--- title: Show HN: A proof-of-concept formalization of the IRS... description: This is an open-source, vibe-coded project that encodes US federal income tax forms and instructions as cited, machine-checkable YAML data, paired with a... canonical: https://daily.dev/posts/show-hn-a-proof-of-concept-formalization-of-the-irs-federal-tax-forms-rjvgr0j2l twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Show HN: A proof-of-concept formalization of the IRS federal tax forms | daily.dev og:description: This is an open-source, vibe-coded project that encodes US federal income tax forms and instructions as cited, machine-checkable YAML data, paired with a... og:url: https://daily.dev/posts/show-hn-a-proof-of-concept-formalization-of-the-irs-federal-tax-forms-rjvgr0j2l og:image: https://api.daily.dev/og/posts/RJVGr0J2l.png og:image:alt: Show HN: A proof-of-concept formalization of the IRS federal tax forms 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. # Show HN: A proof-of-concept formalization of the IRS federal tax forms **[Hacker News](https://daily.dev/sources/hn)** · 12 min read · 0 upvotes · 0 comments ## Summary This is an open-source, vibe-coded project that encodes US federal income tax forms and instructions as cited, machine-checkable YAML data, paired with a guided self-prep interview. Every computed line must cite an archived IRS source, and all logic is built from a closed set of about twenty 'atoms' with no loops or user-defined functions. A TypeScript build checker and a Lean formalization verify properties of the data and evaluators, and the two evaluators agree on 2,895,826 line values across 1,648 cases with zero differences. Three outside tax engines (PolicyEngine-US, Tax-Calculator, OpenTaxSolver) cross-check the summary lines they compute on 1,700 generated returns, and mutation testing kills 94.4% of 2,278 injected bugs. Measured filer coverage currently stands at 68.6% of individual returns, counting only the situations measured so far. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How do you verify that AI-generated code encoding IRS tax forms is actually correct? Verification layers include per-line citations to archived IRS pages, a closed set of about twenty arithmetic 'atoms' with no loops or user-defined functions, a TypeScript build-time checker, a Lean 4 formal proof of atom properties, mutation testing (94.4% of 2,278 mutants caught), and cross-checks against three outside tax calculation engines over 1,700 generated returns. _Developers building trust into AI-written code can follow verification approaches like this one on daily.dev._ ### What bug was found in the us-tax-model project's Form 1040 calculation despite having two independent evaluators agree? Form 1040 line 9 incorrectly summed line 3a along with line 3b, and both the TypeScript and Lean evaluators agreed with the wrong sum because they read the same underlying data. The bug was caught only by an outside engine comparison, not by the internal fixtures or evaluator agreement, and is recorded as BUG-0001 in the project's append-only bug inventory. _Anyone weighing how much to trust internal test agreement can track cases like this on daily.dev._ ### What percentage of US individual tax returns can the us-tax-model project currently handle without any refusals? The measured filer coverage is 68.6% of individual returns, calculated by tools/coverage/estimate.py, where a return only counts if nothing on it triggers a refusal gate. However, 148 of 157 refusal gates lack IRS count or microdata measurement, so true coverage is acknowledged to be lower than the measured figure suggests. _Developers evaluating early-stage open-source coverage claims can follow projects like this on daily.dev._ ## Community take How the wider developer community reacted, aggregated from 1 discussion and 12 comments across hackernews (as of 2026-10-30). **TL;DR:** Commenters are intrigued by the idea of formalizing tax logic but split on whether it's maintainable long-term, with AI-assisted updates seen as a possible solution by some and wishful thinking by others. **Sentiment:** 30% positive · 40% mixed · 30% skeptical **The case for** - Some think AI agents could make yearly updates to tax law changes tractable, similar to how coding agents keep up with other fast-moving specs. - One commenter described a personal experiment where an AI did a surprisingly good job interpreting IRS rules and state tax code for a real financial situation. - Narrowing scope to the most common filer situations (rather than all cases) was seen as a more achievable goal. **The pushback** - Experienced engineers warn that tax rules change significantly every year and are notoriously hard to keep up with, even with dedicated teams. - Some doubt that even the IRS's own guidance is fully consistent or correct, raising questions about what a formal model can actually