I Turned Coding Agents Into a Strategy Game — Ido Salomon, AgentCraft
频道: AI Engineer
视频: https://www.youtube.com/watch?v=YIVkERhy8xo
原文语言: en
统计: 共 12 轮 · Ido 12
[0:21] Ido
Okay. Hi everyone. Uh I'm Solomon. I am the creator of ATMCraft and I also created MCPUI which became MCP apps and MCP steering committee. I've done a bunch of other stuff. Uh but we're here to talk about why we're the bottleneck and why we don't have to be. So, uh as you all know and use every day, agents are amazing. There's so much they can do uh from generating cat memes to B2B SAS apps. So if they are so amazing, we have to ask why aren't we all unstoppable? Why doesn't everyone have an army of agents that just does whatever they want? So it should be relatively easy. I mean you just have a bunch of monitors and you spin up 25 cloud code agents and you're done. So the reality is that it's just not that simple to do something like that. you need to a really want to use agents and secondly once you actually spin them up each one of these agents requires you to steer them direct them review them and when it's at scale especially it's just exhausting we need to find the skills to allow us to do it without burning out so in reality we are the bottleneck and the the cool thing here is that while we are the bottleneck The skills that we need to do these things to uh find ourselves not burning out every time we use an agent are with us all along. Uh if you try to compare it to other stuff that you've been doing outside of programming, for example, playing video games, it's not really that different. Like if if you're playing Warcraft, for example, or Sims, and you have a bunch of different uh units running around, you need to understand what they're doing and supervise all of it. How different is that from from actually inspecting agents? So I took this concept uh too far and I created agent uh which is a game inspired orchestrator uh that tries to bring all that good stuff together into work.
好，大家好。我是 Ido Salomon，AgentCraft 的作者。我还做了 MCP-UI，它后来演变成了 MCP Apps，另外还有 MCP 指导委员会（steering committee）。我还做过一堆别的东西。不过今天我们要聊的是：为什么我们自己成了瓶颈，以及为什么我们其实不必如此。大家都知道，也每天都在用——agent 真的很厉害。它们能干的事太多了，从生成猫咪表情包到做 B2B SaaS 应用都行。既然它们这么厉害，那就得问一句：为什么我们还没个个都势不可挡？为什么不是每个人都有一支 agent 大军，让干啥就干啥？按理说这应该挺容易的：弄一堆显示器，开上 25 个 Claude Code agent，完事。但现实是，想做到这一步根本没那么简单。首先，你得真心想用 agent；其次，一旦你真把它们跑起来，每一个 agent 都需要你去掌舵、指挥、评审，尤其是规模一上来，简直累死人。我们得找到一些技能，让自己能干这些事又不至于累垮。所以实际上，瓶颈是我们自己。而有意思的是，虽然我们是瓶颈，但要做好这些事、不至于每用一次 agent 就累垮一次，所需要的技能其实一直都在我们身上。你拿它跟你在编程之外做过的事比一比，比如打电子游戏，其实没多大区别。比如你在玩 Warcraft（魔兽争霸）或者 Sims（模拟人生），手下有一堆不同的单位跑来跑去，你得搞清楚它们各自在干什么，还得统管全局。这跟真正去盯着 agent 干活，又能差多少呢？所以我把这个想法玩过了头，做出了 AgentCraft——一个受游戏启发的编排器（orchestrator），想把游戏里那些好东西都带进工作里来。
[2:30] Ido
So let's take uh a quick look of what that actually looks like and try and understand the journey that we can go through to really raise the ceiling of how humans and agents can collaborate. So, uh, this is AgentCraft. Uh, there's really a lot to unpack here. Uh, so we'll just go through the basics really quick. Um, what you see here is an agent. Uh, it's an actual representation of a cloud code or openclaw, whatever agent that you can imagine. Uh, it can be detected on your device or you can actually spawn new ones directly from a craft. Uh, so like I said, you have like codecs and uh, open code etc. at your command.
