@DavidKPiano@titanXio@restatedev My cto says temporal so much we have a fake drinking game on every mention that would have killed us all long ago ๐
Just dropped restated temporal vs restate into slack and requested his review. ๐ฟ
Spent nearly 18 months building durable workflows (on top of Qstash) at @titanXio - so last summer when it came time to build what the cool kids now call agent harnesses and loops, it was only natural to reach for @temporalio. Durable state machines, with a registry of deterministic activities that call probabilistic LLMS is the way.
Domain expertise is 100% the moat. If you are an expert in your field and you can make your implicit knowledge more explicit (eg write it down), you will absolutely crush it if you learn how to use these new tools we call โlanguage modelsโ - trust me, whatever you learned in your domain is much more complex than learning how to do the thing with language models - so just dive in, or find a person who knows a bit more that will tug you along
@wesbos One of the most interesting things Ive done professionally was @teropaโs build your own angularjs. The transition from โI know best practicesโ to โI deeply understand why to do things all the ways and the optimal path for my use caseโ was so dope.
Alternative theory: go where your competition is dumb. Why compete with the smartest most well funded people on earth when you can compete against notepad and spreadsheets?
Iโve noticed some startups pivot away from markets that feel overly hot because they think the contrarian move is to go somewhere less crowded.
But the real issue is not the market. It is that the team lacks the domain expertise needed to break through.
That quickly turns into negative self talk
The market is too hot. Every competitor is already doing it. There is no margin left. It is not technically possible.
Then an extremely domain specific team enters the same market and figures out every problem the other teams could not solve.
This is why domain expertise or passion matters so much when choosing a startup idea.