✕ Clear all filters
15 articles

📰 Dev.to · qcrao

15 articles · Updated every 3 hours · View all reads

All Articles 99,885Blog Posts 115,178Tech Tutorials 25,231Research Papers 20,779News 15,817 ⚡ AI Lessons
Why splitting storyboard from rendering beats one mega-prompt for AI comics
Dev.to · qcrao 1mo ago
Why splitting storyboard from rendering beats one mega-prompt for AI comics
The most tempting way to build an AI comic generator is also the worst one: take a paragraph of...
Why I chose human-edited subtitles over AI auto-captions for vocabulary mining
Dev.to · qcrao 1mo ago
Why I chose human-edited subtitles over AI auto-captions for vocabulary mining
When I was prototyping TubeVocab, the obvious shortcut was to use YouTube's auto-generated captions...
Why single-panel editing matters more than perfect first-shot AI comics
Dev.to · qcrao 1mo ago
Why single-panel editing matters more than perfect first-shot AI comics
The biggest misunderstanding about AI comic generation is the expectation that the model should...
Character consistency in AI comics: 3 tricks that beat LoRA training for me
Dev.to · qcrao 1mo ago
Character consistency in AI comics: 3 tricks that beat LoRA training for me
The single thing that breaks an AI-generated comic isn't the art style or the prompt — it's the...
How I cut speech-bubble retries from 70% to 0% with 200 lines of Pillow code
Dev.to · qcrao 1mo ago
How I cut speech-bubble retries from 70% to 0% with 200 lines of Pillow code
If you've ever asked Stable Diffusion or DALL-E to render readable text inside a comic panel, you...
7 prompt engineering tricks that pulled my AI comic costs from $0.20 to $0.038/panel
Dev.to · qcrao 1mo ago
7 prompt engineering tricks that pulled my AI comic costs from $0.20 to $0.038/panel
Six months ago, generating a single 4-panel comic on Comicory cost me ~$0.80 in GPU time and produced...
How I scraped 50k YouTube subtitles in 2 weeks for $7 (and the legal gray zones)
Dev.to · qcrao 1mo ago
How I scraped 50k YouTube subtitles in 2 weeks for $7 (and the legal gray zones)
When I started building TubeVocab, I had a chicken-and-egg problem. I needed a corpus of YouTube...
5 LoRA training pitfalls when you're trying to lock down a comic character
Dev.to · qcrao 1mo ago
5 LoRA training pitfalls when you're trying to lock down a comic character
TLDR: Most "my LoRA works in test prompts but breaks the second I put it in a comic panel" problems...
What I learned squeezing the YouTube Data API v3 quota for a side project
Dev.to · qcrao 1mo ago
What I learned squeezing the YouTube Data API v3 quota for a side project
TLDR: The default 10,000 unit/day quota will burn through in ~10 naive user requests. Three tricks...
The Engineering Challenge of Turning YouTube Into an ESL Corpus
Dev.to · qcrao 2mo ago
The Engineering Challenge of Turning YouTube Into an ESL Corpus
Language learning apps have spent a decade chasing the same pattern: curate a 2,000-word...
Why Character Consistency Is Hard in AI Comic Generation
Dev.to · qcrao 2mo ago
Why Character Consistency Is Hard in AI Comic Generation
When you feed a story prompt into a generic image AI — say, "a detective with a red scarf walks into...