那我们快速看一眼它实际长什么样，也试着理解一下这段旅程——怎样才能真正抬高人和 agent 协作的天花板。这就是 AgentCraft。这里面值得展开的东西真的很多，所以我们只快速过一下基础。你在这儿看到的是一个 agent，它对应的是一个真实的 Claude Code、OpenClaw，或者任何你能想到的 agent。它可以是在你设备上检测到的，你也可以直接在 AgentCraft 里拉起新的 agent。所以就像我说的，Codex、OpenCode 等等，全都听你调遣。
[3:13] Ido
Uh, and these agents can do everything. They're just normal agents. Uh, so I can just prompt them in this nice side panel that also has voice. So there's multimodality here. Uh, and just do something for me.
这些 agent 什么都能干，它们就是普通的 agent。所以我可以直接在这个挺好用的侧边栏里给它们发 prompt，侧边栏还支持语音，所以这里是多模态的。然后就让它帮我干点什么。
[3:27] Ido
Prompt. Prompt.
Prompt。Prompt。
[3:29] Ido
And now these agents can do whatever. Uh, and if you look around you would see that the map has a a bunch of different elements. So you have buildings. These buildings represent functionalities. So for example, if you can manage your plugins or skills, uh you have like integrated terminal, integrated git, uh you can do in your all of your work from within this space. So now that we have this and we understand the basics, let's look at what we can do uh to raise the ceiling. So the first step is obviously visibility. I need to understand what's going on. What are my agents doing? So, I have this nice side panel and it shows me the different agents and what their task is and what the last thing they did is and what they're currently doing. Everything I need to quickly understand who needs my attention and who doesn't. The second part of it is we're not just limited to this. We have a map. So, I can take let's say my file system, project it onto the map, and now I can see exactly what my agents are doing. Not just by looking at a list of stuff. I can actually see okay this agent is now working on this directory or this file and these files are represented as runes on the map. So I can really look at each one and see exactly what each agent did and when. And now I can take this nice information because we are the orchestrator uh and use that to create linkages and now I can create cool stuff like heat maps. So I can really easily see what's going on. And the the other aspect is that now that we have this visibility, that's not really enough. We need to be able to react to what's going on. If someone needs my attention, I need to be able to quickly jump to it.
现在这些 agent 就可以放手去干了。你四处看看，会发现地图上有一堆不同的元素。比如有建筑，这些建筑代表各种功能：你可以管理自己的插件或 skill，还有内置终端、内置 git，你所有的工作都能在这个空间里完成。好，现在这些有了，基础也了解了，我们来看看能做点什么来抬高天花板。第一步显然是可见性（visibility）。我得知道现在发生了什么——我的 agent 都在干嘛？所以我有这个侧边栏，上面列着各个 agent、各自的任务、上一步做了什么、现在正在做什么。我需要的一切都在这儿，能让我迅速看出谁需要我关注、谁不需要。第二部分是，我们不止于此——我们还有地图。比如说，我可以把自己的文件系统投射到地图上，这样就能清清楚楚看到我的 agent 在干什么，而不只是盯着一张列表。我能真正看到：哦，这个 agent 现在正在处理这个目录或者这个文件；这些文件在地图上是以符文（rune）的形式呈现的。所以我可以逐个细看，准确知道每个 agent 在什么时候做了什么。而且因为我们本身就是编排器，我可以拿这些信息建立关联，做出热力图这类很酷的东西，一眼就能看清发生了什么。另一方面，光有可见性其实还不够，我们得能对正在发生的事做出反应。有谁需要我关注，我得能迅速跳过去。
[5:12] Ido
So much like video games or RTS games or Civilization, whatever, you would click uh spacebar and it would just take you to the nearest thing that actually needed your attention. So you can really easily understand what's going on. and jump from one thing to the other and answer questions or approve plans, etc. That was nice. I mean, it's a nice step just to have that visibility, but it's not really enough. It's pretty exhausting to keep looking at what the agents are doing in the granularity of what files are they using and what tools are they using at the moment. So, we need something else. Uh and the reason is that we have like two two main problems here. Uh one is that there's just too much to hold in our heads. Even if I can understand what's going on