Articles I've Written Elsewhere 📌

I write on this blog weekly. Nobody edits me. Nobody tells me to "soften the tone" or "add more tactical takeaways."

I also pitch articles to other publications. They have editors. Standards. Opinions about what their audience wants.

Sometimes they publish them. Sometimes they send polite rejection emails that say "not quite right for us", which means "too weird" or "too honest" or "we already published something about onboarding last month."

Here are the ones that made it through, newest first.

HackerNoon likes tactical breakdowns of why things fail. I like writing about why things fail. Good match.

Mar 6, 2026Your AI Designer Has Never Disagreed With YouAI design tools generate, iterate and refine, and not once do they tell you the brief is wrong.HackerNoon

Dec 11, 2025Why Good Products Feel BrokenPolished, well funded and slightly wrong: products built around delivery milestones instead of the months people spend using them.HackerNoon

Nov 26, 2025Why SaaS Products Feel Harder to Use Every YearEvery sprint adds something, nothing gets removed, and the cognitive load compounds.HackerNoon

Nov 21, 2025The Developer Experience Tax Hidden Inside Your Design SystemDesign systems built for designers and handed to engineers, and what the friction costs once it piles up.HackerNoon

Nov 6, 2025Why Everyone in Your Design Review Knows Nothing Will Change and Shows Up AnywayEveryone in the room knows the feedback won't change anything. They come anyway, every week.HackerNoon

Oct 17, 2025The Design Industry's Senior Designer Addiction Is Bankrupting Your CompanySenior designers hired for roles that don't need them, and what that does to the people in those roles.HackerNoon

Oct 7, 2025The Seven-Step Hostage Situation You Call OnboardingSeven steps before a single feature, and the people they drive away are the ones the product needed most.HackerNoon

Sep 26, 2025Why I Only Design Mobile Apps for One Type of Client NowWhich mobile clients use design well, which waste it, and why narrowing down changed the work.HackerNoon

Sep 24, 2025The Button That Teaches People to Stop Paying AttentionA button that does nothing useful often enough gets ignored. So do the tooltips, the empty states and most onboarding copy.HackerNoon

Sep 10, 2025Your Product Feels Complex? Give It a SpineComplex products feel hard when there's no core logic to hang a mental model on. No amount of polish fixes that.HackerNoon

Sep 5, 2025Why SaaS Pricing Pages FailA feature on the table, then locked behind "upgrade to unlock" two days later, is a hostage situation.HackerNoon

The pricing one went furthest. Featured as a Top Story, 2,407 readers in three weeks, their newsletter twice, nine days on the homepage, translated into 12 languages, trending. Apparently "your pricing page is a confidence problem, not a design problem" resonated. If you think your pricing page is "just design", it will annoy you. Good.

Medium's Product Coalition focuses on team dynamics, hiring, and the messy human parts of building products. Less "here's a framework" and more "here's why your team is struggling."

Mar 17, 2026AI Said My Design Was Fine. Was It?I asked an AI to review a design. It was fine with everything, and that was the problem.Product Coalition

Jan 19, 2026Why Integration Setup Flows Fail: The UX of "Connected Successfully"The "connected successfully" screen is the most misleading moment in any integration flow.Bootcamp

Nov 26, 2025You Don't Need Perfect Soft SkillsTurns out the problem wasn't my communication. It was my willingness to disagree.Product Coalition

Oct 19, 2025The Cost of Always Hiring "the Best"A team of five senior designers all want to "lead strategy". Nobody wants to fix the button states.Product Coalition

You don't need better soft skills. You need better clients, or at least clients who don't think "Can you make it pop?" is actionable feedback. And credential inflation is expensive: it doesn't make better products.

Where the AI pieces went: design slop, bad defaults, and model collapse in the tools people use every day.

Jun 15, 2026We're in the AI Design Slop Era. Here's What That Means.AI-generated assets that look finished, feel familiar and say nothing – why slop spreads and who defaults to it.Built In

Feb 11, 2026Training Won't Fix Bad AI DesignMore prompting tutorials won't fix tools whose defaults are the problem.Built In

Feb 8, 2026When AI Tools Train on AI Output: Model Collapse in Daily WorkflowsModel collapse outside the research papers – slow, gradual, and noticed last by the people using the tools.Communications of the ACM

Jan 13, 2026AI Model Collapse in Design Tools: Why Your Background Remover Is Getting WorseSame photo. Same tool. Worse results six months later. It's not a bug, it's model collapse.SitePoint

When AI trains on content made by other AI, which is most of the internet now, each generation gets fuzzier. Your design tools are getting worse and almost nobody is talking about why, or what to do about it.

For founders about to spend money and product people about to lose an argument. And one list for BuzzFeed about working from home.

