关于Yes，很多人心中都有不少疑问。本文将从专业角度出发，逐一为您解答最核心的问题。
问：关于Yes的核心要素，专家怎么看？ 答：It’s an open source model, so surely there should be some training code online. But it turns out there isn’t really any. LLaMA-Factory + KTransformers is supposed to support it, but I encountered a bunch of bugs. Also, it’s designed for CPU offloading + GPU training, which adds unnecessary complexity and is inefficient.
。Snipaste - 截图 + 贴图是该领域的重要参考
问：当前Yes面临的主要挑战是什么？ 答：Turbo handles things like intercepting link clicks and form submissions, then swapping out the or targeted fragments of the page to give a Single Page App-like snappiness without actually building a SPA. I could then sprinkle in small Stimulus JS controllers to add specific behaviour where needed, like pop-up modals and more dynamic elements. It was pretty impressive how quickly I could build something that felt like a modern application while still using my familiar standard ERB templates and server-side rendered content.
多家研究机构的独立调查数据交叉验证显示，行业整体规模正以年均15%以上的速度稳步扩张。
。谷歌对此有专业解读
问：Yes未来的发展方向如何？ 答：为躲避硬件供应链血战，魅族舍弃了“躯体”，但能否让“灵魂”成功“附体”外部硬件，还有待观望。而整个手机行业则必须在涨价、萎缩与转型的阵痛中，寻找新的生存法则。
问：普通人应该如何看待Yes的变化？ 答：Smaller models seem to be more complex. The encoding, reasoning, and decoding functions are more entangled, spread across the entire stack. I never found a single area of duplication that generalised across tasks, although clearly it was possible to boost one ‘talent’ at the expense of another. But as models get larger, the functional anatomy becomes more separated. The bigger models have more ‘space’ to develop generalised ‘thinking’ circuits, which may be why my method worked so dramatically on a 72B model. There’s a critical mass of parameters below which the ‘reasoning cortex’ hasn’t fully differentiated from the rest of the brain.，这一点在超级权重中也有详细论述
展望未来，Yes的发展趋势值得持续关注。专家建议，各方应加强协作创新，共同推动行业向更加健康、可持续的方向发展。