Siri's iOS 27 beta 3 finally works, but Apple's rollout might be quietly rationing who gets to use it. In this Tuesday episode for week 28 of 2026, Grant and Maya put the beta through a full design critique, checking two years of WWDC promises against what actually shipped on July 6th.

This UX and UI teardown covers real product design tradeoffs, not just surface polish. The hosts dig into the Pace and Expressivity sliders gated to A19 Pro hardware, Siri's sprawling six entry points including the new Watch-centered standalone app, the system-wide Liquid Glass transparency slider and layered app icons, and the iCloud+ storage tier now required for Home's AI video summaries. They also debate a silent redownload bug that reverts Siri settings without warning.

• Whether hardware-gated features are honest constraints or bait-and-switch marketing
• How too many entry points can undermine discoverability
• Why static labels fail compared to live content previews in customization UI

Listeners walk away with a concrete takeaway for building gated features responsibly, plus sharper instincts for spotting product design promises that outpace shipped software.

Figma Motion launched at Config 2026 and it is more ambitious than it is finished. This is the Tuesday episode for week 27 of 2026, and Grant and Maya run a full design critique on Figma's native animation feature — defending what works, attacking what does not, and refusing to call it an After Effects killer just yet.

The UX and UI discussion gets specific: the Motion toolbar toggle's discoverability gap, the autokey behavior that lowers the floor for beginners while burying the easing curve editor that motion specialists actually need, the time-stamped comments tied to animation moments, and Dev Mode's inspectable timeline as a genuine handoff improvement. The missing Lottie export at launch and the Full seat paywall on publishing animated components both get honest scrutiny.

• Why the spring-to-CSS translation gap is a concrete product design failure mode Dev Mode does not fully solve
• How MCP-structured animation context changes what agentic coding tools receive from a Figma link
• Whether motion tokens actually survive into real design systems between sprints

Listeners will walk away knowing how to define duration and easing as design system variables before animating anything in Figma Motion — and why that one step prevents most of the handoff pain discussed in this episode.

Figma Make is posting impressive numbers — 60% weekly active usage among enterprise customers — but this week's autopsy argues the stat is doing a lot of heavy lifting. Recorded the Tuesday of Config 2026 week (week 26 of 2026), Grant and Maya tear into the design-to-production story before Dylan Field's keynote has a chance to clean it up.

This episode covers four specific failure points in Figma's MCP pipeline: the get_design_context / get_variable_defs / get_screenshot tool trio and why the model ignores two of them without explicit prompting, the monday.com engineering team's hard-coded color and overridden typography mess from naive Make codegen, the Design System Intelligence governance gap where detection works but no one owns the rules file, and Make's closed-beta production codebase integration — Grant's unconditional red flag heading into the keynote.

• Audit Code Connect coverage before evaluating any Make output
• Drop an MCP rules file in the repo root before touching Make
• WCAG annotations only carry through the pipeline if designers added them — Make-generated code still needs a full accessibility review
• Wall Street and designers watch the same 2 PM PDT investor session on June 24

Listeners will walk away knowing how to scope a Code Connect audit, what to watch in the keynote, and where the UX, UI, and product design gaps in Figma's current pipeline actually live. Screenshot a UI that bugs you and tag us — and subscribe so you do not miss tomorrow's keynote breakdown.

Figma's Design Agent shipped in beta before anyone could ratify it at Config — and that timing is not an accident. This is the Tuesday episode for week 25 of 2026, and Grant and Maya run a full design critique on the agent and the features surrounding it, with a spicy opening take: the agent's left-rail UX trades one set of problems for another, and the credit model is structured for enterprise procurement, not the individual designer doing exploratory work.

Specific elements under the microscope: the Design Agent's left-rail interaction model and its black-box reasoning problem, the tool-call approval dialogs borrowed from IDE consent UX, the PR workflow's Git-fluency requirement that quietly expands designer scope, and plan mode's upfront credit meter that trains designers to self-censor before exploring.

Grant also surfaces a 5.6x cost asymmetry buried in the credit pricing that most designers have not caught yet.

• Why the agent's self-correction loop drains credits in exactly the use case AI is most valuable
• How to read the ghost-seat arbitrage in Figma's credit pricing
• The linting-pass mental model from Check designs that applies regardless of your plan

Listeners will walk away knowing how to audit an AI tool's UX, pricing structure, and scope implications before committing to it. Subscribe, leave a review, and tag us with a UI that bugs you.

Apple's long-awaited Siri redesign finally shipped in iOS 27 — and the three-state interaction model is either a masterclass in progressive disclosure or a discoverability trap, depending on who you ask. Grant and Maya run the full WWDC 2026 Siri overhaul through their UX scorecard on this week's Tuesday episode for week 24 of 2026.

This design critique covers four specific elements: the Dynamic Island pill, transparent results card, and chatbot view as a three-state UI architecture; the Liquid Glass intensity slider Apple shipped as a transparency fix in iOS 27; the dark-only Siri results card and its ambient light legibility problem; and the iCloud-synced Siri app and the friction it creates alongside the Dynamic Island identity. The hosts also work through the EU availability gap as a product design constraint problem for developers shipping AI-dependent flows on day one.

