Haven’t benchmarked them head-to-head yet, but agentic work is where the real test is. Give all three the same messy, multi-step task with tools + memory + retries and see who actually finishes it with the least babysitting. 😅
genuine question:
if AI can eventually write your code, design your UI, test your product and deploy it…
what exactly should a developer spend most of their time learning?
I keep changing my answer.
Wait… are AI agents about to get their own “brain” for making decisions?
Jev isn’t trying to write better essays.
It’s built to decide things like:
→ which tool to use
→ whether to retry
→ what action an agent should take
AI is slowly moving from “generate” → “decide” → “act.”
That’s crazy.