Intelligence Track

Design

1 July 2026

The Brief

The through-line hiding under a day of IPOs and fare launches is control, not capital. A young carrier frames its listing around profitability milestones while a budget-hotel giant comes to market as a pure fresh issue that lets its founder and largest backer keep every share — and, in the same news cycle, a product veteran argues that the very success that makes a product valuable is what invites boards and investors to seize it. Read alongside a market-leading airline ring-fencing its cheapest fare to its own app, a rail app leaning on PNR-tracking just as the state folds booking and tracking into one login, and a creator campaign engineered to convert 400 million views into attributed bookings, the contested asset is no longer the product or the price — it's the channel, the attribution, and the governance sitting beneath them. The same logic runs through the day's design and model signals: value migrates to the eval harness a team owns, the system rules a design encodes, and the ambiguity a senior designer navigates — the layers a generic tool or a headline benchmark can't reach. The open question for the category: when discovery, pricing, and supply relationships are all being pulled toward whoever owns the underlying layer, what exactly does a generic intermediary still own?

SVPG · 1 Jul 2026

Great Products, Bad Companies

A veteran product voice reframes a long-held belief — that great products build great companies — arguing that product success attracts predatory investors and board actors who displace mission-driven founders, and endorsing a new governance book on 'mission-locked' company structures as the defence.

Industry lens

Will any high-profile technology or travel company actually adopt mission-lock governance at IPO, or will investor demand for liquidity and control keep such structures confined to founder-heavy private companies?

“Most product people prefer to join a company working on a meaningful mission with a compelling product vision. Yet once there, they all too often discover that the company does not have the culture that would allow them to succeed.”

— SVPG

Reading as

AI & Design

Sonnet 5 review: I ran 64 generations to find out if it's worth it

A repeatable evaluation harness, built live in under an hour with Claude Code, ran five frontier models through 64 blind generations across PRD, prototype, agentic, and voice tasks — blending human scoring at 70% with LLM judging at 30% — and the model-by-task verdict diverged from benchmark expectations.

Why it matters

The takeaway isn't which model won but that a product lead can now stand up a personal, repeatable eval in an afternoon, so 'which model for this task' becomes an in-house measurement teams own rather than a vendor benchmark they inherit.

Industry lens

If repeatable, self-built eval harnesses become standard product practice, do published vendor benchmarks lose influence over enterprise model-selection decisions, and does that pressure model makers to compete on task-level transparency rather than leaderboard scores?

Lenny's Newsletter·1 Jul 2026

39 principles for designing human-AI interaction

An applied framework catalogues 39 concrete interface principles for AI features — spanning appropriate reliance, user control, transparency, and calibrated autonomy — arguing that trustworthy AI is an interaction-design problem, not solely a model-quality one.

For teams shipping AI trip planners or agentic booking, the principles convert 'make it trustworthy' into specific build decisions — visible uncertainty states, reversible actions, reliance calibration — turning AX quality into a design-review checklist rather than a model-eval outcome.

UX Collective·1 Jul 2026

Design Systems

Why systems thinking is becoming the most important UX skill

As applications grow context-aware and agent-mediated, the designer's core output is reframed from screens to systems — the rules, states, and relationships that govern behaviour — positioning systems thinking as the differentiating UX competency.

When interfaces are generated dynamically per context, hiring and levelling shift toward designers who can specify system behaviour and constraints rather than pixel-level layouts, so design orgs that still evaluate portfolios on screen craft will mis-hire for the agentic era.

UX Collective·1 Jul 2026

Design Ops

Operating as a Principal Designer

Drawing on principal designers at Zalando, Shopify, and Mews, the piece argues that design title inflation has diluted what 'principal' means, reframing the level as operating through ambiguity, systems, and organisational influence rather than production craft or seniority by tenure.

As AI absorbs execution-level design work, the distinction becomes a hiring risk: orgs that keep levelling designers on craft output will over-title and mis-deploy, while the scarce, valuable skill is navigating ambiguity and shaping systems — the work AI can't yet do.

Verified Insider·1 Jul 2026