Stop building AI agents that demand perfect context. Start building agents that make imperfect context productive. Curiosity is more important than comprehensiveness for meaningful human-AI interaction.
I think the point of CS at most universities isn’t to teach the students how to code, but to understand the environment and logic that goes into making code exist in the first place. Maybe?
It’s like going to college to learn about making furniture, but all the classes are about how trees grow and the intracacies of different wood types. Like bro, I’m just trying to build tables, teach me how to use a hammer, saw, CNC machine.
A lot left to be desired when I’d say 60% of our graduating class couldn’t build a working program if it meant life or death. Yet, they might be able to explain Big O, tree traversal, etc.
We have an insurance-tech startup that automates compliance, lead generation, lead communication, claims, and policies for insurance agencies.
Our ML models learn what type of leads certain agents do well with, and distribute them accordingly. We also automate 10DLC compliance using AI.
there are no decent calendars built with @shadcn, so we are going to open source our calendar. create calendars, sync with oauth, custom db tables, coming soon.