and I can see everything, I can't really do like 20 things at the same time. I can't even think of the next thing I want to do if I have this, you know, huge coding project or whatever. So I need to find some way to fix that. Uh so we can I mean we can tell the agents uh you know go through my codebase, find out what the next things I need to do are, create these tasks for me and I just click a button, I accept the quest and there's some agent that does it for me which is fun. Uh it's cool. Uh but the next problem here is that okay now I have 20 stuff going at the same time and I have to babysit each and every one and that is completely exhausting. So you could ask okay we have agents agents are amazing as we've established so I can just go ahead and tell the agent to do it for me now it's called software factories but at the time it was just doing stuff autonomously uh so you can give it like a general sense of what you want it would spawn an orchestrator and break it down for you and run it all in well in this case it's a local uh container so everything is just done in isolation and I don't need to do any babysitting the agent does it for me. On
所以就跟电子游戏、RTS（即时战略游戏）或者 Civilization（文明）之类的一样，你按一下空格键，它就会直接把你带到最近一个真正需要你处理的地方。这样你就能很轻松地掌握全局，从一件事跳到另一件事，回答问题、批准计划等等。这挺好的，光是有了这种可见性，就已经是不错的一步了，但还不够。一直盯着 agent 在干什么，还细到它们此刻在用哪些文件、调哪些工具，这相当累人。所以我们需要别的东西。原因是这里有两个主要问题。第一，要装在脑子里的东西实在太多了。就算我能搞懂发生了什么、什么都看得见，我也没法同时干 20 件事。要是手上有个巨大的编程项目之类的，我连下一步想做什么都想不过来。所以我得想办法解决这个问题。我们可以让 agent 去过一遍我的代码库，找出我接下来要做哪些事，帮我建成任务；我只要点个按钮，接下任务（quest），就会有某个 agent 替我去做，挺好玩的，也挺酷的。但接下来的问题是：好，现在我同时有 20 件事在跑，每一件都得我像保姆一样盯着，这简直累死人。那你可能会问：好，我们有 agent，前面也说了 agent 很厉害，那我直接让 agent 替我干不就行了？这东西现在叫软件工厂，但在当时就只是让它自主干活。你只要大致告诉它你想要什么，它就会拉起一个编排器，帮你把任务拆解好，然后全部放在——嗯，这个例子里是一个本地容器——里面跑，所以一切都是隔离着完成的，我完全不用当保姆，agent 替我搞定。在此
[7:10] Ido
top of that, you have loops, which everyone really likes right now. So, you can do cool things like just tell it to, I don't know, scan Twitter for cool stuff you want to build or scan your GitHub, and it will just create these nice uh features for you in the background without you needing to do too much up front. The bad thing is that I mean this is nice uh but what do I do when I have again I don't know I would say 20 but even five things to review at the same time. Uh it's very very hard to get that out of the way. So what I did is create uh this like review kit. So I get the entire changes I need and I get uh what files changed if you want to go like file by file or you can get just get visual evidence. So you can actually see videos of what changed or photos of what changed so you can really quickly understand what you want or even run a few uh instances in parallel and just pick the best implementation out of these. So now that we have this we have uh visibility and we have autonomy uh it's pretty cool like we reached a pretty good place in raising the ceiling of what we can do with agents but it's still pretty exhausting. I mean it's it was fun. Uh, but now I have to ask myself, do I really need to be the only one doing this? It's pretty lonely to just talk to agents all day and trying to and you know, if you look at uh what a team does, for example, if it's just me, that's fine. But if I have a team, I might not even be the right person to review what's going on. So, I need something else. Uh, and that something is collaboration. We can just have other people join us. So the way that looks in agent craft is that I can create these uh you can call it you know war halls war room whatever uh host them it's like completely local but you can do like tunnels and people can join you h so I can create a room and let's say I have a product designer in my team in this case uh it's my wife and I can approve uh her
