7

Source document

[2025-04-28 07:38] lisamegawatts: roo code is agent team in ide, they suggest if you need another agent on your team you can input a job posting for that agent it will create the prompt for it [2025-04-28 07:40] lisamegawatts: i have it running autonomously in vs code with a small local model, will report back on results :p [2025-04-28 18:15] traves_theberge: Interesting 🤔 resumes as instructional prompts lol 😂 savage [2025-04-28 18:15] traves_theberge: Guess I’m going Christmas shopping for resumes 🫡 [2025-04-28 18:17] traves_theberge: Guys let’s write a conversation framework for JSON resume [2025-04-28 18:32] traves_theberge: 🔥🔥 [2025-04-28 20:35] ajaxdavis: ahh i finally get it. it is a great idea [2025-04-28 20:36] ajaxdavis: someone make me an agent registry.jsonresume.org/thomasdavis [2025-04-28 21:08] traves_theberge: ill work on it [2025-04-28 21:08] traves_theberge: im willing to work somthing like this out. gotta do it for work anyway [2025-04-28 21:10] traves_theberge: its more about How we want to build that entitiy and how to manage the History as Skills, and Tools ect. I would be intreasting to have a small vector embedding pipeline [2025-04-28 21:10] traves_theberge: some thinking to do [2025-05-07 10:39] ajaxdavis: https://github.com/plexe-ai/plexe [2025-05-08 05:30] ajaxdavis: https://github.com/AnswerDotAI/rerankers [2025-05-09 08:03] lisamegawatts: https://github.com/choyingw/Cross-Modal-Perceptionist [2025-05-09 08:04] lisamegawatts: i actually think you could predict face geometry and body with sufficient voice data [2025-05-09 15:40] jonathan.poczatek: Is this the time that I pitch the grand idea because everybody's ready?? I mentioned this a while back because it made a lot of sense to me for some reason. We're talking about agentic frameworks and stuff like that, and I think there was a discussion about what he's an agent, how people define it and what not [2025-05-09 15:42] jonathan.poczatek: Ultimately, the test that I was using to determine if these agenda frameworks made any sense, try to understand what an agent would be doing as mapped to your own behaviors. Like real world stuff. That's why Google's task-based agent model made a lot of sense, it was very simple to apply that way of modeling agentic behavior to our own human behavior, like in our workspace. Org charts and shit [2025-05-12 03:08] ajaxdavis: https://www.primeintellect.ai/blog/intellect-2-release [2025-05-14 03:31] ajaxdavis: learnt hard way at work. those openai usage limits actually are just alerts lol (files: image.png) [2025-05-14 03:32] ajaxdavis: I don't have auto recharge on my personal account or would have fell for it too [2025-05-14 03:47] traves_theberge: 0_o no way!