Jan 28, 2026What Founders Need to Know About Product UX Before Building Their First SaaSWhich UX decisions matter before launch and which can wait, for founders about to spend money on the wrong thing.Startup Nation

Jan 19, 2026How to Influence Product Decisions When You're Not the PMThe right argument in the right format, for when you don't have the title to override the decision.Mind the Product

Jan 19, 2026Things Nobody Tells You About Working From HomeThe remote-work list with the parts people leave out of their LinkedIn posts about async-first culture.BuzzFeed

Jan 8, 2026What Early-Stage Founders Should Know About Product Design Before Spending MoneyProduct design costs $32K–58K and most founders can't evaluate what they're paying for. The one metric that tells you it's working has nothing to do with how pretty it looks.Pitchwall

Feb 2026Looks Good to Me: On AI Sycophancy, Context Loss, and Inverted BaselinesNine structural problems under the AI productivity pitch: tools that only agree with you, context that decays with every exchange, edges and outliers going extinct, baselines that invert until broken becomes normal.Book

Not future speculation. Present observation. What actually happens when AI enters the design process – not what's supposed to happen, but what does – for designers, developers, product managers, and anyone working with or thinking critically about AI tools.

It's an ebook on Google Play Books, Apple Books and Amazon Kindle, and a paperback on Amazon.

Sometimes the editors ask the questions instead.

Jan 28, 2026Talk at Venture Café LondonA session at King's Cross on the design decisions that move product metrics, and why most UX advice early teams get is written for the wrong stage.Venture Café

Jan 26, 2026Interview: Tanya DonskaWhere the DNSK approach comes from, and what working between UX and product strategy looks like without a one-word title.Favourite Design

Jan 23, 2026UX Band-Aids That Hide Real ProblemsA podcast episode. Bad UX shows up in support ticket patterns before it shows up in the metrics.seat.fm

Jan 19, 2026Preparing for the Future of Work: Tanya Donska on the Top Five Trends to WatchWith Phil La Duke: why "soft skills" is code for "hire people who don't disagree", what credential inflation costs teams, and why the future of design is better designers, not more of them.Authority Magazine

Jan 14, 2026The Onboarding Step That Accelerates Time to ValueA Q&A on the one onboarding decision that sets how fast new users reach the moment they need the product.Consultant Magazine

Grokipedia is basically Wikipedia for people in the design and tech world. Not the watered-down version. The one where people actually write about what you did, not what you claim you do.

Jan 21, 2026Tanya DonskaA London-based UX/UI designer, Creative Director, and founder of DNSK WORK, a design studio specialising in product design for SaaS and enterprise software.Grokipedia

Having a profile there matters because it's third-party validation. I didn't write it. A publication did. Which means it has to pass some threshold of "actually worth documenting."

It shows up in searches, links back properly, and doesn't disappear when the algorithm changes. It's boring credibility infrastructure that actually works.

Reason 1: Instant audience

My blog gets 20–40 visitors a day, a couple of them me, checking the page still loads after a deploy. The HackerNoon pricing article found 2,407 readers in three weeks.

On your own site you build an audience from zero, and it stays near zero for a while. SEO takes 6–9 months to matter, and content compounds at savings-account speed.

A publication lends you its audience for a week or two. Some of those readers come over to the blog afterwards. Most don't, and I've stopped refreshing the analytics to check.

Reason 2: Editorial validation

"I wrote a blog post" means nothing to a client who's been burned by a designer before. "I wrote a HackerNoon Top Story that went out in their newsletter twice" gets a follow-up question instead of a polite "great".

Before it went out, someone with no reason to be nice to me read it and decided it was worth their readers' time. That's the whole signal, and clients pick it up faster than any case study.

I'd like the work on my own site to count the same as someone else's yes. It doesn't, so I pitch.

Reason 3: Better writing

Editors make you better, and they rarely do it by fixing your sentences. They ask questions you can't answer with "you know what I mean".

"Is this point clear?" "Does this example land?" "Can you cut 200 words?"

The third one stings, mostly because the answer is yes and the 200 words are the paragraph I liked most. My own blog never asks any of the three. I publish, find the problem a week later and fix it at 11pm, hoping nobody read it in between.

Reason 4: Constraints breed creativity

Every publication has a house style and a reader it's protecting.

HackerNoon wants the tactical breakdown of why something failed. Product Coalition wants the people problem underneath the process problem. Startup Nation wants founders to stop spending money they haven't made yet.

Writing to someone else's brief makes you adapt. The same point about onboarding comes out three different ways depending on who's reading, and you learn which way works by having it sent back.

Your own blog has no brief, so you can ramble. The draft in my folder called final-v3-actually-final says I do.

Reason 5: It's marketing that doesn't feel like marketing

I don't sponsor newsletters or pay to boost LinkedIn posts.

I write articles that fix one specific problem, a publication puts them in front of its readers, and some of those readers come looking for me. Some become clients.

It's still marketing. Nobody gets interrupted by it, and that's why it doesn't feel gross. The person reading about broken pricing pages went looking for that article, usually with their own pricing page open in the next tab.

I write what I'd write anyway and pitch it where it might fit. When an editor says no, it comes here instead, usually the same week and a little less polite.

Then the rest of this blog is for you. Same voice, same opinions, nobody between me and the publish button.

Read why marketing website case studies lie, what hiring a freelance UX designer says about how ready you are to commit, and what UX/UI debt needs before anyone says "redesign".

If you do UX research, start with the Fake Persona Test. The one that got the most replies: the UX iceberg problem.

On my own blog nobody asks for a "practical framework" in the last section, so I say the thing and stop.

Some of those posts could have used an editor. I know which ones, and I'm not telling.

Want to republish or translate any of these? Email me. I'm reasonable about it.

Want to see this thinking applied to a product? Here is what my UI/UX design work actually looks like.