My cofounder Kanyes and I have been power users of a lot of second brain type software like Notion and Obsidian for years, and tried fine-tuning GPT-2 back in 2020 to solve this exact problem of giving AI access to all context of my life so it can do work for me more intelligently. We were too early then, and spent the next 4 years building robots at Matic. Earlier this year, we realized this was perhaps the right time to take a second crack at this problem, so we quit our jobs and built Hyper. It's super exciting that we finally shipped a self-serve version that you can try out yourself from the website!
Happy to answer any questions below, technical or otherwise, and hope this is something useful!
May I ask how did you make those? For my videos I just use windows snipping tool and they look okay ish, but very amateur. It seems like yours was professional video editor work and not just recording using the tool.
Additionally: in what ways does the product differ from a prompt on a chat + pumping information there? It has MCP connections/orchestrations that facilitate this process, making it more automatic after the initial setup?