之上，还有循环（loop），这是现在大家都特别喜欢的东西。你可以做些很酷的事，比如让它去——我也不知道——刷刷 Twitter 找你想做的酷东西，或者扫一遍你的 GitHub，然后它就会在后台帮你把这些功能做出来，你前期不用投入太多。麻烦的是——这当然很好——但当我同时有，我也说不好，就说 20 件吧，哪怕只有 5 件东西要评审，我该怎么办？要把这些都消化掉，真的非常非常难。所以我做了这么一个评审工具包（review kit）。我能拿到需要的全部改动，能看到哪些文件变了——如果你想逐个文件过的话；或者你也可以直接看可视化的证据：你能直接看到改动的视频或者截图，很快判断这是不是你想要的；甚至可以并行跑几个实例，从中挑出最好的那个实现。所以现在我们有了可见性，也有了自主性（autonomy），挺酷的——在抬高用 agent 能做成事的天花板上，我们已经走到了一个相当不错的位置，但它仍然挺累人的。我是说，确实挺好玩的。但现在我得问问自己：真的只能由我一个人来干这件事吗？整天只跟 agent 说话，挺孤独的。而且你看看一个团队是怎么运作的——比如说，如果只有我一个人，那没问题；但如果我有个团队，我可能根本就不是评审这些东西的合适人选。所以我需要别的东西，那就是协作（collaboration）。我们可以直接让其他人加入进来。在 AgentCraft 里它是这样的：我可以创建这种——你可以叫它作战大厅、作战室（war room），随便怎么叫——然后自己托管。它完全是本地的，但你可以打隧道（tunnel），这样别人就能加入你。所以我可以建一个房间，假设我团队里有一位产品设计师——在这个例子里就是我太太——我可以批准她进来——
[9:16] Ido
otherwise she'll be mad uh so I can approve her and she does her thing she designs whatever and I as an engineer can not only see that that's that's her um I can not only see that on my map uh I can actually take that and follow off off of it
不然她会生气的。所以我批准她进来，她就去做她的事，设计点什么都行；而我作为工程师，不仅能看到——那个就是她——我不仅能在自己的地图上看到她做的东西，还能直接拿过来接着往下做。
[9:35] Ido
prompt prompt
Prompt。Prompt。
[9:36] Ido
yeah that's just like her um so I can take that design follow up on it implement it while she keeps working uh and the cool thing here is that not only can I hand off or fork and kind of have this shared workspace so I'm not really bound anymore or to Git. I can use Git, but I'm not really bound to something. Everyone is working together. Uh you can see that there's also some kind of chat between not only between the people within the room, but also between agents. So we can have this really nice notice board of exactly what each person is doing, what each agent doing. They're all aware of each other and they can collaborate much more closely uh than you normally would be able to do. Uh so for example, here I can say, okay, I took off this design task. Uh so now we have oh by the way this is also from devices you can have like mobile and and if you like telegram you can use telegram whatever floats your boat. So we have visibility and we have autonomy and we have collaboration so we managed to raise the ceiling. Uh it's now possible especially if you're power users to do really cool stuff and use more agents and do more things. But the interesting thing here was, you know, if I looked at the feedback I got, it's not just that power users were able to do more. The cooler thing for me was that people that had even nothing to do with agents or nothing to do with productivity, uh, come and say, uh, okay, my kids really loved it and they use it all the time to to orchestrate agents, which they didn't do until now.
对，这就很像她。所以我可以拿过这个设计，接着往下推进、把它实现出来，同时她还能继续干她的活。这里很酷的一点是，我不仅可以把工作交接出去，或者 fork 一份，还等于有了一个共享的工作空间，所以我不再被绑死在 Git 上。我可以用 Git，但不会被什么东西绑住，大家是一起干活的。你还能看到这里有一种聊天，不只房间里的人之间能聊，agent 之间也能聊。所以我们会有一块很好用的公告板，清楚写着每个人在做什么、每个 agent 在做什么。大家都清楚彼此在干嘛，协作起来比平常紧密得多。比如在这里，我可以说：好，这个设计任务我接手了。所以现在我们有了——哦对了，顺便说一句，这也能从各种设备上用，你可以用手机；如果你喜欢 Telegram，也可以用 Telegram，你爱用啥就用啥。所以我们有了可见性、有了自主性，还有了协作，天花板算是被我们抬高了。现在，尤其如果你是重度用户，已经可以做很酷的事、用更多的 agent、干更多的活了。但有意思的是，看看我收到的反馈，不光是重度用户能做更多事了。对我来说更酷的是，一些跟 agent、跟生产力压根不沾边的人跑来跟我说：我家孩子特别喜欢它，一天到晚都在用它编排 agent——这是他们以前从来没做过的。
[11:10] Ido
