a tech journalist, one of which was was a huge hit at the Golden Gate Breakfast Club: Dylan’s speech on the future of technology

Poetry and deep listening: In the summer, I became a Certified Listener Poet: a nonprofit called The Good Listening Project trained me to hold space for people in health care (providers and patients alike), listen to them deeply, and then write poems for them. This might seem tangential to my tech content work but it comes from the same well: Being curious, asking good questions, listening deeply, and weaving stories that move hearts and minds. TGLP

This is just a sampling. Tweney Media worked with many other companies throughout the year, including agencies Mission North and PANBlast Communications, to create industry-leading content and to hone the abilities and writing quality of staff members.

I’m sharing all this work to show that there’s still a strong need for human-centered content creation. Sometimes, the right prompt can make a big difference. But other times, you need the help of a seasoned editor (and, occasionally, a team of veteran writers) to create writing that makes a real difference, as our clients demonstrate.

If your organization is struggling to create content that stands apart from the sea of AI slop, let’s talk. My consultancy, Tweney Media, has helped PWC, GitLab, New Relic, Bloom Energy, and Mayfield elevate their brands, differentiate their viewpoints, and expand their audiences through thought leadership. Our writing, editing, and coaching work combines journalistic rigor with tech fluency with deep listening. We’d love to help you stand out from the crowd.