Mariusz
banner
mariusz.cc
Mariusz
@mariusz.cc
Product / Service / UX Design, Leadership, Ethics, Yelling at clouds. ⚡️ Languages. ⚡️ Taking photos. ⚡️ Screamy music ⚡️ Often seen dumping links to what I read. 🔴⚫️

I don’t know what HD is but my doctor says I have eighty of them.
Pinned
Mariusz @mariusz.cc · Nov 29
I think I've taken close to 10 thousand photos with many different cameras at different budgets when doing street and travel photography and yet this old photo taken with a shitty phone camera might still be my favorite.
AI interfaces become truly intelligent when they anticipate rather than react. Instead of waiting for users to articulate perfect prompts, contextual UI patterns can surface intent naturally. Users spend less time explaining and more time refining the results.

uxdesign.cc/how-can-ai-u...
How can AI UI capture intent?
Exploring contextual prompt patterns that capture user intent as it is typed
uxdesign.cc
December 11, 2025 at 9:20 AM
Reposted by Mariusz
With the news that the US wants to have five years of your social media history allow me to remind you of the tool I built to scrub your skeets
GitHub - Gorcenski/skeeter-deleter: A script for auto-deleting Bluesky posts
A script for auto-deleting Bluesky posts. Contribute to Gorcenski/skeeter-deleter development by creating an account on GitHub.
github.com
December 11, 2025 at 1:46 AM
Reposted by Mariusz
People keep wondering why they're going after Venezuela. It's because they want to do regime change and control its oil, then use a big tract of land there to keep a tax free "network state" that looks like Dubai but operates like Rhodesia. They are saying this out loud! Believe them!
December 10, 2025 at 6:57 PM
Reposted by Mariusz
there's like a curve graph for how the badness of UI relates to consolidation and competitiveness in the market. is there like a named law for this yet
The way tech companies constantly do stuff that nobody asked for really tells you a lot about their bigger ideology. They tell themselves that you will like it over time. What they mean is you’ll learn to endure it, or you’ll give up. quality of UI is not actually anywhere on their radar
new iOS update is sleek and also defaults in Music to “you must want songs transition as they would on a dance floor.” so many people will wonder why they can’t just listen to an album. Apps > Music > Song Transitions > toggle off.
December 9, 2025 at 7:51 PM
Reposted by Mariusz
ur replies when you say something bad about LLMs
December 10, 2025 at 12:22 PM
“Good enough” doesn’t mean sloppy. It means understanding what matters right now versus what might matter later. You need code that works reliably for your current users and can evolve as you learn more.

www.trevorlasn.com/blog/good-en...
Good Enough Is a Strategy
Your competitors will eat your lunch while you refactor
www.trevorlasn.com
December 8, 2025 at 2:51 PM
UX folks typically gravitate to conventional metrics, not because they are the best for measuring UX but because they’re easy to collect and report. Yet, despite their ease of collection, conventional analytics make it almost impossible to tell a story.

articles.centercentre.com/how-the-righ...
How the Right UX Metrics Show Game-Changing Value - UX Articles by Center Centre
There’s a story I want to tell you. It’s a story that emerged out of my recent work with a UX team. They used this story to excite their executives on how recent UX improvements made a substantial…
articles.centercentre.com
December 8, 2025 at 9:43 AM
Why do they all have Slavic last names 😭
Lasota’s friends, colleagues, and family talk about how a NASA intern and Google employee wound up at the center of a bizarre string of murders across the country.

Rolling Stone reports ↓
The Radicalization of Ziz Lasota: How an AI Doomer Became an Accused Cult Leader
bit.ly
December 6, 2025 at 4:05 PM
Definining components in data form rather than passing unstructured artifacts between collaborators unlocks many opportunities across a component’s life cycle.

medium.com/@nathanacurt...
Components as Data
How to define components independent of platforms to scale a system’s impact
medium.com
December 5, 2025 at 2:41 PM
Lofi YouTube is trying too hard by now.
December 5, 2025 at 11:12 AM
You are creative when you see things others don’t. Not necessarily new visuals, but new correlations. Connections between concepts. Problems that aren’t obvious until someone points them out. And you can’t see what you’re not exposed to.

hvpandya.com/exposure?ref...
Exposure
For equal amount of design skills, your exposure to the world determines how effective of a designer you can be.
hvpandya.com
December 5, 2025 at 9:44 AM
It’s not always possible to change your organization’s business process (selling heads), but you can influence it and try to measure your designs so that your “project for a month” turns into a “short-term product” with goals and measurable results.

uxplanet.org/measuring-ux...
Measuring UX: Your First Step Towards Objective Evaluation
Have you ever experienced a situation where a good design was rejected just because the decision maker woke up on the wrong side of the…
uxplanet.org
December 4, 2025 at 2:39 PM
Design thrives when it helps product and engineering discover, frame, and de-risk the right problems. But when “product” becomes the dominant paradigm used to construct an operating model, calling design “product design” doesn’t help differentiate the unique value.

medium.com/design-bootc...
The value of Design in a product organisation
Clickbait opening: There’s no such thing as Product Design
medium.com
December 4, 2025 at 9:20 AM
The “don’t be evil” was a marketing thing at best anyway, they were always fine to work with things like law enforcement in totalitarian countries.
Googles new motto: We’re evil. Deal with it.
It’s actually up to governments to regulate AI and stop it from destroying lives.
December 3, 2025 at 6:28 PM
That’s the paradox: the plan became obsolete, but the act of planning together made us capable of executing even as everything changed.

www.productparty.us/p/the-planni...
The Planning Paradox: Why your plans are useless, but planning isn't.
How the act of planning drives teams to launch while the plan itself becomes obsolete.
www.productparty.us
December 1, 2025 at 2:51 PM
Reposted by Mariusz
We can't really say this enough...

