<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Neil Williams</title><description>Neil Williams writes about legal technology, AI tools, current work, and the projects he makes.</description><link>https://neilofneils.com/</link><item><title>Building an internal AI litigation workbench</title><link>https://neilofneils.com/posts/building-a-legal-ai-workbench-not-another-chatbot/</link><guid isPermaLink="true">https://neilofneils.com/posts/building-a-legal-ai-workbench-not-another-chatbot/</guid><description>A local-first AI workbench for litigation teams, organized around files, progress, work products, source context, and human review instead of model choice and prompt tricks.</description><pubDate>Fri, 05 Jun 2026 14:45:00 GMT</pubDate></item><item><title>DishCache: Saving the Recipe, Not the Life Story</title><link>https://neilofneils.com/posts/building-dishcache-a-better-home-for-recipes/</link><guid isPermaLink="true">https://neilofneils.com/posts/building-dishcache-a-better-home-for-recipes/</guid><description>DishCache is a recipe app built around a simple idea: save recipes you care about, capture the useful parts from links, and strip away the bloated blog narrative that gets in the way.</description><pubDate>Fri, 03 Apr 2026 18:30:00 GMT</pubDate></item><item><title>Making a discovery production queryable with local retrieval</title><link>https://neilofneils.com/posts/turning-a-discovery-production-into-something-an-llm-can-actually-query/</link><guid isPermaLink="true">https://neilofneils.com/posts/turning-a-discovery-production-into-something-an-llm-can-actually-query/</guid><description>Notes from building a local pipeline for a Relativity-style export: parse the load file, normalize the text, index useful chunks, and make the model answer from retrieved documents.</description><pubDate>Fri, 03 Apr 2026 13:00:00 GMT</pubDate></item></channel></rss>