--- title: The Brain Sandwich | daily.dev description: An engineering manager shares a colleague's simple framework for working with AI coding agents called the 'Brain Sandwich': use your brain to understand the... canonical: https://daily.dev/posts/the-brain-sandwich-1xznwdvve twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: The Brain Sandwich | daily.dev og:description: An engineering manager shares a colleague's simple framework for working with AI coding agents called the 'Brain Sandwich': use your brain to understand the... og:url: https://daily.dev/posts/the-brain-sandwich-1xznwdvve og:image: https://api.daily.dev/og/posts/1xzNwDvVe.png og:image:alt: The Brain Sandwich 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. # The Brain Sandwich **[Terrible Software](https://daily.dev/sources/terriblesoftware)** · 2 min read · 36 upvotes · 8 comments ## Summary An engineering manager shares a colleague's simple framework for working with AI coding agents called the 'Brain Sandwich': use your brain to understand the problem first, delegate the implementation to AI second, then use your brain again to verify, simplify, and explain the result before sending it for review. The goal is taking full advantage of AI assistance without letting engineering judgment atrophy or shipping unreviewable, misunderstood code. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is the brain sandwich framework for using AI coding agents? It is a three-step workflow: use your own brain to understand the problem first, then delegate the implementation to an AI agent, then use your brain again to review, simplify, and explain the resulting diff before sending it for review. The rule of thumb is that if you can't explain what changed and why, it isn't ready for someone else to review. _daily.dev surfaces practical workflows like this for developers figuring out how to pair human judgment with AI agents._ ## Community take How the wider developer community reacted, aggregated from 1 discussion and 44 comments across hackernews (as of 2026-10-19). **TL;DR:** The framework gets a generally warm reception as a sane middle ground for using AI coding agents, but a long sub-thread argues over whether true understanding can really be delegated and whether the industry is sliding toward code nobody comprehends. **Sentiment:** 30% positive · 35% mixed · 35% skeptical **The case for** - The brain-first/brain-last structure gives people a sense of agency and personal responsibility for AI-generated code. - Several say this mirrors how they already work: plan and scrutinize before letting the agent implement, then review carefully. - Some argue AI can be reviewed and questioned effectively, similar to pair-programming with a very fast, tireless teammate. **The pushback** - Skeptics argue you can't really understand work you didn't do yourself, likening full delegation to hiring an electrician who never learns from running the wires. - Some say real-world incentives (delivery pressure, management apathy) make the discipline impractical despite being a good idea in theory. - Others worry AI-driven delegation is deskilling engineers and leading toward codebases no one truly understands, like legacy COBOL systems no one can safely touch. - One commenter contends the actual utility of AI for many is shipping without understanding at all, making calls for discipline unrealistic. **By community** - hackernews (heated): A long back-and-forth debate dominates between those who think the sandwich approach is sound practice and those who argue delegating understanding to AI is fundamentally a losing proposition, alongside lighter appreciative comments and jokes about the naming. **Hottest debate:** Whether you can genuinely understand and be responsible for code/work that an AI did for you, versus this being a convenient fiction that erodes engineering competence over time. **Open questions** - Can reviewing and questioning an AI's output substitute for the learning that comes from doing the work yourself? - Will this discipline actually survive delivery pressure and management incentives in most organizations? - As AI capabilities improve, will the concerns about brownfield complexity and understanding become moot or worse? **Highlights** > The whole actual, not just the "openly admitted", utility of AI is to allow people to ship while not understanding the work, and not even caring that much about it. So such posts will make as much change as "trade, but be ethical" would make on a late 80s Wall Street trading floor: "Before you tell the agent to start building anything, you go there and read and understand it first." Yeah, not gonna happen. 