--- title: You can just choose how many bugs you want now | daily.dev description: AI coding tools have made finding bugs nearly free, since agents can surface as many correctness, performance, accessibility, or reliability issues as you ask... canonical: https://daily.dev/posts/you-can-just-choose-how-many-bugs-you-want-now-yybwzht7p twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: You can just choose how many bugs you want now | daily.dev og:description: AI coding tools have made finding bugs nearly free, since agents can surface as many correctness, performance, accessibility, or reliability issues as you ask... og:url: https://daily.dev/posts/you-can-just-choose-how-many-bugs-you-want-now-yybwzht7p og:image: https://api.daily.dev/og/posts/yybwzHt7p.png og:image:alt: You can just choose how many bugs you want now og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # You can just choose how many bugs you want now **[Read the Tea Leaves](https://daily.dev/sources/nolanlawson)** · 7 min read · 0 upvotes · 0 comments ## Summary AI coding tools have made finding bugs nearly free, since agents can surface as many correctness, performance, accessibility, or reliability issues as you ask for. Yet the overall polish of software hasn't improved, largely due to organizational incentives (the preventable problem paradox, pressure to ship visible features) and reviewer fatigue from sifting through endless AI-generated bug reports. Suggested approaches to avoid diminishing returns include seeking simplifications rather than piling on fixes (avoiding 'epicycles'), investing in strong test suites so agents can grind toward correctness, and simplifying system design (e.g. favoring multi-page apps over single-page apps) so whole bug classes become impossible. The piece argues that in the near term, developers still add value by steering agents toward simpler, more maintainable code rather than trusting agents to manage arbitrary complexity. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### Why hasn't software quality improved even though AI coding tools can find many more bugs? Organizational incentives still reward visible feature shipping over invisible bug prevention, a dynamic sometimes called the preventable problem paradox: fixing a visible incident makes you a hero, but preventing it in the first place goes unnoticed. Management often equates AI productivity gains with more features rather than more polish, and reviewers get mentally exhausted triaging endless AI-generated bug reports, so many just tune them out. _Anyone weighing AI-assisted code review against real quality gains can track these debates as they unfold on daily.dev._ ### What is the epicycle problem when using AI agents in a review-and-fix loop? Running an agent in a loop that repeatedly finds and patches issues tends to bolt on fix after fix without ever simplifying the underlying design, mirroring how pre-Copernican astronomers added circles upon circles to planetary orbit models rather than recognizing a simpler heliocentric explanation. Agents are described as poor at making the dramatic simplifications needed to avoid this, so code can devolve into a spaghetti mess of one-off patches. _Developers refining their AI code review workflow can follow practical takes like this on daily.dev._ ### How can having a strong test suite help when letting an AI agent fix bugs automatically? A sufficiently thorough test suite lets an agent grind through failures until it reaches near 100% pass rate, giving reasonable confidence in a bug-free implementation, as demonstrated when vibe coding a W3C IndexedDB API against the Web Platform Tests suite. This works because that test suite was honed over years by independent browser implementers covering many edge cases, a bar most first-party codebases don't meet. _Teams building test-driven AI workflows can keep up with real-world experiments like this via daily.dev._ ## Similar posts on daily.dev - [How Software Quality Changes When AI Enters the Build Loop](https://daily.dev/posts/how-software-quality-changes-when-ai-enters-the-build-loop-llzgc7n9q) · Medium · 0 upvotes · 0 comments - [Managing the Hidden Overhead of AI Software Engineering](https://daily.dev/posts/managing-the-hidden-overhead-of-ai-software-engineering-odnh0uxvl) · DEV · 1 upvotes · 0 comments - [Human Maintainable Code in the Age of AI Agents](https://daily.dev/posts/human-maintainable-code-in-the-age-of-ai-agents-ilzajja2s) · Atomic Spin · 12 upvotes · 1 comments --- Tags: [#testing](https://daily.dev/tags/testing), [#architecture](https://daily.dev/tags/architecture), [#ai-coding](https://daily.dev/tags/ai-coding), [#vibe-coding](https://daily.dev/tags/vibe-coding) [View this post on daily.dev](https://daily.dev/posts/you-can-just-choose-how-many-bugs-you-want-now-yybwzht7p) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"You can just choose how many bugs you want now","url":"https://daily.dev/posts/you-can-just-choose-how-many-bugs-you-want-now-yybwzht7p","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/you-can-just-choose-how-many-bugs-you-want-now-yybwzht7p"},"datePublished":"2026-08-16T16:54:01.189Z","dateModified":"2026-09-13T21:09:42.718Z","description":"AI coding tools have made finding bugs nearly free, since agents can surface as many correctness, performance, accessibility, or reliability issues as you ask...","image":"https://media.daily.dev/image/upload/s--VDukGCjf--/f_auto/v1722860399/public/Placeholder%2002","thumbnailUrl":"https://media.daily.dev/image/upload/s--VDukGCjf--/f_auto/v1722860399/public/Placeholder%2002","isAccessibleForFree":true,"articleSection":"Read the Tea Leaves","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"Read the Tea Leaves","logo":"https://media.daily.dev/image/upload/s--hCPaPAOM--/f_auto/v1727446484/logos/nolanlawson","url":"https://daily.dev/sources/nolanlawson"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/you-can-just-choose-how-many-bugs-you-want-now-yybwzht7p","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"testing,architecture,ai-coding,vibe-coding","timeRequired":"PT7M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Read the Tea Leaves","item":"https://daily.dev/sources/nolanlawson"},{"@type":"ListItem","position":3,"name":"You can just choose how many bugs you want now"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/you-can-just-choose-how-many-bugs-you-want-now-yybwzht7p#faq","mainEntity":[{"@type":"Question","name":"Why hasn't software quality improved even though AI coding tools can find many more bugs?","acceptedAnswer":{"@type":"Answer","text":"Organizational incentives still reward visible feature shipping over invisible bug prevention, a dynamic sometimes called the preventable problem paradox: fixing a visible incident makes you a hero, but preventing it in the first place goes unnoticed. Management often equates AI productivity gains with more features rather than more polish, and reviewers get mentally exhausted triaging endless AI-generated bug reports, so many just tune them out. Anyone weighing AI-assisted code review against real quality gains can track these debates as they unfold on daily.dev."}},{"@type":"Question","name":"What is the epicycle problem when using AI agents in a review-and-fix loop?","acceptedAnswer":{"@type":"Answer","text":"Running an agent in a loop that repeatedly finds and patches issues tends to bolt on fix after fix without ever simplifying the underlying design, mirroring how pre-Copernican astronomers added circles upon circles to planetary orbit models rather than recognizing a simpler heliocentric explanation. Agents are described as poor at making the dramatic simplifications needed to avoid this, so code can devolve into a spaghetti mess of one-off patches. Developers refining their AI code review workflow can follow practical takes like this on daily.dev."}},{"@type":"Question","name":"How can having a strong test suite help when letting an AI agent fix bugs automatically?","acceptedAnswer":{"@type":"Answer","text":"A sufficiently thorough test suite lets an agent grind through failures until it reaches near 100% pass rate, giving reasonable confidence in a bug-free implementation, as demonstrated when vibe coding a W3C IndexedDB API against the Web Platform Tests suite. This works because that test suite was honed over years by independent browser implementers covering many edge cases, a bar most first-party codebases don't meet. Teams building test-driven AI workflows can keep up with real-world experiments like this via daily.dev."}}]} ```