uh or you know someone that flunked out of college because he played Starcraft and now you can suddenly do really productive stuff with it. Uh and a bunch of other examples of kind of bringing communities that so far were kind of behind into this new world of productivity. So we need to raise the ceiling but we also need to uh lower the floor. We need to have some kind of way to take no probably around 90% of the people in the world and bring them into this hygienic future without having them burn out. So again, I took it too far. Um, and there's this uh currently named TBD um potentially loopers. Uh the idea here is okay why don't why do we need something like you know Warcraft which is potentially very complicated uh especially or might be even daunting if you're not into the granularity of what files are there are or how uh complex the UI is and kind of try to find a lower denominator of what makes stuff accessible especially now that agents are much more capable so instead of granularity or having like exactly what files they're working on maybe I can have it more of a mobile game level uh have projects that they keep returning to and use loops and other stuff that rely on agents being more autonomous to have like a very basic level of I prompt I get the result I don't really care perhaps about the way I can see it if I want to and I have this really nice interactive and cool experience that makes me want to come back to this and obviously I mean this is one visualization uh it can really resonate with some people but might not might not present to others because we're not all the same and what may appear like amazing and easy to one person can appear very difficult to another. Uh but here uh I'm trying to look at a way to make this more customizable. Try and understand what it is that you as a person need in order to understand and want to use these tools. So this is a very very early draft. Um but if you
又或者，有人因为打 StarCraft（星际争霸）大学挂科退了学，现在突然能用它干点真正有生产力的事了。还有一堆别的例子，都是把那些到目前为止有点被落在后面的群体，带进这个新的生产力世界。所以我们既需要抬高天花板，也需要降低门槛（地板）。我们得找到某种办法，把——不，大概是世界上 90% 左右的人——带进这个 agentic 的未来，又不让他们累垮。所以，我又一次玩过头了。于是有了这个东西，目前名字还是 TBD（待定），可能会叫 loopers。这里的想法是：我们为什么非得要一个像 Warcraft 那样的东西呢？它可能非常复杂，甚至让人望而生畏——尤其如果你并不关心有哪些文件这种颗粒度，或者 UI 有多复杂的话。所以就想试着找一个更低的公分母，看看到底是什么让东西变得容易上手，尤其是现在 agent 已经强大多了。与其追求颗粒度、非得精确看到它们在改哪些文件，也许我可以把它做成更像手游那种层级：有一些让人一次次回来照看的项目，用上循环和其他依赖 agent 更加自主的机制，做到一种最基础的层级——我发 prompt，我拿结果，过程我可能不太在乎，想看的话也能看；而且还有一种很好玩、互动性强又很酷的体验，让我想一次次回来用它。当然，这只是其中一种可视化方式，它可能会让一部分人很有共鸣，但对另一些人可能就不对味，因为我们每个人都不一样，对一个人来说很棒、很简单的东西，在另一个人眼里可能非常难。不过在这里，我在尝试找一种办法，让它更加可定制，试着去理解：你作为一个具体的人，需要什么才能看懂这些工具、并且愿意去用它们。所以这还是一个非常非常早期的草稿。不过如果你
[13:22] Ido
want to use it, please let me know. So we're the bottleneck. I think we've established that uh pretty well until now. But the skills that we need to stop being the bottleneck and enable this new future are already with us. So we are the bottleneck, but we don't have to be. And I hope that you all can take what's relevant to you and your life and apply it forward. So I would ask you for two things. One, agent is already out there. Like you can just npx install agent as a website uh and do your thing. Uh this new thing that I just showed you uh potentially loopers uh is still experimental. I'm still trying to understand the interest and what people are actually looking for to uh get the floor that they want or the ceiling that they want. So, please uh go here and express your interest and tell me what you're looking for so I can know what to do with this. And thank you. I got it.
想用的话，请告诉我。所以，我们就是瓶颈——我想到目前为止，这一点已经讲得挺清楚了。但让我们不再当瓶颈、开启这个新未来所需要的技能，其实早就在我们身上了。所以，我们是瓶颈，但我们不必是。希望大家能把其中跟自己、跟自己生活相关的部分带走，用到往后去。所以我想请大家做两件事。第一，AgentCraft 已经放出来了，你可以直接用 npx 安装 AgentCraft，也有网站，然后就可以自己玩起来了。至于我刚才给你们看的这个新东西——可能会叫 loopers——还在实验阶段。我还在摸索大家的兴趣在哪、真正想要的是什么，好让大家得到自己想要的那个门槛，或者想要的那个天花板。所以请大家到这里来登记一下兴趣，告诉我你想要什么，这样我才知道这东西接下来该怎么做。谢谢大家。好，收到。