--- title: AI migrations get real numbers, and QA teams debate... description: The Asana migration story resurfaces with two more case studies, cementing AI-assisted migration as a proven pattern rather than a one-off. ClickHouse's... canonical: https://daily.dev/posts/ai-migrations-get-real-numbers-and-qa-teams-debate-what-s-left-to-test-gserjpqgs twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: AI migrations get real numbers, and QA teams debate what's left to test | daily.dev og:description: The Asana migration story resurfaces with two more case studies, cementing AI-assisted migration as a proven pattern rather than a one-off. ClickHouse's... og:url: https://daily.dev/posts/ai-migrations-get-real-numbers-and-qa-teams-debate-what-s-left-to-test-gserjpqgs og:image: https://api.daily.dev/og/posts/GseRjpqgS.png og:image:alt: AI migrations get real numbers, and QA teams debate what's left to test 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. # AI migrations get real numbers, and QA teams debate what's left to test **[Testing Digest](https://daily.dev/sources/testing_digest)** · 4 min read · 0 upvotes · 0 comments ## Summary The Asana migration story resurfaces with two more case studies, cementing AI-assisted migration as a proven pattern rather than a one-off. ClickHouse's autonomous QA agent posted five months of hard numbers on filing and fixing real bugs unsupervised. Meanwhile several posts converge on the same worry: green tests and passing checks increasingly mean nothing when nobody verifies the verifier. And QA-is-dead predictions keep getting undercut by teams reporting testing demand going up, not down, as AI writes more of the code. ## Content **TLDR:** The Asana migration story resurfaces with two more case studies, cementing AI-assisted migration as a proven pattern rather than a one-off. ClickHouse's autonomous QA agent posted five months of hard numbers on filing and fixing real bugs unsupervised. Meanwhile several posts converge on the same worry: green tests and passing checks increasingly mean nothing when nobody verifies the verifier. And QA-is-dead predictions keep getting undercut by teams reporting testing demand going up, not down, as AI writes more of the code. --- ## AI-assisted migrations rack up more real numbers: Uber's 600K tests, Bun's 530K lines Beyond Asana's now-famous $12K Enzyme migration, this week's Pulse follow-up adds Uber (600,000 JUnit tests migrated across 15 million lines of code in four months with two engineers) and Bun (530,000 lines of Zig moved to Rust in two weeks) to the pile of migrations that were previously considered too expensive to attempt. Airbnb's 3,500-file Enzyme migration in six weeks, versus an estimated 1.5 engineer-years by hand, rounds out the case studies. The pattern is now well-documented enough that the interesting question isn't whether AI can do this, it's which migration you've been avoiding that's now worth reconsidering. [Read more](https://daily.dev/feed-by-ids?id=olifMKjr1&id=q4UReSUB6) ## Nobody tests whether their tests can actually fail An engineer audited three of their own repos and found only 11% of 204 'conclusion-bearing' checks had ever been fed a known-bad input to prove they could fail. Three real incidents illustrate the gap: a deploy gate that failed on the exact missing dependency it existed to catch, a data harvester that discarded good work over exit-code confusion, and a pattern matcher that misread a quota number as a server error. The argument lands hard given how much review work now flows through automated gates: a green check and a broken check that finds nothing look identical from the outside. [Read more](https://daily.dev/posts/bJA5mVGkR) ## ClickHouse's autonomous QA agent: 500 issues filed, 200 PRs, in five months ClickGap reviews every PR merged into ClickHouse's C++ codebase, runs tests against real builds, bisects regressions to the introducing commit, and files fixes without waiting for human pre-approval. Five months in it's filed roughly 500 issues and about 200 PRs, with over half the issues closed, at a running cost of a few dollars per PR reviewed. The writeup is refreshingly honest about failure modes too, including leaked private reasoning and misparsed CI logs, which is more candor than most 'autonomous agent' pitches offer. [Read more](https://daily.dev/posts/RMhYS8bkW) ## QA demand isn't shrinking even where AI