If you go to https://t.co/k2PToaXnUx make sure you scroll down to this part. It’s what helped it click for me.
Agents are forced to design a decoupled architecture using Commands and Events.
Tests are then just a bunch of “Given, When, Thens” which AI is so good at.
The `effect/net` family is growing. We've now got Datagram sockets (UDP) and a DNS resolver in core with accompanying data types and schemas for dns records & host / domain name handling: https://t.co/BTUhJIuX4o and https://t.co/37nIsPvL3h
@theo Speaking of `ssh`, you guys still use the host's native ssh client? I've got a zero dependency Effect built-in ssh client coming up some time later today :-P. Scary, I know.
Streaming today at 2:30 PM ET with @maxwellbrown of @EffectTS_ and @samgoodwin89 of @alchemy_run.
We’ll be having some fun building with the CAFE stack and probably getting distracted. Come hang out!
(will repost with the link)
Streaming today at 2:30 PM ET with @maxwellbrown of @EffectTS_ and @samgoodwin89 of @alchemy_run.
We’ll be having some fun building with the CAFE stack and probably getting distracted. Come hang out!
(will repost with the link)
Introducing: The next evolution of Polar ✨
We’re building the most ambitious metering & billing substrate for AI software.
— Events as the core primitive. A blazing fast, local-first reducer architecture to achieve zero latency metering & entitlement resolution.
— Your business as a definition, version controlled & native to agentic workflows.
— Definition branching for simulating pricing changes, model forecasts & easy deployment to production.
— Signals as business guardrails. Switch to a cheaper AI model when customer’s credits run below a threshold.
— Identities as nested entities under the Customer model. Built for agents & services acting on behalf of the customer.
Coming soon.
Learn more → https://t.co/pEdxMjNsTZ