<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/the-self-improving-loop-breakthrough-in-avalonia-hlxpizmbs" -->

---
title: The Self-Improving Loop Breakthrough in Avalonia | daily.dev
description: After 6 months of failure, my model just taught itself to build pixel-perfect Avalonia UI — with no screenshots, no vision model, and no help from me.  This is...
canonical: https://daily.dev/posts/the-self-improving-loop-breakthrough-in-avalonia-hlxpizmbs
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: The Self-Improving Loop Breakthrough in Avalonia | daily.dev
og:description: After 6 months of failure, my model just taught itself to build pixel-perfect Avalonia UI — with no screenshots, no vision model, and no help from me.  This is...
og:url: https://daily.dev/posts/the-self-improving-loop-breakthrough-in-avalonia-hlxpizmbs
og:image: https://api.daily.dev/og/posts/hLxPiZmbs.png
og:image:alt: The Self-Improving Loop Breakthrough in Avalonia
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 Self-Improving Loop Breakthrough in Avalonia

**[AngelSix](https://daily.dev/sources/angelsix)** · 11 min read · 0 upvotes · 0 comments

## Summary

After 6 months of failure, my model just taught itself to build pixel-perfect Avalonia UI — with no screenshots, no vision model, and no help from me.

This is the follow-up to Model Training Primer, and this is the breakthrough. For the first time I have a truly self-contained, self-improving loop where my custom model can see layout data, catch its own padding and margin mistakes, fix them, and keep learning indefinitely. It even found a bug in how thickness tokens render.

If you've ever tried to get AI to make Avalonia UI you'll know it's absolutely awful — until now.

WHAT YOU'LL LEARN:
- Why primers, docs, rules and frontier models all failed miserably for Avalonia UI
- How God's Amul mechanical code catches the classic LLM Avalonia mistakes
- How I wired XAML compile errors from Rider into an auto-harness
- How my headless frame buffer engine lets a non-vision model see layout as structured data
- The live demo: self-detected search bar padding fix + font boundary understanding with zero visual input
- The thickness token bug my model found on its own
- What's next: indefinite training toward a full theme engine that can reproduce gorgeous complex UI verbatim

---
TIMESTAMPS:
00:00 - This doesn't look impressive, but it's massive
00:53 - Why every model fails at Avalonia UI
01:27 - First thing that worked: God's Amul
02:30 - Next level: XAML compile error harness
02:51 - Why screenshots + vision models can't work
03:40 - First pixel-perfect breakthrough
04:29 - No UI, no screenshots, it fixed padding by itself
05:12 - It found a bug in thickness tokens
05:50 - Done entirely in code, no vision
07:15 - I can now leave training running indefinitely
10:12 - Testing on Javier's gorgeous Avalonia design
11:44 - Amazing things coming soon

---
LINKS:
GitHub: https://github.com/angelsix
Website: https://www.angelsix.com
Avalonia UI: https://www.avaloniaui.net

#AvaloniaUI #LLMTraining #LocalLLM #CSharp #DotNet

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://www.youtube.com/watch?v=e2FfsKCT0ro>

---

Tags: [#machine-learning](https://daily.dev/tags/machine-learning), [#llm](https://daily.dev/tags/llm)

[View this post on daily.dev](https://daily.dev/posts/the-self-improving-loop-breakthrough-in-avalonia-hlxpizmbs)

```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 Self-Improving Loop Breakthrough in Avalonia","url":"https://daily.dev/posts/the-self-improving-loop-breakthrough-in-avalonia-hlxpizmbs","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/the-self-improving-loop-breakthrough-in-avalonia-hlxpizmbs"},"datePublished":"2026-10-04T13:04:27.781Z","dateModified":"2026-10-04T21:56:42.991Z","description":"After 6 months of failure, my model just taught itself to build pixel-perfect Avalonia UI — with no screenshots, no vision model, and no help from me.\n\nThis is...","image":"https://i.ytimg.com/vi/e2FfsKCT0ro/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/e2FfsKCT0ro/sddefault.jpg","isAccessibleForFree":true,"articleSection":"AngelSix","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":"AngelSix","logo":"https://media.daily.dev/image/upload/s--h1ikmN_u--/f_auto,q_auto/v1791119063/logos/angelsix?_a=BAMAMicg0","url":"https://daily.dev/sources/angelsix"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/the-self-improving-loop-breakthrough-in-avalonia-hlxpizmbs","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"machine-learning,llm","timeRequired":"PT11M","video":{"@type":"VideoObject","name":"The Self-Improving Loop Breakthrough in Avalonia","description":"After 6 months of failure, my model just taught itself to build pixel-perfect Avalonia UI — with no screenshots, no vision model, and no help from me.\n\nThis is...","thumbnailUrl":"https://i.ytimg.com/vi/e2FfsKCT0ro/sddefault.jpg","uploadDate":"2026-10-04T13:04:27.781Z","duration":"PT11M","url":"https://api.daily.dev/r/hLxPiZmbs","embedUrl":"https://www.youtube.com/embed/e2FfsKCT0ro"}}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"AngelSix","item":"https://daily.dev/sources/angelsix"},{"@type":"ListItem","position":3,"name":"The Self-Improving Loop Breakthrough in Avalonia"}]}
```