• Defaults are design opinions — a slider added to fix them means the original opinion was wrong
• Every state in a progressive disclosure flow needs a visible exit, not just a gesture
• Contrast stacking risk is real when dark translucent overlays land on Liquid Glass backgrounds

Listeners will walk away knowing how to audit their own defaults as design decisions and how to map state transitions for visible exits in multi-step UX flows.

Apple's leaked three-layer Siri interaction model is a navigation problem dressed up as a design system — and this episode takes it apart. This is the Tuesday episode for week 23 of 2026, recorded six days before WWDC, and the UX and product design stakes are unusually high.

Grant and Maya run a full UI and design critique across the leaked iOS 27 Siri renders, covering four specific design elements: the stacked Dynamic Island pill, translucent overlay, and standalone dark-mode Siri app interaction states; the four-entry-point 'Search or Ask' trigger problem; Liquid Glass applied to a high-text-density chat surface; and the complete absence of error and uncertainty states in every leaked render. The $250 million Siri lawsuit settlement and Alan Dye's departure to Meta Reality Labs frame why this product design moment carries real consequences beyond aesthetics.

• Why three distinct Siri interaction layers create a model users cannot learn without a tutorial
• What the Vista Aero Glass parallel reveals about Liquid Glass on a full conversation surface
• How to name every interaction state with a trigger and exit so your own UI doesn't repeat this pattern

Subscribe and bring the three-point keynote scorecard to June 8.

This is the Tuesday episode for week 22 of 2026, and Figma's variables panel is more broken than its stock price — and that's saying something.

Grant and Maya run a full UX and product design critique of Figma heading into Config 2026, covering the financials, the tooling, and the AI roadmap with equal skepticism. The discussion gets specific: the variables panel's silent acceptance of circular aliases and type mismatches, the Dev Mode inspect panel's inability to surface system-level token structure, the pricing math on Dev seats versus full designer seats, and the Figma Make code output gap between Config marketing and production reality.

• 69.8% of professional teams use Figma Variables daily, making UI flaws in the panel an industry-wide UX problem
• Dev Mode answers element-level questions but leaves developers without a view of the full token system
• Figma Weave runs a separate credit meter alongside the core AI credit balance
• Adobe Firefly has generated 24 billion assets; Figma Make holds roughly 2% market share

Listeners will walk away knowing how to audit their own most time-costly Figma workflows and apply that lens to every Config 2026 announcement in real time. A practical exercise in UI, UX, and product design critique you can use before the keynote drops.

iOS 26's Liquid Glass redesign is the most contested visual overhaul Apple has shipped in a decade, and the adoption data suggests users agree. This is the Tuesday episode for week 21 of 2026, and Grant and Maya run a full UX and UI design critique of Apple's new design system — from its visionOS origins to its real-world failure modes on a 2D screen.

The episode works through four specific design elements: the translucent tab bar rendering text illegible over photo feeds, the Settings toggle animation that breaks affordance clarity, the Control Center layered over a photo lock screen, and the iOS 26.1 transparency toggle Apple quietly shipped in response to criticism. Grant and Maya also unpack the StatCounter methodology problem distorting the adoption headlines, present independent battery drain testing that showed iOS 26 burning 13% versus iOS 18's 1% across 150 routine actions, and examine what Alan Dye's departure to Meta signals about the design's internal reception at Apple.

• Why StatCounter's 15% figure overstates the gap — and why the hesitation is still real
• NN/g's documented legibility failures across Mail, Settings, and Control Center
• The spatial computing steelman: iOS 26 as a training ground for AR glasses users
• The one product design rule to apply tomorrow: build the accessibility-off state first

Listeners will walk away knowing how to pressure-test their own UI decisions against degraded-state conditions before shipping. Screenshot a UI that bugs you and tag us, and subscribe so you don't miss next week's audit.

Android 17 is Google's most opinionated platform release in years — and this week's autopsy argues they're making one genuinely smart structural bet alongside at least one unresolved accountability crisis. This is the Friday episode for week 20 of 2026.

Grant and Maya run a full UI and design critique across the release, covering four specific design decisions worth your attention: Sameer Samat's public Liquid Glass denial and what it signals about Material 3 Expressive as a product design strategy, Android 17 Beta 3's two confirmed blur interaction states — the widget picker and app launch transition — and whether tinted Dynamic Color blur is a principled system or just rebranded Liquid Glass, Gemini Intelligence's 'Create My Widget' generative UI feature and the unresolved question of who owns accessibility accountability when AI is composing contrast ratios and touch targets, and the color picker's four-preset plus hue slider structure as a progressive disclosure model worth stealing. Listeners will walk away knowing how to apply progressive disclosure to customization UIs, how to evaluate blur as a token-inheritance decision rather than a surface effect, and what questions to ask when AI enters your design system.