<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Generative AI on Ashish Saini</title><link>https://ashishsaini.work/tags/generative-ai/</link><description>Recent content in Generative AI on Ashish Saini</description><generator>Hugo -- 0.154.0</generator><language>en-us</language><lastBuildDate>Sat, 06 Jun 2026 10:00:00 +0530</lastBuildDate><atom:link href="https://ashishsaini.work/tags/generative-ai/index.xml" rel="self" type="application/rss+xml"/><item><title>From still renders to a walkable building: image-to-video and Claude are rewriting how architecture gets shown</title><link>https://ashishsaini.work/posts/walkable-building-image-to-video-claude/</link><pubDate>Sat, 06 Jun 2026 10:00:00 +0530</pubDate><guid>https://ashishsaini.work/posts/walkable-building-image-to-video-claude/</guid><description>Image-to-video models invent the motion between architectural renders; Claude turns the footage into a scroll-driven walk. A small experiment that cost about $4.5 and an afternoon.</description></item><item><title>From 73% to 100%: a GenAI photoshoot pipeline that cleared the catalogue</title><link>https://ashishsaini.work/posts/73-to-100-genai-photoshoot-pipeline/</link><pubDate>Tue, 26 Nov 2024 10:15:00 +0530</pubDate><guid>https://ashishsaini.work/posts/73-to-100-genai-photoshoot-pipeline/</guid><description>A quarter of the catalogue had no model photo, and the physical photoshoot couldn&amp;#39;t catch up. We closed the gap in a month with Stable Diffusion, ComfyUI, and a lot of unglamorous QA.</description></item></channel></rss>