@paraschopra The problem is agents training data is an average of all the content we have. Taste is always cutting edge, whats common becomes boring and out of taste. This way llms always keep catching up
@tamarajtran@typesafeai As far as i know jev only offers input limit of 64k max context window so this might not be useful in production for this usecase at the moment
if you havenβt tried @ao_build in a while, itβs worth another look.
the team and contributors have been putting a lot of work into the UI and overall dev experience over the past month. nice to get feedback like this from people.
try it on something youβre building, not just a quick look at the dashboard. would love to hear what works for you and what still feels rough :)