Taste & Craft: A Conversation with Tuomas Artman, CTO Linear & Gergely Orosz, @pragmaticengineer

Key Concepts

• Product Engineering: A role that combines technical implementation with deep product sense, customer empathy, and design intuition.
• Tasteful Software: Software built with high attention to detail, user experience (UX), and aesthetic quality, rather than just feature volume.
• Zero Bug Policy: A methodology where bugs are treated as the highest priority, fixed immediately upon discovery, and never relegated to a long-term backlog.
• Quality Wednesdays: A recurring team practice where engineers proactively identify and fix minute UI/UX imperfections.
• AI-Assisted Development: Using AI agents to handle repetitive tasks (like bug fixes) while acknowledging that AI currently lacks "taste" and human-centric design intuition.

---

1. The "Feature Factory" Trap and the AI Pendulum

The speakers argue that the industry is currently trending toward a "feature factory" model. Because AI tools (like Claude Code) make it trivial to ship code, companies are tempted to ship every requested feature immediately.

• The Risk: Shipping without deep thought leads to convoluted, confusing software.
• The Philosophy: Great products are born from saying "no" to 999 ideas to focus on one perfect solution.
• The Shift: Previously, engineering difficulty acted as a natural "gate" that forced developers to think before building. With AI, that gate is gone, making intentionality more important than ever.

2. Maintaining Quality in a Hyper-Growth Environment

Drawing from experiences at Uber, the speakers highlight the danger of prioritizing metrics (like revenue) over quality.

• The "Gradual Slip": When companies prioritize speed to outpace competitors, quality degrades slowly. Users may not leave immediately, but they will eventually migrate to higher-quality alternatives.
• Measuring Quality: It is notoriously difficult to measure quality via A/B testing. Instead, it requires a cultural commitment to excellence, such as the "two-pixel" standard observed in early Uber development.

3. Methodologies for High-Quality Engineering

Linear employs specific frameworks to maintain their high bar:

• Quality Wednesdays: A 30-minute weekly meeting where 25+ engineers present one small quality fix they found themselves. This practice has resulted in 2,500–3,000 minor improvements, fostering a culture where engineers are constantly scanning for "polish" issues.
• Zero Bug Policy:
  • Process: When a bug is reported, it is assigned immediately. It becomes the engineer's top priority.
  • The Trade-off: If the backlog is large, the company stops new feature development entirely for a period (e.g., three weeks) to clear the debt.
  • Benefit: It prevents the "backlog rot" where hundreds of bugs accumulate, and it delights users who see their reports fixed within hours.

4. The Role of AI Agents

While AI is powerful, the speakers emphasize its current limitations:

• AI lacks "Taste": AI can write code, but it cannot judge if an animation feels "natural" or if a two-second load time is "too slow" for a specific user context.
• Current Utility: AI is highly effective at fixing bugs (10% of Linear’s bugs are currently fixed by AI without human intervention) and summarizing customer feedback.
• The Future: The speakers predict that AI will eventually handle 100% of routine bug fixes, allowing engineers to focus entirely on product strategy and UX.

5. Hiring and Culture

Linear’s hiring process is designed to filter for "Product Engineers":

• The Week-Long Trial: Every candidate works on a full-week, paid project. This allows the team to see how the candidate handles a project from start to finish.
• Customer Exposure: Engineers are immersed in a "fire hose" of customer feedback. They have access to Slack channels with enterprise customers and searchable recordings of all customer meetings. This ensures that even engineers building for large corporations understand the end-user's pain points.

6. Future Outlook: The Evolution of the Engineer

The speakers predict that the role of the software engineer will shift toward becoming a "Product Engineer."

• The Shift: As AI automates data piping and routine coding, the value of an engineer will lie in their ability to act as a "mini-PM"—understanding customer needs, designing intuitive experiences, and making product decisions.
• Actionable Advice for Growth:
  1. Get closer to customers: Engage directly with user feedback.
  2. Build for yourself: The best way to learn product sense is to build and ship your own projects.
  3. Study Design: Read resources like Apple’s Human Interface Guidelines to develop a foundational understanding of good UX.

Synthesis

The core takeaway is that in an era of AI-driven speed, quality and intentionality are the ultimate competitive advantages. By implementing rigorous internal practices like "Quality Wednesdays" and a "Zero Bug Policy," and by shifting the engineering focus from mere code production to holistic product ownership, companies can avoid the trap of building bloated, low-quality software and instead create products that users prefer over time.