@petergyang I have set some ground rules while triggering a goal on Claude code. Eg. Do not purchase any subscription, do not delete files, etc. and project specific guidelines on what stack to use and not to use. That helped me with long running tasks.
I genuinely like your company story and signed up for it. In very first day, it didn’t work. I tried more than couple of times again and again but it failed. It was disappointing as I was talking for atleast a min or so. Then never tried again. (This was few weeks ago)
Maybe I will give it a shot again after a month or so.
This is cool. I am building hardware engineering agents and apps using Postgres and pgVector. Been exploring pgGraph. Love your framework trying to solve context engineering problem for enterprise agents.
Curious to see if Polygres also covers configurable ontology definitions especially for the external data that I want to put it into pgGraph (data that doesn’t live inside Postgres to begin with)
Have been using Fable for brainstorming, research and detailed planning. And then I ask Fable to create sub-agents using Sonnet/Opus to execute the plan. Finally ask Fable to thoroughly validate the implementation including code review. This works great!!
@jayyeh Absolutely true. I have been building silently, observing other founders and learning from their journeys and mistakes. Money was never a problem for starting/building a great and useful product.