Hey, good morning! I’m feeling a bit scattered this morning, so I’m trying to organize my thoughts and figure out my day.
I’m pretty stoked about what I accomplished yesterday. I spent the day building tiny AI bots, essentially creating small AI agents. I came up with a cool chat interface that lets me switch between different Large Language Model (LLM) modes inside my Next.js app. With this setup, I can easily switch between API routes, each featuring a different system prompt and AI persona. In the same chat, I can talk to a CTO, a CEO, and even a copywriting assistant.
This tool has been super useful for me as I’m working on a pitch deck for my AI leasing assistant startup. It’s made streamlining the pitch a lot easier, and I plan to use it more today to polish the copy for each slide. The idea is simple yet powerful: dump all my raw content into the LLM, and let it figure out how to articulate my thoughts in an effective and eloquent way. Properly prompted, the LLM can be much more articulate than I could ever hope to be.
The concept of feeding LLMs raw data extends beyond just pitch decks. Yesterday, I also thought about how this technology could revolutionize resume writing. Imagine dumping a bunch of ideas and experiences into an LLM, and having it generate a polished resume for you. This could be game-changing, especially for students who might not have the resources to pay for professional resume services.
I want to experiment with creating a resume writer. The idea is simple: input your experiences, education, skills, and generate a resume that can be downloaded as a PDF or shared as a web page. Users should be able to edit sections, delete experiences, and add new ones—much like editing a LinkedIn profile. Core features would include sections for experiences, education, skills, a bio, and headline generators. Maybe even placeholders for images and dates.
Technically, this wouldn’t be too challenging. Using React for the UI and Tailwind for styling, we could create sections for different parts of the resume: summary, work experience, education, etc. Each section would be generated by mapping over the data provided. I believe an MVP could be put together in a day, and we could start commercializing this idea soon after.
The main reason I’m pursuing this is to see if others find it as useful as I do. It’s all about the comfort of dumping a ton of data into an LLM and having it generate the end product for you. People are tired of filling out endless input fields, and LLMs can handle this much faster and more efficiently. Plus, the inference speeds of modern LLMs are incredible. I’m excited to experiment with this idea and see where it takes me. Let’s see if we can come up with an MVP today and take it from there!