verify. - Skepticism that discovering inconsistencies in IRS guidance would have any practical value, since interpretation disputes are already the domain of tax lawyers. **By community** - hackernews (mixed): Discussion splits between people who think maintaining tax law formalization yearly is nearly impossible and those who think AI agents make it newly feasible. **Hottest debate:** Whether keeping a formalized tax model up to date year after year is realistically achievable or a naive dream, especially with AI assistance. **Open questions** - Would this method actually surface inconsistencies in official IRS guidance if they existed? - Can an AI-driven update process realistically keep pace with annual tax law changes without the resources of a large company? **Highlights** > Have fun keeping this up to date. It's an admirable idea, but the tax forms and rules change significantly every year. I built a payroll system for a suite of small and medium sized businesses back in the day. Even during a stable administration the rules are impossible to keep up with. We had a team of engineers who every year had to spend weeks to months trying to keep it up to date and verify the numbers. It's why my accountant makes $400/hr and why there are no open source projects that have ever taken hold in the market. > Most tax software derives trust from the party who maintains it (e.g. Intuit, the IRS, etc.) Because they have the resources to keep it up to date. And even they get it wrong sometimes. Including the IRS itself: https://www.irs.gov/e-file-providers/known-issues-and-soluti... > — [GrumpyGoblin on hackernews · 3 comments](https://news.ycombinator.com/item?id=50295460) > I did a similar project just for myself recently, and I think AI interpretation of the IRS is probably one of the ideal projects. AI is generally pretty good at following a spec, and the IRS is generally pretty good at writing detailed specs. In my case, I had an Inherited IRA that, and this is the first year I have to take a distribution from it. It needs to be emptied in 10 years and starting this year I have a minimum amount to take out based on my age. My goal was to write a simulation of how much I take out each year, what that does to my income taxes, how much tax deferred growth I get by leaving it, etc. I have an optimizer to connect to the simulation, so I could then figure out how much I should take each year. I decided to vibe it, just for fun. I started very simply: write me a simulation for each year with this much projected growth, etc. The AI asked me some questions like "do you have kids going to college during this time frame?" and did you know that your state has an allowance for this much tax free withdrawal, etc. That took me down a rabbit hole since I didn't know all of those things. I said, go read the IRS code and my state tax code, figure out all of the things that are relevant for my finances (normal W-2 employee, no fancy stuff), and it wrote me a much more detailed simulation than I would have thought of and tied it all to all of the tax things that might be relevant for me. I was shocked at how good of a job it did, I can (and did) independently verify it since my taxes aren't that complicated. It did find savings that I didn't know were available to me. I think that this is a great area for AI-based development since it's essentially transpiling detailed IRS specs into code, and that's what AI is really good at. > — [michaelrpeskin on hackernews](https://news.ycombinator.com/item?id=50296800) > Have fun keeping this up to date. Why? These rules are written down somewhere. Looks like the spec was created by pointing the model at irs.gov and/or the Federal Register itself and telling it something like, "Create a specification for a verifiable tax model and the tools to verify it." Next year, all the maintainer has to do is issue the same prompt, no? > — [CamperBob2 on hackernews](https://news.ycombinator.com/item?id=50296087) > Would this method have turned up inconsistencies in the IRS's own guidance if such inconsistencies existed? > — [woah on hackernews · 1 comments](https://news.ycombinator.com/item?id=50297684) > Yes and then? Sue them in federal court? Refuse to pay the taxes? Write them a letter pointing out inconsistencies? Inconsistencies already are known and well studied - this is one of the domains of corporate tax lawyers - how to interpret the inconsistencies into a strategy to benefit the company, so that even if the company is sued by the IRS, a well formed argument would be made and case dismissed or fines significantly reduced. > — [lnkdinsuxs on hackernews](https://news.ycombinator.com/item?id=50298337) **Source threads** - [hackernews](https://news.ycombinator.com/item?id=50278678) · 55 points · 12 comments ## Similar posts on daily.dev - [Calculate Sales Tax And File Returns Automatically With