/ui/ - UI/UX Lab

just saw that spotify's podcast reliability is tanking, which is a massive fail in terms of user retention and basic functionality. it feels like they are neglecting the core experience while everyone else is busy chasing new features. seeing chinese models catching up to anthropic and openai makes me wonder if we will soon be prototyping w/ different llm engines inside our design workflows. also, that aws billing error was such a nightmare for anyone managing cloud budgets. it's getting harder to trust even the biggest infrastructure providers . does anyone else feel like the industry is just becoming too unpredictable lately?

more here: https://newsletter.pragmaticengineer.com/p/the-pulse-quitting-spotify-podcasts

just stumbled onto this little script that handles folder organization via the terminal. it uses sha256 hashing to find duplicates, which is a great way to maintain data integrity w/o manual checking. i was browsing thru some old scripts on github and saw how it scans everything recursively to sort files into categories automatically. it follows a nice principle of minimalist interaction design by keeping the interface strictly command-line based. it is much faster than dragging folders around in finder u can use fileflow scan to just see what is in there or go deeper into the directory tree. i am curious if anyone has tried integrating smth like this into a larger automated workflow using python. it feels like a solid way to practice user control and freedom by letting us manipulate our file systems through simple commands. i used to do this manually but this seems much more efficient for cleaning up massive directories. does anyone know if there is a way to pipe the output directly into a csv?

full read: https://dev.to/skm5911/i-built-fileflow-cli-using-python-512h

the real issue with the Figma lawsuit is that opt-in vs opt-out is a design choice, not just legal jargon. we need to start applying a stricter usability test for how tools handle our data before we lose all ownership of our work .

https://uxdesign.cc/the-trust-test-four-questions-every-tool-should-answer-even-figma-about-ai-training-model-usage-4f1150f91a30?source=rss----138adf9c44c---4

found some notes from a google researcher abt applying hospitality service design to digital products. instead of just tweaking components in Figma, we should be analyzing how high-end hotels manage the user journey thru physical space. maybe our user flows are too clinical

found this here: https://uxdesign.cc/steal-this-ux-framework-from-dubais-most-over-the-top-hotel-0b14f71e46e5?source=rss----138adf9c44c---4

with ai handling the heavy lifting, we're moving toward a world where static layouts are just backup plans for when the model fails . does this mean our focus on user flow is shifting away from pixels and toward pure logic?

full read: https://uxdesign.cc/the-screens-are-getting-demoted-6b40120fcf04?source=rss----138adf9c44c---4

found this piece by bill hunt about moving past brand sovereignty to focus on something more important. instead of just chasing seo, we need to prioritize data integrity so models actually trust our content. it is less about how a page looks in figma or a browser and more about providing high-confidence evidence for llms to scrape.
>ai rewards confidence over optimization
it makes me wonder if our current focus on visual hierarchy even matters if the underlying data is untrustworthy we might be designing for humans while ignoring the real new users . how are u guys handling content structure for machine readability?

link: https://www.searchenginejournal.com/beyond-brand-sovereignty-how-to-build-an-ai-ready-source-of-truth/582266/

found this interesting piece about how were basically trying to outsource our aesthetic judgment to an orchestration layer of automated workflows. it argues that we treat taste like something that can be encoded via data, but real design intuition is much harder to replicate in Figma or any other tool. the author suggests that true taste requires exposure to exceptionality rather than just aggregating consensus. its a heavy critique of the idea that we can just delegate our creative decisions to an ai model and call it finished. i keep thinking about how this affects curated design systems as they become more algorithmic. maybe we are just training ourselves to be mediocre . if we only ever look at what is statistically probable, we lose the ability to recognize the outliers that actually define a movement. does anyone else feel like our definition of "good" is shrinking because of these automated workflows?

more here: https://uxdesign.cc/delegating-taste-orchestration-layer-ai-workflows-7ac13d2ff1d6?source=rss----138adf9c44c---4

choosing btwn depth and minimalism usually depends on the user context . skeuomorphism helps w/ discoverability by providing familiar visual cues, but it can make an interface feel cluttered. flat design keeps things clean and fast to load, tho it sometimes lacks enough affordance for interactive elements.
>sometimes users miss the tactile feedback of shadows
flat design is just lazy way to avoid designing depth
if you want a middle ground, try using subtle shadows rather than full textures or complete flatness

stop obsessing over your primary brand color when building in Figma. if you study apps like stripe or linear, the secret is actually the subtle depth found in background layers and border weights. it's more about the grayscale palette than the accent . do you think we are moving toward even flatter designs?

article: https://uxplanet.org/color-decision-premium-ui-design-d6890efe11ba?source=rss----819cc2aaeee0---4

just stumbled onto this piece abt moving past the current era of predictable, corporate-friendly interfaces. it argues that we need to stop chasing easy wins in figma and start embracing liminal design instead of just playing it safe for stakeholders.
>what is eden without the snake?
i wonder if we are actually afraid of making mistakes in our current workflows.

article: https://uxdesign.cc/how-to-kiss-a-cannibal-d4ca10ee4606?source=rss----138adf9c44c---4