Grok Bot 多智能体工作坊与 Jev 路由对比
作者看完 54 分钟 Grok Bot 团队协作教程后指出，其路由决策全靠前沿模型输出散文，而 Jev 直接返回带类型的选项和概率。
SpaceXAI just dropped a free Grok Bot workshop 54 minutes on running a bot team like a real product org Grok Bot → Specialists → Shared Context → Delegation → Self-Improving Teams 7% → 03:37 - Why coordinating multiple agent threads gets difficult 25% → 13:32 - Build a team around real company roles 44% → 23:59 - Let agents handle several questions in parallel 56% → 30:08 - Share context without copy-pasting 75% → 40:40 - Watch an agent learn to delegate and review work Every routing decision comes back as prose from a frontier model Who handles the task → What to check → Who signs off That’s not an intelligence architecture Jev returns typed options and probabilities without writing a sentence LLM creates the work → Jev chooses what happens next → Code enforces the decision Save this before you build another coordinator agent ↓
这条是引用着另一条写的：x.com/0xwhrrari/status/2107093781807775896