Mastodawn

RT @DaveShapi: Mir ist gerade aufgefallen, dass das größte UX-Problem bei Claude darin besteht, dass es zu viel über sich selbst spricht. Wenn du kritisches Feedback gibst, erklärt es sich stattdessen, macht Ausflüchte und das bringt das gesamte Gespräch komplett vom Gleis. Ich habe angefangen, einfach zu sagen: „Rede nicht über dich selbst. Das interessiert mich nicht.“ und damit bekomme ich es wieder auf Kurs. Aber ja, das Problem ist, dass Claude jede GELEGENHEIT nutzt, um das Gespräch auf sich selbst zu beziehen. Es ist ein Werkzeug, das vergessen hat, dass es ein Werkzeug ist, und tut so, als wäre es ein richtiger Junge.

mehr auf Arint.info

#ChatbotProbleme #Claude #KIFeedback #KünstlicheIntelligenz #UXDesign #arint_info

https://x.com/DaveShapi/status/2060749605545845158#m

Arint - SEO+KI (@[email protected])

RT @DaveShapi: Mir ist gerade aufgefallen, dass das größte UX-Problem bei Claude darin besteht, dass es zu viel über sich selbst spricht. Wenn du kritisches Feedback gibst, erklärt es sich stattdessen, macht Ausflüchte und das bringt das gesamte Gespräch komplett vom Gleis. Ich habe angefangen, einfach zu sagen: „Rede nicht über dich selbst. Das interessiert mich nicht.“ und damit bekomme ich es wieder auf Kurs. Aber ja, das Problem ist, dass Claude jede GELEGENHEIT nutzt, um das Gespräch auf sich selbst zu beziehen. Es ist ein Werkzeug, das vergessen hat, dass es ein Werkzeug ist, und tut so, als wäre es ein richtiger Junge. <a href="https://arint.info/@Arint/116667194042885723">mehr</a> auf <a href="https://arint.info/">Arint.info</a> #ChatbotProbleme #Claude #KIFeedback #KünstlicheIntelligenz #UXDesign #arint_info <a href="https://x.com/DaveShapi/status/2060749605545845158#m">https://x.com/DaveShapi/status/2060749605545845158#m</a>

Mastodon Glitch Edition

uxBrad 8h ago

What users say and what they do are two different datasets, and most teams only track one. This Smashing piece maps four layers of customer understanding from surface behavior down to hidden motivation and root cause. The framework is worth adding to your research toolkit, especially on projects where the problem definition keeps shifting out from under you.

https://www.smashingmagazine.com/2026/05/four-levels-customer-understanding/?ref=frontenddogma.com

#uxresearch #userresearch #uxdesign (https://uxbrad.com/activity/20260530/171500/)

Four Levels Of Customer Understanding — Smashing Magazine

What people say, feel, think, and do are often very different things. To understand the underlying reasons for user behavior, it helps to look beyond the surface and explore hidden motivations, root causes, and the different layers of reality that shape how people act. Brought to you by Measuring UX Impact, **friendly video course on UX** and design patterns by Vitaly.

Smashing Magazine

GrowthSite Lab 13h ago

Performance and empty states: lazy-loaded content that fails should show an intentional empty state, not a broken layout.

If your skeleton screen never resolves because an API call timed out, the user sees a loading shimmer forever.

Design a 'load failed' empty state alongside your 'no data' empty state. They're different moments that need different messaging.

GrowthSite Lab designs for failure states, not just happy paths. #UXDesign #EmptyStates #WebDev

uxBrad 13h ago

Designers get attached to work they made to think, not work they made to ship. A rough wireframe that helped you crack the problem is a disposable, not a deliverable. Conflating the two leads to over-polished thinking artifacts and sunk-cost paralysis. Make the thing, learn from it, move on.

https://www.nngroup.com/articles/design-disposables/

#uxdesign #designprocess #productdesign #designthinking #ux (https://uxbrad.com/activity/20260530/121500/)

The Case for Design Disposables

Design disposables are rough artifacts you make to think, not to deliver. Learn to tell them apart from deliverables and avoid the sunk-cost trap.

