--- title: Fixing the PR Bottleneck — Matt Pocock, AIHero | daily.dev description: A conference talk from AI Hero's Matt Pocock argues that AI coding agents have made the pull-request review bottleneck worse by flooding repos with more code... canonical: https://daily.dev/posts/fixing-the-pr-bottleneck-matt-pocock-aihero-jyn4aj6sm twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Fixing the PR Bottleneck — Matt Pocock, AIHero | daily.dev og:description: A conference talk from AI Hero's Matt Pocock argues that AI coding agents have made the pull-request review bottleneck worse by flooding repos with more code... og:url: https://daily.dev/posts/fixing-the-pr-bottleneck-matt-pocock-aihero-jyn4aj6sm og:image: https://api.daily.dev/og/posts/jyN4aJ6sm.png og:image:alt: Fixing the PR Bottleneck — Matt Pocock, AIHero 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. # Fixing the PR Bottleneck — Matt Pocock, AIHero **[AI Engineer](https://daily.dev/sources/aidotengineer)** · 22 min read · 0 upvotes · 0 comments ## Summary A conference talk from AI Hero's Matt Pocock argues that AI coding agents have made the pull-request review bottleneck worse by flooding repos with more code ('software factories' need 'brakes' or you get a 'slop cannon'). He proposes a three-layer quality pipeline: cheap automated checks (linting, tests, type checking), automated review by a separate sub-agent loaded with team-specific coding standards, and finally human review reserved for high-risk changes. He shows real examples of AI-generated 'tautological' tests that just reassert implementation details and tests that can never fail due to over-mocking, arguing deep-module design reduces these failure modes. He advises against putting coding standards in the implementation agent's context (it's already overloaded) and instead using a dedicated code-review agent that commits fixes directly rather than leaving comments. He also introduces a new 'PR skill' for generating human-friendly PR summaries (using one-way/two-way door risk framing and pseudocode diagrams) and a 'retro' skill that mines past sessions to continuously improve coding standards and automated checks. New skill versions are shipping via aihero.dev/skills. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is a tautological test and why do AI coding agents tend to write them? A tautological test is one that merely reasserts the implementation instead of verifying real behavior, such as asserting a character limit constant equals 280 when the code itself defines that constant as 280. AI agents, including Opus, produce these because they tie tests too closely to internal structure rather than testing through a stable interface, making the tests fail on any rename or refactor. _Anyone reviewing AI-generated test suites can track these code quality patterns and fixes on daily.dev._ ### Why should coding standards be kept out of an AI implementation agent's context and put in a separate review agent instead? Implementation agents are already overloaded with exploring the codebase, writing changes, and debugging, so adding coding standards on top degrades their performance. A separate code-review sub-agent, given only a diff and a coding-standards file, is comparatively underloaded and can enforce style and quality rules far more effectively, following a red-green-refactor style split across two context windows. _Teams tuning multi-agent coding workflows can follow discussions like this on daily.dev._ ### How can a PR description help reviewers decide how much scrutiny a pull request needs? Classify each PR as a one-way door (hard to reverse, e.g. expensive migrations, data loss, or an email blast to 60,000 people) or a two-way door (easily revertible), and summarize the blast radius and merge danger at the top of the PR. Combined with pseudocode or diagram summaries of the change, this lets reviewers skip deep review on low-risk two-way-door changes. _Developers refining PR review practices for AI-heavy codebases can find more approaches like this on daily.dev._ ## Similar posts on daily.dev - [How to Unblock Your AI PR Review Bottleneck: A Tech Lead’s Guide to Building a Codebase-Aware Reviewer](https://daily.dev/posts/how-to-unblock-your-ai-pr-review-bottleneck-a-tech-lead-s-guide-to-building-a-codebase-aware-review-hurtxskqp) · freeCodeCamp · 1 upvotes · 0 comments - [The Human Bottleneck](https://daily.dev/posts/the-human-bottleneck-hxoxu6uht) · Scott Logic · 1 upvotes · 0 comments - [AI Code Review Bottleneck 2026: Why PRs Are Stuck in