--- title: Waterfall 2.0: Controlling LLM-Driven Software... description: An engineer proposes 'Waterfall 2.0', a revival of classic stage-gated software development adapted for LLM-driven engineering. The five stages—vision, ADRs,... canonical: https://daily.dev/posts/waterfall-2-0-controlling-llm-driven-software-development-with-stage-gated-discipline-twzkyevnp twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Waterfall 2.0: Controlling LLM-Driven Software Development with Stage-Gated Discipline | daily.dev og:description: An engineer proposes 'Waterfall 2.0', a revival of classic stage-gated software development adapted for LLM-driven engineering. The five stages—vision, ADRs,... og:url: https://daily.dev/posts/waterfall-2-0-controlling-llm-driven-software-development-with-stage-gated-discipline-twzkyevnp og:image: https://api.daily.dev/og/posts/TWZkYEVnp.png og:image:alt: Waterfall 2.0: Controlling LLM-Driven Software Development with Stage-Gated Discipline 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. # Waterfall 2.0: Controlling LLM-Driven Software Development with Stage-Gated Discipline **[DevOps.com](https://daily.dev/sources/devops)** · 10 min read · 6 upvotes · 0 comments ## Summary An engineer proposes 'Waterfall 2.0', a revival of classic stage-gated software development adapted for LLM-driven engineering. The five stages—vision, ADRs, prototypes, architectural skeleton, and code plus tests—are compressed from months into hours or days because LLMs can regenerate artifacts quickly. Backtracking, once painful in traditional Waterfall, becomes cheap and encouraged, likened to bubble sort swapping elements until convergence. Central to the approach is a shared knowledge corpus combining requirements, ADRs, code, schemas, diagrams, and telemetry that gives the LLM full context. The engineer's role shifts from writing artifacts to orchestrating the LLM, validating outputs, and making trade-off decisions—a return to what the author calls the 'combi-shaper' role that predates specialization in software engineering. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What are the five stages of the Waterfall 2.0 methodology for LLM-driven development? The five stages are vision, architecture decision records (ADRs), prototypes, architectural skeleton, and code plus tests. Each stage runs through an LLM at high speed: vision takes about 30-60 minutes, ADRs 2-3 hours, prototypes 1-2 days, architectural skeleton about 2 hours, and code plus tests 0.5-2 days. _daily.dev surfaces workflow ideas like this for engineers structuring AI-assisted development pipelines._ ### Why is backtracking to earlier stages considered cheap in an LLM-driven Waterfall process, unlike classic Waterfall? Backtracking is cheap because LLMs can regenerate architecture, prototypes, and code quickly once assumptions change, turning a correction that once took weeks into a process taking a few hours. This flips the economics of classic Waterfall, where late-stage fixes were expensive, into an approach where discovering problems early is rewarded rather than punished. _Engineers weighing structured versus agile workflows for AI-assisted coding can track approaches like this on daily.dev._ ### What testing strategy is recommended for the code and tests stage of an LLM-driven development pipeline? A diamond testing model is recommended: minimal unit tests only where they add unique value, heavy emphasis on integration tests for broad coverage, and end-to-end tests reserved for critical paths only. Since generated source code becomes relatively transient and easy to regenerate, automated tests serve as the primary guarantee of system stability. _daily.dev helps developers compare testing strategies suited to fast, AI-generated codebases._ ## Similar posts on daily.dev - [Software Development in 2026](https://daily.dev/posts/software-development-in-2026-q2wl4qawy) · Beam Bloggers Webring · 4 upvotes · 0 comments - [Software 2.0: Planning and Verifying a Greenfield Project](https://daily.dev/posts/software-2-0-planning-and-verifying-a-greenfield-project-dx9yva0ku) · Aaronontheweb · 1 upvotes · 0 comments - [The next hit: How LLMs change the way engineers work](https://daily.dev/posts/the-next-hit-how-llms-change-the-way-engineers-work-y3ba4zyqx) · RUBYLAND · 1 upvotes · 0 comments - [Spec-Driven Development: The Waterfall Strikes Back](https://daily.dev/posts/spec-driven-development-the-waterfall-strikes-back-xzhbdo7yv) · marmelab · 91 upvotes · 10 comments --- Tags: [#architecture](https://daily.dev/tags/architecture), [#mcp](https://daily.dev/tags/mcp) [View this post on daily.dev](https://daily.dev/posts/waterfall-2-0-controlling-llm-driven-software-development-with-stage-gated-discipline-twzkyevnp) ```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":"Waterfall 2.0: Controlling LLM-Driven Software Development with Stage-Gated Discipline","url":"https://daily.dev/posts/waterfall-2-0-controlling-llm-driven-software-development-with-stage-gated-discipline-twzkyevnp","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/waterfall-2-0-controlling-llm-driven-software-development-with-stage-gated-discipline-twzkyevnp"},"datePublished":"2026-08-21T11:40:52.644Z","dateModified":"2026-09-14T07:31:18.478Z","description":"An engineer proposes 'Waterfall 2.0', a revival of classic stage-gated software development adapted for LLM-driven engineering. The five stages—vision, ADRs,...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/64c71325b4b93d8e8b7d12524e2178a7?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/64c71325b4b93d8e8b7d12524e2178a7?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"DevOps.com","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":"DevOps.com","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/db8f2265cff0416c878c6e7e92bb8715","url":"https://daily.dev/sources/devops"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/waterfall-2-0-controlling-llm-driven-software-development-with-stage-gated-discipline-twzkyevnp","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":6},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"architecture,mcp","timeRequired":"PT10M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"DevOps.com","item":"https://daily.dev/sources/devops"},{"@type":"ListItem","position":3,"name":"Waterfall 2.0: Controlling LLM-Driven Software Development with Stage-Gated Discipline"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/waterfall-2-0-controlling-llm-driven-software-development-with-stage-gated-discipline-twzkyevnp#faq","mainEntity":[{"@type":"Question","name":"What are the five stages of the Waterfall 2.0 methodology for LLM-driven development?","acceptedAnswer":{"@type":"Answer","text":"The five stages are vision, architecture decision records (ADRs), prototypes, architectural skeleton, and code plus tests. Each stage runs through an LLM at high speed: vision takes about 30-60 minutes, ADRs 2-3 hours, prototypes 1-2 days, architectural skeleton about 2 hours, and code plus tests 0.5-2 days. daily.dev surfaces workflow ideas like this for engineers structuring AI-assisted development pipelines."}},{"@type":"Question","name":"Why is backtracking to earlier stages considered cheap in an LLM-driven Waterfall process, unlike classic Waterfall?","acceptedAnswer":{"@type":"Answer","text":"Backtracking is cheap because LLMs can regenerate architecture, prototypes, and code quickly once assumptions change, turning a correction that once took weeks into a process taking a few hours. This flips the economics of classic Waterfall, where late-stage fixes were expensive, into an approach where discovering problems early is rewarded rather than punished. Engineers weighing structured versus agile workflows for AI-assisted coding can track approaches like this on daily.dev."}},{"@type":"Question","name":"What testing strategy is recommended for the code and tests stage of an LLM-driven development pipeline?","acceptedAnswer":{"@type":"Answer","text":"A diamond testing model is recommended: minimal unit tests only where they add unique value, heavy emphasis on integration tests for broad coverage, and end-to-end tests reserved for critical paths only. Since generated source code becomes relatively transient and easy to regenerate, automated tests serve as the primary guarantee of system stability. daily.dev helps developers compare testing strategies suited to fast, AI-generated codebases."}}]} ```