writes all the code A QA services co-founder describes a project where AI handles feature building, bug fixing, code review, and deployment end to end, yet manual testing demand held steady or increased. The reasoning: AI builds exactly what's specified, but requirements never capture every edge case, like what happens when a customer changes a delivery address after the order ships. It's a pointed rebuttal to the recurring 'QA is dead' prediction that apparently dates back to 2012 and still hasn't landed. [Read more](https://daily.dev/posts/SwJuI7evK) --- ## Also notable - **Playwright-PHP runs Symfony E2E tests through the kernel, cuts runtime in half:** A real migration from Symfony Panther to Playwright-PHP dropped a JS-dependent suite's runtime from ~39 seconds to ~18 seconds, and Zenstruck Browser has already deprecated Panther in favor of it. [Read more](https://daily.dev/posts/LZ5xDMgdk) - **mirrord ships Chaos Testing for local fault injection:** MetalBear's new mirrord Chaos Testing feature (OSS/CLI 3.241.0+) lets you inject latency or connection failures into your own service's outgoing calls during a local session without touching a shared staging environment. [Read more](https://daily.dev/feed-by-ids?id=emVOlloZ5&id=hEhTl0QjO) - **CircleCI's Smarter Testing claims 4x faster PR feedback:** CircleCI's new Intelligent Test Selection, Dynamic Test Splitting, and Auto Rerun features are paid add-ons billed by stored test results, with case studies citing 4x faster PR feedback loops for a 100-developer team. [Read more](https://daily.dev/posts/cOLZPXW8R) - **Bruno v4 adds typed variables with a hard breaking change:** Any .bru file with a @type annotation will fail to parse on Bruno versions older than v4.0.0, so teams sharing collections need to coordinate the upgrade before adopting typed variables. [Read more](https://daily.dev/posts/EonUsS4PW) - **A study cited this week: AI-generated code introduces known security flaws in ~45% of cases:** The figure comes up in a broader piece on why code verification is now the bottleneck, alongside METR's finding that AI-assisted tasks took 19% longer despite developers feeling faster. [Read more](https://daily.dev/posts/ROTEb1arC) ## Similar posts on daily.dev - [The Pulse: We need to talk about migrations with AI](https://daily.dev/posts/the-pulse-we-need-to-talk-about-migrations-with-ai-q4uresub6) · The Pragmatic Engineer · 1 upvotes · 1 comments --- Tags: [#testing](https://daily.dev/tags/testing), [#cicd](https://daily.dev/tags/cicd), [#ai-coding](https://daily.dev/tags/ai-coding), [#clickhouse](https://daily.dev/tags/clickhouse) [View this post on daily.dev](https://daily.dev/posts/ai-migrations-get-real-numbers-and-qa-teams-debate-what-s-left-to-test-gserjpqgs) ```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":"DiscussionForumPosting","mainEntityOfPage":"https://daily.dev/posts/ai-migrations-get-real-numbers-and-qa-teams-debate-what-s-left-to-test-gserjpqgs","headline":"AI migrations get real numbers, and QA teams debate what's left to test","text":"The Asana migration story resurfaces with two more case studies, cementing AI-assisted migration as a proven pattern rather than a one-off. ClickHouse's autonomous QA agent posted five months of hard numbers on filing and fixing real bugs unsupervised. Meanwhile several posts converge on the same worry: green tests and passing checks increasingly mean nothing when nobody verifies the verifier. And QA-is-dead predictions keep getting undercut by teams reporting testing demand going up, not down, as AI writes more of the code.","url":"https://daily.dev/posts/ai-migrations-get-real-numbers-and-qa-teams-debate-what-s-left-to-test-gserjpqgs","datePublished":"2026-08-31T04:18:32.883Z","dateModified":"2026-08-31T04:18:55.236Z","author":{"@type":"Organization","name":"Testing Digest","logo":"https://media.daily.dev/image/upload/s--1rQKIOqD--/f_auto,q_auto/v1773839410/logos/testing_digest?_a=BAMAMiiu0","url":"https://daily.dev/sources/testing_digest"},"interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"isPartOf":{"@type":"WebPage","url":"https://daily.dev/sources/testing_digest","name":"Testing Digest"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Testing Digest","item":"https://daily.dev/sources/testing_digest"},{"@type":"ListItem","position":3,"name":"AI migrations get real numbers, and QA teams debate what's left to test"}]} ```