1 in 100 might do that. > — [coldtea on hackernews · 3 comments](https://news.ycombinator.com/item?id=50143555) > If you had an electrician that told you "Oh, I don't learn by running wires, but if I didn't immediately see where a wire lead, I'd ask someone who knew, and then I read wiring diagrams after the other person did the work", would you hire them? > — [0c3ca83 on hackernews · 1 comments](https://news.ycombinator.com/item?id=50145062) > I completely agree. While the best outcome may obviously involve lots of thinking, from what I've seen an experienced, the AI tools we have today put people on a slope to becoming "meat proxies". Some of the brighter, stronger engineers may have a less steep slope, but I'm convinced that with time they too will be meat proxies. AI is attacking on too many fronts for it to be otherwise: - AI generates too much code that overwhelms humans - AI deskills users concerning the gritty details of coding by having them not do it themselves - Having most of one's co-workers become meat proxy will put pressure for others to convert - Lack of effort or expertise invested into codebase leads people to care less; caring less -> not caring about optimal outcome -> not expending brain power on problems we don't care about. I think since we as an industry are still on this slope, we're still having systems propped up by people actually using their brains. I personally think this situation is regrettable. We should as an industry consider if the direction we're going truly is something we want. > — [uludag on hackernews](https://news.ycombinator.com/item?id=50144363) > This approach appears valuable for personal growth as an engineer, yet it demands discipline that may not yield returns in corporate settings. Specifically: > The rule of thumb is: if you can’t explain what’s changing and why, it’s not ready for someone else to take a look I have observed firsthand that the metric or "understanding" can be irrelevant to management. When delivery dominates, understanding can be deferred to future agents who will gladly consume your codebase and convince you they grasp it, and further, that you do too. If I invest substantial effort in my PRs only for them to be /review'd, does it still matter? Perhaps reviewers should practice brain sandwich as well. The worst offenders seem to sit at either extreme: fully AI firms that treat understanding as a human weakness and prioritize outcomes alone, or large corporate teams building internal CRUD apps whose middle management merely pretends to value good engineering to satisfy finicky senior engineers. Not a criticism on the authors work at all, I like the idea (and I don't think they were trying to solve anything to do with work culture), but ultimately if my company doesn't find my practice valuable, I can't easily justify taking it on. Especially if I have to sell to them that I'm going to be spending _more_ time delivering. Fun times! > — [jmarlow47 on hackernews · 1 comments](https://news.ycombinator.com/item?id=50141454) > We are clearly getting to the point, or past the point, where humans will understand the code base at all. I fear we are building a tower of "it kind of works this way" (maybe it does or maybe it doesn't), in a couple years people won't know what they can change, they won't get much of the architecture. If one can't understand the monstrosity that was built up over time, where the code base is enormous, millions and millions of lines and components, all with a few buggy corners that the tests didn't touch, and a human didn't have the insight to write a test that understands and verifies the corner cases, it will be like a magic machine that we can't change because we don't understand it and we can't afford years of effort. Some may hope their LLM will be able to understand it. We'll have things that are buggy in ways that become very difficult to understand or fix, and no one can afford the effort. Remember when we had all those huge old programs from the 60s or early 70s, before there were tests maybe, or unit tests. Code written at a lot of human effort, but before the ideas of software engineering made it more tractable to build large programs reliably, testable. We are creating a new problem for the people that take over our new code, just like the "vast for the time" 1960s huge cobol program can't be changed. > — [Alive-in-2025 on hackernews · 2 comments](https://news.ycombinator.com/item?id=50144713) **Source threads** - [hackernews](https://news.ycombinator.com/item?id=50139998) · 60 points · 44 comments ## Community discussion Top comments from developers on daily.dev. **@elmarchavez** · 2 upvotes > Such a nice framework, the brain sandwich is catchy too. Hopefully, more devs would do the same though. Skill atrophy is real if not used. It all boils down to discipline. **@ricktonoli** · 2 upvotes > Awesome framework, sharing with my team next week. I also love the reference to Meat Proxy, which I'm afraid some of them might be trending towards. **@souvikguria** · 1 upvotes > nothing new.. resonsible devs have already been doing this since the very beginning.. **@ristotoldsep** · 1 upvotes > Brain before, AI in the middle, brain after. surprisingly rare recipe. **@dragonifr** · 1 upvotes > better than the current one ingredient ai sandwich ## Similar posts on daily.dev - [When you should delegate to AI \(and when you shouldn't\)](https://daily.dev/posts/when-you-should-delegate-to-ai-and-when-you-shouldn-t--vulgppgsn) · LeadDev · 0 upvotes · 0 comments - [A.I. Should Elevate Your Thinking, Not Replace It](https://daily.dev/posts/a-i-should-elevate-your-thinking-not-replace-it-a7si2mcn1) · Hacker News · 0 upvotes · 0 