Using Jev for between-step choices in a Hermes workflow
The post proposes using the Decisions API or Jev in Hermes to judge tool selection, model routing, failed-job priority and whether a draft needs review, with Hermes executing the chosen action; the author notes the 10x speedup applies to decision calls and real gains need testing.
Up to 10× faster decisions. Here’s where Hermes could use that. OpenAI says its new Decisions API makes decisions up to 10× faster than GPT-6 Luna through the Responses API. Think about the small choices inside an agent workflow: → Which tool should handle this task? → Which model should get this request? → Which failed job needs attention first? → Is this draft ready—or does it need review? A proposed Hermes setup: use Decisions API or Jev to evaluate those choices, then let Hermes execute the selected action. That’s where I’d test it: faster choices between steps. The 10× claim is about decision calls. Actual Hermes speed gains would need testing. Which would you want first: tool selection, model routing, or draft review?
This post quotes another post: x.com/1633874951508721686/status/210757338222918…