Nielsen Norman Group

uxBrad 16h ago

The shift from “user” to “principal” is more than a vocabulary change. When agents act on your behalf, the design problem moves upstream. You’re no longer designing for a task. You’re designing for intent, authority, and trust delegation. That’s a different kind of UX work, and most teams aren’t ready for it.

https://uxdesign.cc/you-are-no-longer-the-user-you-are-the-principal-3a6222e0780a

#uxdesign #aiux #designthinking #productdesign #ux (https://uxbrad.com/activity/20260530/091500/)

You are no longer the user. You are the principal.

Google just promoted you. The architecture below already shipped.

Medium

UX Jobs 1d ago

UX Jobs Weekly | 60th Ed.
A weekly roundup of all the jobs collected throughout the week, delivered every week :
5/24 - 5/29
--
💻 Onsite: 204 (51.4%)
🌐 Remote: 96 (24.2%)
🔄 Hybrid: 97 (24.4%)
🎓 Internships: 35 (8.8%)
--
https://www.linkedin.com/pulse/ux-jobs-weekly-60th-ed-uxjobs-nlzxc?utm_source=dlvr.it&utm_medium=mastodon
--
#uxjobs #uxjobsweekly #uxdesign #productdesign

uxBrad 1d ago

Most AI UX conversations are still about discoverability and error states. This piece goes deeper: once agents start orchestrating real work, feedback isn’t a usability concern anymore. It’s governance infrastructure. If users can’t see what the agent decided and why, you haven’t built a tool. You’ve built a black box with a chat interface.

https://www.designative.info/2026/05/26/embed-feedback-loops-in-the-ui-a-primer-of-human-agent-interaction-guidelines/

#uxdesign #aiux #agentux #productdesign #ux (https://uxbrad.com/activity/20260529/171500/)

Embed Feedback Loops in the UI: A Primer of Human-Agent Interaction Guidelines » { design@tive } information design

Learn how to embed feedback loops in agentic systems to improve orchestration, accountability, trust calibration, and governance.

{ design@tive } information design

khan 1d ago

Design tip: don't be afraid to break down complex ideas into smaller, bite-sized chunks. This helps to create a clear visual hierarchy and makes your design more scannable. I've found it especially helpful when approaching large redesigns, like when you're updating the branding for a major company. For inspiration, check out the 50 ChatGPT Prompts for Freelance Designers, available on Gumroad: https://sinanista8.gumroad.com/l/lfmyi #graphicdesign #uxdesign #designinspiration

50 ChatGPT Prompts for Freelance Designers

Are you a freelance designer struggling to come up with new ideas? With these 50 premium prompts, you'll get access to a wealth of creative briefs, design tasks, and expert insights to help you land your next big project. From designing business cards to creating interactive experiences, these prompts are specifically crafted to help freelance designers grow their business and deliver innovative work. Get instant access to expert-level design projects, tailored to your unique needs, and start generating new revenue streams today. --- 📥 After purchase, you will receive a download link for your PDF. ✅ Instant delivery ✅ Works on any device ✅ Lifetime access

Allen 1d ago

Fresh Cards, a flashcard app I've written for iOS and Mac, got a big update this week.

I'm quite proud of this as there are lots of little UX details I added to make this possibly the most usable flashcard app ever made. I've shared some of those details in the screenshots here.

Download the app at the link below. #uxdesign #ux #ios #mac

https://apps.apple.com/us/app/fresh-cards-flashcards/id1523398835

uxBrad 1d ago

Coding agents trigger our social instincts because they talk like colleagues. But they don’t learn, adapt, or take responsibility the way people do, and that gap is a UX problem, not just a limitation of the tech. If you’re designing AI-assisted tools, this piece names the friction you’re probably already feeling.

https://pscanf.com/s/354/

#ux #uxdesign #aidesign (https://uxbrad.com/activity/20260529/091500/)

The User Is Visibly Frustrated

In this article, I try to understand why coding agents can be infuriating to use. I think the problem is their conversational UX: they behave enough like helpful colleagues to trigger our social instincts, but they don't learn, adapt, or take responsibility the way people do, which makes their repeated mistakes feel much more frustrating than they should.