<?xml version='1.0' encoding='utf-8'?>
<rss version="2.0"><channel><title>Agnieszka Besz — The notebook</title><link>https://besz.me/writing.html</link><description>Engineering, product direction, AI and leadership. Written from experience.</description><language>en-gb</language><item><title>Before You Remove Code Review, Name the Risk It Controls</title><link>https://besz.me/posts/before-you-remove-code-review-name-the-risk.html</link><guid>https://besz.me/posts/before-you-remove-code-review-name-the-risk.html</guid><description>Code review controls several different risks. A safer process routes changes to the evidence and human attention they actually need.</description><pubDate>Thu, 10 Sep 2026 00:00:00 +0000</pubDate></item><item><title>Use AI to Find Evidence, Not Score Engineers</title><link>https://besz.me/posts/use-ai-to-find-evidence-not-score-engineers.html</link><guid>https://besz.me/posts/use-ai-to-find-evidence-not-score-engineers.html</guid><description>AI can help managers recover evidence and challenge their reasoning, but it should never score engineers or make employment decisions.</description><pubDate>Tue, 01 Sep 2026 00:00:00 +0000</pubDate></item><item><title>Make Improvement Plans Winnable—or Don't Pretend</title><link>https://besz.me/posts/make-improvement-plans-winnable.html</link><guid>https://besz.me/posts/make-improvement-plans-winnable.html</guid><description>A performance improvement plan should diagnose the real problem, define observable expectations and give the employee a genuine chance to pass.</description><pubDate>Mon, 24 Aug 2026 00:00:00 +0000</pubDate></item><item><title>Feedback Is the Performance System</title><link>https://besz.me/posts/feedback-is-the-performance-system.html</link><guid>https://besz.me/posts/feedback-is-the-performance-system.html</guid><description>Annual reviews cannot replace clear, timely feedback grounded in real work, role expectations and evidence collected while it is fresh.</description><pubDate>Wed, 19 Aug 2026 00:00:00 +0000</pubDate></item><item><title>A Career Ladder Should Reduce Surprises, Not Create Bureaucracy</title><link>https://besz.me/posts/career-ladder-should-reduce-surprises.html</link><guid>https://besz.me/posts/career-ladder-should-reduce-surprises.html</guid><description>Career ladders should make expectations, evidence and promotion decisions easier to explain without turning growth into a checklist.</description><pubDate>Thu, 13 Aug 2026 00:00:00 +0000</pubDate></item><item><title>AI Makes Building Easy. Production Is Hard.</title><link>https://besz.me/posts/ai-makes-building-easy-production-is-hard.html</link><guid>https://besz.me/posts/ai-makes-building-easy-production-is-hard.html</guid><description>AI makes code, specifications and demos cheaper to produce. Production still depends on review, testing, controlled authority and human ownership.</description><pubDate>Mon, 10 Aug 2026 00:00:00 +0000</pubDate></item><item><title>Technical Direction Is Not a Technology Decision</title><link>https://besz.me/posts/technical-direction-is-not-a-technology-decision.html</link><guid>https://besz.me/posts/technical-direction-is-not-a-technology-decision.html</guid><description>Architecture debates go nowhere when the product direction is unclear. Once the outcome and tradeoffs are explicit, many technology choices become easier.</description><pubDate>Sat, 20 Jun 2026 00:00:00 +0000</pubDate></item><item><title>Prioritisation Is Constraint Management</title><link>https://besz.me/posts/prioritization-is-constraint-management.html</link><guid>https://besz.me/posts/prioritization-is-constraint-management.html</guid><description>Most teams have more good ideas than capacity. The job is to find the constraint holding the business back and put enough attention on it to make a difference.</description><pubDate>Sat, 20 Jun 2026 00:00:00 +0000</pubDate></item></channel></rss>