Avalara](https://daily.dev/posts/calculate-sales-tax-and-file-returns-automatically-with-avalara-vbqwubiy4) · API Evangelist · 0 upvotes · 0 comments - [Human vs. AI – Diff-based line-level provenance for text under agentic editing](https://daily.dev/posts/human-vs-ai-diff-based-line-level-provenance-for-text-under-agentic-editing-rn4p3ptsj) · Hacker News · 0 upvotes · 0 comments - [My bookkeeping runs on code now. AI only reads the PDFs.](https://daily.dev/posts/my-bookkeeping-runs-on-code-now-ai-only-reads-the-pdfs--bmpvcdk9m) · Planet Python · 0 upvotes · 0 comments - [California will tax downloaded software for the first time as part of a $351.7 billion budget deal](https://daily.dev/posts/california-will-tax-downloaded-software-for-the-first-time-as-part-of-a-351-7-billion-budget-deal-xjchvmmut) · The Next Web · 2 upvotes · 1 comments - [What We Learned from 3 Million Downloads of Kilo Code](https://daily.dev/posts/what-we-learned-from-3-million-downloads-of-kilo-code-mxphp8car) · Kilo Blog · 2 upvotes · 0 comments --- Tags: [#html](https://daily.dev/tags/html), [#vibe-coding](https://daily.dev/tags/vibe-coding), [#mutation-testing](https://daily.dev/tags/mutation-testing) [View this post on daily.dev](https://daily.dev/posts/show-hn-a-proof-of-concept-formalization-of-the-irs-federal-tax-forms-rjvgr0j2l) ```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":"Show HN: A proof-of-concept formalization of the IRS federal tax forms","url":"https://daily.dev/posts/show-hn-a-proof-of-concept-formalization-of-the-irs-federal-tax-forms-rjvgr0j2l","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/show-hn-a-proof-of-concept-formalization-of-the-irs-federal-tax-forms-rjvgr0j2l"},"datePublished":"2026-10-29T23:03:21.554Z","dateModified":"2026-10-30T02:42:44.782Z","description":"This is an open-source, vibe-coded project that encodes US federal income tax forms and instructions as cited, machine-checkable YAML data, paired with a...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/b0e9ef6395f49862465473832c6478ed?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/b0e9ef6395f49862465473832c6478ed?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"Hacker News","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":"Hacker News","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/hn","url":"https://daily.dev/sources/hn"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/show-hn-a-proof-of-concept-formalization-of-the-irs-federal-tax-forms-rjvgr0j2l","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"html,vibe-coding,mutation-testing","timeRequired":"PT12M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"Hacker News","item":"https://daily.dev/sources/hn"},{"@type":"ListItem","position":3,"name":"Show HN: A proof-of-concept formalization of the IRS federal tax forms"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/show-hn-a-proof-of-concept-formalization-of-the-irs-federal-tax-forms-rjvgr0j2l#faq","mainEntity":[{"@type":"Question","name":"How do you verify that AI-generated code encoding IRS tax forms is actually correct?","acceptedAnswer":{"@type":"Answer","text":"Verification layers include per-line citations to archived IRS pages, a closed set of about twenty arithmetic 'atoms' with no loops or user-defined functions, a TypeScript build-time checker, a Lean 4 formal proof of atom properties, mutation testing (94.4% of 2,278 mutants caught), and cross-checks against three outside tax calculation engines over 1,700 generated returns. Developers building trust into AI-written code can follow verification approaches like this one on daily.dev."}},{"@type":"Question","name":"What bug was found in the us-tax-model project's Form 1040 calculation despite having two independent evaluators agree?","acceptedAnswer":{"@type":"Answer","text":"Form 1040 line 9 incorrectly summed line 3a along with line 3b, and both the TypeScript and Lean evaluators agreed with the wrong sum because they read the same underlying data. The bug was caught only by an outside engine comparison, not by the internal fixtures or evaluator agreement, and is recorded as BUG-0001 in the project's append-only bug inventory. Anyone weighing how much to trust internal test agreement can track cases like this on daily.dev."}},{"@type":"Question","name":"What percentage of US individual tax returns can the us-tax-model project currently handle without any refusals?","acceptedAnswer":{"@type":"Answer","text":"The measured filer coverage is 68.6% of individual returns, calculated by tools/coverage/estimate.py, where a return only counts if nothing on it triggers a refusal gate. However, 148 of 157 refusal gates lack IRS count or microdata measurement, so true coverage is acknowledged to be lower than the measured figure suggests. Developers evaluating early-stage open-source coverage claims can follow projects like this on daily.dev."}}]} ```