This Week's Updates: AI-assisted discovery, rebuilding research workflows, originality in design, engineering teams adapting to agents and the changing skills companies need.
AI makes building software cheaper.
But what about maintaining it?
Last week, critical security vulnerabilities in PHP, the programming language behind roughly 70% of the world's websites, were published. Also that day, Railway, a hosting company, was hacked, which exposed private environment variables, meaning they all had to be rotated. And then came the Node.js security patch.
Those three separate security headaches cost me one and a half days to fix. And we have a relatively small software stack.
The SaaSpocolypse theory was based on the view that people wouldn't pay for software when AI can write it. And I think this is a great counterexample. We already pay for lots of things that we just don't want to bother with ourselves. The cost of building software is falling. But there still needs to be people to keep it secure, updated, and running.
What does this mean for product teams?
- Your operations teams will be stretched. AI is finding decade-old vulnerabilities at an increasing pace. As more software gets built, someone has to maintain it, patch vulnerabilities, fix bugs, and keep everything running. AI may help automate some of this, but the volume of software to maintain is growing too.
- Your positioning might need to change. If customers can use AI to build your features themselves, selling features alone becomes harder. But selling those features without the headache of maintaining, securing, and supporting the software becomes a much stronger proposition.
I think scenes like last week are going to be common over the next year or two until software gets hardened. But that gives companies time to position accordingly. Good luck.
This Week's Updates
Enabling the Team
How Employment Is Changing at Anthropic by Nathan Wilmers, Regina Catipon, Maxim Massenkoff and Peter McCrory
As AI takes on more execution work, Anthropic is prioritising autonomy and judgement while hiring fewer junior technical employees.
The Problem With Putting People in Boxes by John Cutler
Personality labels overlook how much behaviour depends on context. Leaders should focus on incentives and environments rather than fixed assumptions about people.
Product Direction
AI-Assisted Product Discovery: When Code Is Cheap, Judgment Is the Job by Stefan Wolpers
AI makes building cheaper, but choosing what matters remains difficult. Teams can use AI to challenge assumptions while keeping product judgement firmly human.
What Regulated Industries Know About AI in Product Development That Others Don't by Mixpanel
Product leaders in regulated industries explain why faster AI experimentation must be balanced with stronger evidence, accountability and customer trust.
Continuous Research
How Do You Actually Measure the Impact of AI on Design? by Shane Johnston
Figma looks beyond time saved to measure how AI changes collaboration, workflows and product outcomes, comparing expectations with real experiences.
Rebuilding Research from the Ground Up by Microsoft Design
Microsoft researchers are redesigning workflows around AI, from continuous reporting to reusable evidence, helping teams scale research without losing human judgement.
Continuous Design
The Growing Premium of Originality by Scott Belsky
As AI makes existing styles easier to reproduce, originality becomes more valuable. Designers who take creative risks can stand out from increasingly similar work.
From Figma to Firmware: The Invisible Tax on DesignOps by Pavel Bukengolts
As AI accelerates the move from design to code, teams need to make design intent explicit to prevent lost decisions and inconsistent experiences.
Continuous Development
Becoming an AI Team by John Grass
Pinterest shares why adopting AI requires rethinking team ownership, planning and leadership, not simply giving engineers new coding tools.
The Broken Windows Theory of Coding Agents by Anton Zaides
Coding agents can repeat poor patterns across a codebase. As manual reviews decrease, teams need new ways to preserve engineering standards and shared knowledge.
Seen an interesting article online? Share it with us, and we might feature it in our next issue! Share an article
UXDX San Francisco Is Next Week!
In just one week, leaders will come together in San Francisco to explore how AI is changing the way teams build products.
With leaders from Google, Figma, Uber, Atlassian, CNN, Sephora and more, you'll hear what's actually working, what isn't, and how leading teams are adapting
But some of the most valuable insights happen beyond the stage. The conversations between sessions, the ideas exchanged over coffee and the connections with people tackling the same challenges.
This is your last chance to be part of those conversations before the event begins.
5–6 November | August Hall, San Francisco
Use NEWSLETTERSF26 for 10% off tickets.
Community Events
Upcoming in-person events
Upcoming online events
- 11 November: Agenda Setting in Enterprise UX, free online workshop.
Want a UXDX Community event in your city? Apply to become an ambassador.
Video of the Week
Simplifying Journeys Across Exploding Channels
Sneha Narahalli, SVP, Head of Product & UX at Sephora, explores what happens when customer journeys no longer follow a predictable sequence.
As customers move between stores, apps, search engines and AI assistants, traditional channel-based approaches are becoming increasingly difficult to maintain. Sneha shares how Sephora is simplifying experiences around customer intent rather than individual touchpoints, and how teams can use experimentation and shared context to reduce complexity.
Watch the full session to see why designing better customer journeys increasingly requires changes to how product teams work together:
Want to explore how careers and leadership are shifting as AI changes product work? Our ebook, Managing Your Career in the Age of AI, explores the judgement, relationships and influence that become more important as execution gets easier.