https://suriya.cc/ideas/Recent content in Ideas on Suriya's siteHugoen-usThis work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.Sat, 25 Apr 2026 18:40:16 -0400https://suriya.cc/ideas/research/Sun, 04 Feb 2024 00:00:00 +0000https://suriya.cc/ideas/research/<h4 id="ai-guardrail-comparison">ai guardrail comparison</h4> <ul> <li>There are multipel LLM guardrails, evaluating them would be something</li> </ul> <h4 id="privacy-detector-dataset-and-custom-model">privacy detector dataset and custom model</h4> <ul> <li>There is no dataset to test a model for it&rsquo;s privacy Capabilities</li> <li>A custom model implementation that tests this information against an LLM would also be helpful</li> </ul>https://suriya.cc/ideas/product/Sat, 16 Jul 2022 00:00:00 +0000https://suriya.cc/ideas/product/<p>Inspiration comes from many places. I love jotting them down here.</p> <h3 id="small-app-ideas">small app ideas</h3> <ul> <li>virtual bookshelf where you can showcase your books. <ul> <li>I wonder if this can be self serve. like I can just take a video of a book, convert it into a gaussian cloud and create a virtual library.</li> <li>Are Gaussian models there, where I can run them client side.</li> <li>the goal is to get stripe press like aesthetics for books.</li> <li>I&rsquo;m not sure about the legality.</li> <li>It might be easier to have a book shape and attach the front back etc.</li> </ul> </li> <li>link storage and recording as a telegram bot or email etc with the ability to show case in a website</li> <li>Something with Audio models such as <a href="https://fugatto.github.io/">fugatto</a> <ul> <li>Isolate tracks</li> <li>humming -&gt; Song</li> <li>Unlimited song/rap/metallic generation.</li> <li>Introduce a product</li> <li>Introduce</li> </ul> </li> <li>Ideas with <a href="https://www.anthropic.com/news/model-context-protocol">Model Context Protocol</a> enabling LLM Agents with cool tools <ul> <li>Object storage for models</li> <li>Memory layer for any agent</li> <li>Auth layer for agents to authenticate into a page. It is quite obvious how this can even be extended into a secret manager for agents as when these things grow there is a need to manage private and personal data. Maybe credit card information, etc.</li> </ul> </li> <li>Inventory management with Video models. Going from video -&gt; list of inventory. Gemini can do it, tested.</li> <li>talk with a youtube video <ul> <li>Simple extension to inject a chat window + summarization</li> </ul> </li> <li>Talking with a document</li> <li>Unstructured data to structured data</li> <li>Image or Video cropping with SAM2 or <a href="https://yangchris11.github.io/samurai/">SAMURAI</a> models.</li> <li>Video Editing LLM going from a long form video</li> <li>LLamacoder for any LL <ul> <li><a href="https://github.com/nutlope/llamacoder">https://github.com/nutlope/llamacoder</a> </li> </ul> </li> <li>async prompt processing APIs</li> <li>Python library for overarching batch prompts</li> <li>Release Youtube chrome Extension</li> <li>Browser automation for testing, automatically booking tickets etc.</li> <li>native MP3 player, with a remote library</li> </ul> <h2 id="problems-i-have">problems i have</h2> <ul> <li> <p>Foss app to send pictures to a widget in someone else&rsquo;s phone</p>