> Anastasia Berg [at UCL Irvine] said that new research — and what she's hearing directly from colleagues across various industries — shows that employees who heavily rely on AI are losing core skills at a startling rate.

www.businessinsider.com/ai-tools-are...
AI tools are 'deskilling' workers, philosophy professor says
A philosophy professor warns that AI reliance is weakening workers' judgment, creativity, and problem-solving.
www.businessinsider.com
November 30, 2025 at 6:09 PM
Users overvalue what they already have by a factor of three—the familiarity, the muscle memory, the sense of control. And companies overvalue what they’re offering by a factor of three—because they built it, they know every feature, they see the potential.

eleganthack.com/ux-is-your-m...
UX Is Your Moat (And You’re Ignoring It)
Last week, Google released Nano Banana Pro, their latest image generator. The demos looked impressive. I opened Gemini to try it. Then I had a question I needed to ask. Something unrelated to image…
eleganthack.com
December 1, 2025 at 9:43 AM
I’m convinced that people who are surprised average people are putting sensitive data into AI models instead of going to a doctor have not experienced majority of doctors.

As someone who dealt with some medical issues - most doctors will prescribe ibuprofen if you come in missing a leg.
November 30, 2025 at 11:12 PM
Reposted by Mariusz
If you are checking LLM-extruded drivel against a real source anyway, why not just use that source? Why not skip the LLM part entirely?

(it's because in practice, no one actually checks; these things are designed to discourage you from going "off tool")
AI Chatbots Discourage Error Checking
AI hallucinations threaten the usefulness of LLM-generated text in professional environments, but today’s LLMs encourage users to take outputs at face value.
www.nngroup.com
November 30, 2025 at 8:58 AM
Reposted by Mariusz
If only we had learned from every previous time this happened that doing something on the cheap is only cheaper if that thing actually works; when it inevitably doesn't work you will have to pay for the failure and then pay full price (plus a rush fee) to have the job done properly.
"Astonishingly, the textbook companies had invested the equivalent of $567 million in the project, to meet the government’s commitment of $850 million."

At what point are actual monetary costs too high, these days?

futurism.com/future-socie...
South Korea's Experiment in AI Textbooks Ends in Disaster
The South Korean government's billion-dollar AI textbook initiative has been a complete failure from start to finish.
futurism.com
November 30, 2025 at 9:17 AM
“Back in the day there was no such thing as autism”

Yeah bro and that’s why grandpa spent entire day in his shed building train models after a minor argument.
Do you have a relative who keeps talking about how "RFK Jr. has a point about vaccines and autism?" Good news. My book We're Not Broken is on sale and Harper Collins has free shipping in the continental U.S. until December 8. Buy it here.
www.harpercollins.com/products/wer...
We're Not Broken
“This book is a message from autistic people to their parents, friends, teachers, coworkers and doctors showing what life is like on the spectrum. It’s also...
www.harpercollins.com
November 29, 2025 at 11:17 AM
A skill is a document (often markdown) containing instructions, constraints, and domain knowledge, stored in a designated directory that Claude can access through simple file-reading tools.

claude.com/blog/improvi...
Improving frontend design through Skills | Claude
Best practices for building richer, more customized frontend design with Claude and Skills.
claude.com
November 28, 2025 at 2:41 PM
For years, design systems chased efficiency — faster builds, fewer inconsistencies, more adoption. The next wave is intelligence: systems that understand the meaning behind their components so they can adapt.

www.designsystemscollective.com/beyond-compo...
Beyond Components: Designing for Meaning in an Agent-Assisted Era
Design systems were supposed to free us. They promised consistency, velocity, and sanity in the age of infinite screens.
www.designsystemscollective.com
November 28, 2025 at 9:44 AM
An efficient, well-oiled product team that ships at high velocity can only do so when design can fit its process into the sprint-based system, and development is also on board with that process.

uxdesign.cc/let-designer...
Let designers think
How “Thinking” + “Designing” need to be practiced outside AI.
uxdesign.cc
November 27, 2025 at 2:39 PM
Top-down traditionalists see AI as a way to amplify the output of their feature factories. Evidence-guided development enthusiasts believe that AI will help us overcome human biases and make us better at observing the world and testing our ideas.

itamargilad.com/product-disc...
Four Product Discovery Models: A Practical Map - Itamar Gilad
A clear map of the four ways companies decide what to build. Why command-and-control isn't the best and how to avoid gradual drift.
itamargilad.com
November 27, 2025 at 9:20 AM