comments - [Software Engineering Principles Help Get More out of AI](https://daily.dev/posts/software-engineering-principles-help-get-more-out-of-ai-g2d3tqsex) · Telerik · 0 upvotes · 0 comments --- Tags: [#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/the-brain-sandwich-1xznwdvve) ```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":"The Brain Sandwich","url":"https://daily.dev/posts/the-brain-sandwich-1xznwdvve","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/the-brain-sandwich-1xznwdvve"},"datePublished":"2026-10-02T14:00:29.209Z","dateModified":"2026-10-19T00:29:17.424Z","description":"An engineering manager shares a colleague's simple framework for working with AI coding agents called the 'Brain Sandwich': use your brain to understand the...","image":"https://i0.wp.com/terriblesoftware.org/wp-content/uploads/2026/10/cHJpdmF0ZS9sci9pbWFnZXMvd2Vic2l0ZS8yMDIyLTA1L2lzMTk5NzYtaW1hZ2Uta3d2eWVzNXYuanBn.webp?fit=1024%2C683\u0026ssl=1","thumbnailUrl":"https://i0.wp.com/terriblesoftware.org/wp-content/uploads/2026/10/cHJpdmF0ZS9sci9pbWFnZXMvd2Vic2l0ZS8yMDIyLTA1L2lzMTk5NzYtaW1hZ2Uta3d2eWVzNXYuanBn.webp?fit=1024%2C683\u0026ssl=1","isAccessibleForFree":true,"articleSection":"Terrible Software","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":"Terrible Software","logo":"https://media.daily.dev/image/upload/s--tcoB-OS2--/f_auto,q_auto/v1768208174/logos/terriblesoftware","url":"https://daily.dev/sources/terriblesoftware"},"commentCount":8,"discussionUrl":"https://daily.dev/posts/the-brain-sandwich-1xznwdvve","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":36},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":8}],"keywords":"code-review,vibe-coding","timeRequired":"PT2M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Terrible Software","item":"https://daily.dev/sources/terriblesoftware"},{"@type":"ListItem","position":3,"name":"The Brain Sandwich"}]} {"@context":"https://schema.org","@type":"WebPage","@id":"https://daily.dev/posts/the-brain-sandwich-1xznwdvve","comment":[{"@type":"Comment","text":"Such a nice framework, the brain sandwich is catchy too. Hopefully, more devs would do the same though. Skill atrophy is real if not used. It all boils down to discipline.","datePublished":"2026-10-02T15:03:40.584Z","url":"https://daily.dev/posts/1xzNwDvVe#c-3iUimeecj","author":{"@type":"Person","name":"Elmar Chavez","url":"https://daily.dev/elmarchavez","image":"https://media.daily.dev/image/upload/s--7PjH979T--/f_auto/v1780994221/avatars/avatar_bK85rJb2OXXID2LX2wi5y?_a=BAMAMiWQ0"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2}},{"@type":"Comment","text":"Awesome framework, sharing with my team next week. I also love the reference to Meat Proxy, which I’m afraid some of them might be trending towards.","datePublished":"2026-10-02T22:57:41.884Z","url":"https://daily.dev/posts/1xzNwDvVe#c-TMks0hTNE","author":{"@type":"Person","name":"Rick Tonoli","url":"https://daily.dev/ricktonoli","image":"https://media.daily.dev/image/upload/s--1mKwVOZD--/f_auto/v1745813607/avatars/avatar_54gVqSdbL9DajhGSKNFpY"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":2}},{"@type":"Comment","text":"nothing new… resonsible devs have already been doing this since the very beginning…","datePublished":"2026-10-03T08:28:04.557Z","url":"https://daily.dev/posts/1xzNwDvVe#c-uxEUz5zCp","author":{"@type":"Person","name":"Souvik Guria","url":"https://daily.dev/souvikguria","image":"https://lh3.googleusercontent.com/a/ACg8ocLFwm_Fsy4SIXre8Chc0f_lN-_GltKwko7TFX-8K9RjCR2KH5BZCA=s96-c"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}},{"@type":"Comment","text":"Brain before, AI in the middle, brain after. surprisingly rare recipe.","datePublished":"2026-10-05T12:28:37.531Z","url":"https://daily.dev/posts/1xzNwDvVe#c-VLFG2v6ny","author":{"@type":"Person","name":"Risto Tõldsep","url":"https://daily.dev/ristotoldsep","image":"https://lh3.googleusercontent.com/a/ACg8ocLDWc6mZn0JwNmXXw6WY0L_HJ6pegzRttooC5VgtXESHSj3MNYx=s96-c"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}},{"@type":"Comment","text":"better than the current one ingredient ai sandwich","datePublished":"2026-10-03T14:15:26.035Z","url":"https://daily.dev/posts/1xzNwDvVe#c-td84jMkYd","author":{"@type":"Person","name":"dragonFR","url":"https://daily.dev/dragonifr","image":"https://lh3.googleusercontent.com/a/ACg8ocK6-w9SpUgH-F8Suee7ODLs7PDiq87HjNVD5x070sGWwA=s96-c"},"interactionStatistic":{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1}}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/the-brain-sandwich-1xznwdvve#faq","mainEntity":[{"@type":"Question","name":"What is the brain sandwich framework for using AI coding agents?","acceptedAnswer":{"@type":"Answer","text":"It is a three-step workflow: use your own brain to understand the problem first, then delegate the implementation to an AI agent, then use your brain again to review, simplify, and explain the resulting diff before sending it for review. The rule of thumb is that if you can't explain what changed and why, it isn't ready for someone else to review. daily.dev surfaces practical workflows like this for developers figuring out how to pair human judgment with AI agents."}}]} ```