Boach, Bugbergs and bottlenecks_ going deeper with AI, not just faster
Categories: Podcasts , MOT This Week in Testing
The Bug Berg challenge uses iceberg metaphors to reveal hidden quality issues like technical debt and poor test data beneath visible bug reports. Participants also examined how AI accelerates code generation but shifts bottlenecks to review phases, creating a quality gap and faster silos.
MOT This Week in Testing
MOT - This week in Testing - Varied hosts, group chat, often with community questions and involvement. Show notes have a full transcript.
Episode Details
- Show Notes: https://www.ministryoftesting.com/podcasts/this-week-in-testing?wchannelid=czgwdadw2c&wmediaid=ebow0om92y
- Published: 2026-11-20T16:07:56Z
- Duration: 59:36
- Author: Unknown
Overview
The discussion centers on the Bug Book challenge, an initiative inviting participants to create “bug-bergs” - iceberg-style diagrams illustrating what is visible versus hidden in software quality work. Several community-authored bug-bergs are examined: Justin Holsgrove’s maps root causes like environments, test data, technical debt, and poorly defined acceptance criteria beneath visible symptoms such as bug reports and “works on my machine” responses. Melissa’s reframes testing beyond test cases and automation coverage to include planning, execution, continuous improvement, and stakeholder expectations. Simon’s maps quality-professional skills - creativity, critical thinking, risk awareness above water; care, listening, empathy, systems thinking, and joy below - using AI tools (Claude and Gemini) to generate the diagram.
A recurring theme is how AI reshapes quality work. Participants note that AI empowers customers to surface issues themselves, shifting bottlenecks from code production to review and triage. The “leaf blower” metaphor captures this: faster code generation merely pushes the bottleneck downstream. A quality person describes developers completing all sprint work on day one, creating a quality gap where testing cannot keep pace. Another speaker warns that AI can create “faster silos” when individuals implement tools without team coordination.
Accessibility advocacy is reframed through testability: because AI-generated test cases become flaky, framing accessibility standards as a testability fix gains organizational traction. Community meetups in London, Brighton, and Manchester reinforce that in-person conversations surface realities people withhold online, adding context that online profiles cannot.
What If
-
What if you mapped your product’s recurring bug reports as a bug-berg before your next release cycle?
- Move: List every visible symptom (bug reports, hotfixes, user workarounds, “works on my machine” replies) above the waterline, then force yourself to enumerate at least five hidden causes below it - environment gaps, unseeded test data, undefined acceptance criteria, technical debt, and untested edge-case scenarios.
- Why Now?: The community challenge closes on the 30th, and the exercise mirrors how customers now use AI to surface issues you may not have anticipated, giving you a structured pre-release audit.
- Expected Upside: A proposal: you may catch root-cause clusters before they become customer-facing AI-generated complaint reports, shifting you from reactive triage to proactive prevention.
-
What if you reframed your flaky AI-generated test suite as a testability problem and fixed it via accessibility standards?
- Move: Identify the three most flaky AI-generated test cases, trace each failure to a missing accessibility attribute (ARIA labels, focus order, semantic markup), and implement the accessibility fix as the testability remediation.
- Why Now?: Sean reports this framing is “getting a lot more traction” in organizations where direct accessibility arguments previously stalled, making it a viable lever for solo developers who lack a dedicated accessibility advocate.
- Expected Upside: A possibility: you reduce flake rates while simultaneously improving accessibility compliance, gaining two quality wins from a single corrective action.
-
What if you ran a personal “leaf blower audit” on your AI-assisted workflow this week?
- Move: Track where AI accelerates your output (code generation, documentation drafting) and explicitly log where the bottleneck lands - review time, triage queues, or production bug volume. Write down the single downstream step that is now overloaded.
- Why Now?: The speakers describe AI shifting bottlenecks rather than eliminating them; without a personal audit you may be unknowingly pushing your own quality debt into the next sprint or into customer hands.
- Expected Upside: A proposal: identifying the shifted bottleneck lets you allocate your limited solo time to the actual constraint rather than the visible speed gain, preventing the “quality gap” from widening.
Takeaway
-
Before each release, sketch a bug-berg listing visible symptoms above the waterline and at least five hidden root causes (environments, test data, technical debt, acceptance criteria, untested scenarios) below it, mirroring Justin Holsgrove’s root-cause mapping.
-
When AI-generated test cases turn flaky, implement the relevant accessibility standard as the fix and frame the work as a testability improvement to gain stakeholder buy-in, as Sean reports this framing is gaining organizational traction.
-
Audit your AI-assisted workflow weekly: log where speed gains land and where the bottleneck shifts (review, triage, production bugs), so you address the real constraint rather than the visible output rate.
-
Before adopting an AI-generated tool or framework solo, share it with your team or community first to prevent the “faster silos” pattern where one person’s Claude-generated solution overrides others’ work without coordination.
-
Attend at least one in-person meetup or chapter event per month; participants report that face-to-face conversations surface operational realities and personal context that online channels withhold, adding context you cannot get from profiles or posts.
For a PDF of longer Software Testing Podcast Episode Summaries with Briefing Notes and more detailed summary notes, visit EvilTester Patreon Podcast Summaries.