Review](https://daily.dev/posts/ai-code-review-bottleneck-2026-why-prs-are-stuck-in-review-o7qcycgq0) · Alex CloudStar · 2 upvotes · 1 comments - [Agentic Code Review](https://daily.dev/posts/agentic-code-review-0ue8fcsa1) · Addy Osmani · 147 upvotes · 14 comments --- Tags: [#testing](https://daily.dev/tags/testing), [#ai-agents](https://daily.dev/tags/ai-agents), [#code-review](https://daily.dev/tags/code-review), [#vibe-coding](https://daily.dev/tags/vibe-coding) [View this post on daily.dev](https://daily.dev/posts/fixing-the-pr-bottleneck-matt-pocock-aihero-jyn4aj6sm) ```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":"Fixing the PR Bottleneck — Matt Pocock, AIHero","url":"https://daily.dev/posts/fixing-the-pr-bottleneck-matt-pocock-aihero-jyn4aj6sm","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/fixing-the-pr-bottleneck-matt-pocock-aihero-jyn4aj6sm"},"datePublished":"2026-09-26T00:09:30.784Z","dateModified":"2026-09-26T20:09:48.978Z","description":"A conference talk from AI Hero's Matt Pocock argues that AI coding agents have made the pull-request review bottleneck worse by flooding repos with more code...","image":"https://i.ytimg.com/vi/LlgiOCmFG_w/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/LlgiOCmFG_w/sddefault.jpg","isAccessibleForFree":true,"articleSection":"AI Engineer","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":"AI Engineer","logo":"https://media.daily.dev/image/upload/s--u5PucxNT--/f_auto/v1724338940/logos/aidotengineer","url":"https://daily.dev/sources/aidotengineer"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/fixing-the-pr-bottleneck-matt-pocock-aihero-jyn4aj6sm","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"testing,ai-agents,code-review,vibe-coding","timeRequired":"PT22M","video":{"@type":"VideoObject","name":"Fixing the PR Bottleneck — Matt Pocock, AIHero","description":"A conference talk from AI Hero's Matt Pocock argues that AI coding agents have made the pull-request review bottleneck worse by flooding repos with more code...","thumbnailUrl":"https://i.ytimg.com/vi/LlgiOCmFG_w/sddefault.jpg","uploadDate":"2026-09-26T00:09:30.784Z","duration":"PT22M","url":"https://api.daily.dev/r/jyN4aJ6sm","embedUrl":"https://www.youtube.com/embed/LlgiOCmFG_w"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"AI Engineer","item":"https://daily.dev/sources/aidotengineer"},{"@type":"ListItem","position":3,"name":"Fixing the PR Bottleneck — Matt Pocock, AIHero"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/fixing-the-pr-bottleneck-matt-pocock-aihero-jyn4aj6sm#faq","mainEntity":[{"@type":"Question","name":"What is a tautological test and why do AI coding agents tend to write them?","acceptedAnswer":{"@type":"Answer","text":"A tautological test is one that merely reasserts the implementation instead of verifying real behavior, such as asserting a character limit constant equals 280 when the code itself defines that constant as 280. AI agents, including Opus, produce these because they tie tests too closely to internal structure rather than testing through a stable interface, making the tests fail on any rename or refactor. Anyone reviewing AI-generated test suites can track these code quality patterns and fixes on daily.dev."}},{"@type":"Question","name":"Why should coding standards be kept out of an AI implementation agent's context and put in a separate review agent instead?","acceptedAnswer":{"@type":"Answer","text":"Implementation agents are already overloaded with exploring the codebase, writing changes, and debugging, so adding coding standards on top degrades their performance. A separate code-review sub-agent, given only a diff and a coding-standards file, is comparatively underloaded and can enforce style and quality rules far more effectively, following a red-green-refactor style split across two context windows. Teams tuning multi-agent coding workflows can follow discussions like this on daily.dev."}},{"@type":"Question","name":"How can a PR description help reviewers decide how much scrutiny a pull request needs?","acceptedAnswer":{"@type":"Answer","text":"Classify each PR as a one-way door (hard to reverse, e.g. expensive migrations, data loss, or an email blast to 60,000 people) or a two-way door (easily revertible), and summarize the blast radius and merge danger at the top of the PR. Combined with pseudocode or diagram summaries of the change, this lets reviewers skip deep review on low-risk two-way-door changes. Developers refining PR review practices for AI-heavy codebases can find more approaches like this on daily.dev."}}]} ```