Weekly Digest

What matters in AI for designers — this week.

5 things to know · 3 tools to try · 1 skill to learn

01

5 Things to Know

Ideas and shifts worth your attention this week.

01

10 UI patterns that won't survive the AI shift

UX

A useful gut-check on which conventions are quietly becoming legacy: dropdown-heavy filters, empty states that just sit there, multi-step forms, dashboards that wait to be read. As intent moves into language and agents, screens stop being the destination. Worth auditing your current product against this list — not to rip everything out, but to spot patterns you're still defaulting to out of habit rather than user need.

Read more →
02

Mapping AI presence to user intent

UX

Not every interaction needs a copilot. This piece argues for matching AI's visibility to what the user is actually trying to do — invisible when they want flow, present when they want a partner, explicit when they want control. It's a sharper framework than 'add a chat panel' and gives you language to push back on stakeholders who want AI sprinkled across every screen. A keeper for design reviews.

Read more →
03

Nine accessibility myths, debunked with data

Insight

Stephanie Walter's roundup goes after the talking points that still derail accessibility conversations — that it's expensive, niche, or only matters for compliance — with actual numbers. Bookmark it for the next time a PM asks if alt text is really worth the sprint. The same issue surfaces an underrated point: a lot of a11y failures aren't design problems, they're engineering ones, which changes how you should be advocating internally.

Read more →
04

The forgotten-conversation problem in AI chat

UX

Chat interfaces are great at the message you just sent and terrible at the one from twenty turns ago. This piece names a UX gap most teams ignore: users lose track of what the AI knows, agreed to, or already tried. If you're designing anything conversational — assistants, support, onboarding — it's a prompt to design memory as an interface element, not a backend feature. Think summaries, pinned context, visible state.

Read more →
05

Staff designers aren't there to ship the best work

Career

After rewriting seven Staff and Principal portfolios, the author lands on something useful for anyone eyeing senior IC roles: the job stops being about your pixels. It's about lifting the ceiling on everyone else's. If your portfolio still leads with hero shots and solo case studies, you're applying for a Senior role under a Staff title. Reframe around leverage, ambiguity, and decisions you unblocked — not screens you owned.

Read more →
02

3 Tools to Try

Practical AI tools worth your time this week.

Inspiration

Extra

Visit →

From ex-Pinterest folks, Extra is a new visual collection tool aimed squarely at how designers actually gather references — moodboards, screenshots, fragments — without Pinterest's algorithmic noise. If you've been hacking together inspiration in Figma files, Are.na, and a screenshots folder, this is worth a look this week.

Color

Perplexity for color systems

Visit →

Not a new tool, but a new use: a designer walks through using Perplexity to build a perceptually uniform, accessibility-checked color triad — and documents where the AI gets confused. Useful as both a workflow recipe and a reality check for building tokens or refreshing a palette.

Productivity

TinyStart

Visit →

A tiny, fast launcher for macOS — think Spotlight stripped back to the essentials. If you're stacking AI tools, design apps, and browser tabs, a keystroke-driven launcher is one of those quiet productivity upgrades that compounds. Free, lightweight, and the kind of thing you install once and forget you're using.

03

1 Skill to Learn

A concrete skill you can start practicing today.

Workflow

Audit your product for AI-era UI debt

Before redesigning anything, learn to spot the patterns in your current product that won't age well: filter-heavy interfaces, dead empty states, dashboards that don't react to intent. A structured audit you can run in an afternoon — useful ammo for roadmap conversations.

  1. 1Pick three core flows and screenshot them. Mark every dropdown, filter, and multi-step form — these are your highest-risk legacy patterns.
  2. 2For each screen, write the user's actual intent in one sentence. Flag anywhere the UI makes them translate that intent into clicks.
  3. 3Identify two places where AI presence should be invisible, one where it should assist, and one where it should be fully explicit.
  4. 4Turn the audit into a one-page doc with before/after sketches and bring it to your next product review as a discussion starter.

Get this in your inbox every week.

Free. Weekly. No noise — just signal.