Practical AI · 379
Abundant Intelligence at the Midwest AI Summit Part 2
丰饶智能：中西部 AI 峰会实录（下）
Midwest AI Summit 三场演讲实录：International Mission Board 用一份 AI 章程界定组织「会用／不会用」AI 的边界；Noblis 拆解组织级 AI 治理的四个关键决策，从授权、问责到技术债；AeroCore Technologies 展示如何用 AI agent、text-to-SQL 与视觉模型打通数据访问、自动化工单核验。
00:0000:47:53
00:01
Welcome to the Practical AI Podcast, where we break down the real-world applications of artificial intelligence
欢迎来到Practical AI Podcast，我们在这里拆解人工智能在现实世界中的应用，
00:08
and how it's shaping the way we live, work, and create.
以及它如何塑造我们的生活、工作和创作方式。
00:12
Our goal is to help make AI technology, practical, productive, and accessible to everyone.
我们的目标是帮助AI技术变得实用、高效，并且让每个人都能接触到。
00:18
Whether you're a developer, business leader, or just curious about the tech behind the buzz,
无论你是开发者、企业领导者，还是只是对热潮背后的技术感到好奇，
00:22
you're in the right place.
你来对地方了。
00:24
Be sure to connect with us on LinkedIn, X or Blue Sky
记得在LinkedIn、X或Blue Sky上关注我们，
00:27
to stay up-to-date with episode drops, behind-the-scenes content, and AI insights.
以便及时获取节目更新、幕后内容和AI洞察。
00:32
You can learn more at practicalai.fm. Now, on to the show.
你可以在practicalai.fm了解更多。现在，进入节目。
00:41
Welcome to another episode, the Practical AI Podcast. I'm Chris Benson, your co-host.
欢迎收听另一期Practical AI Podcast。我是Chris Benson，你的联合主持人。
00:46
And this week we're going to bring you three more of our speakers from the Midwest AI Summit.
这周我们将为你带来另外三位来自中西部AI峰会的演讲嘉宾。
00:53
They are three speakers who are dealing with AI issues on an everyday basis
他们是三位日常处理AI问题的演讲者，
00:58
that many of us are coping with. Those three speakers are Don Barger,
这些问题也是我们许多人正在应对的。这三位演讲者是来自国际宣教委员会的Don Barger、
01:03
from the International Mission Board, Mark Lay, from Nobleus, and Darren Taley,
来自Nobleus的Mark Lay，以及来自Aerocore Technologies的Darren Taley。
01:09
from Aerocore Technologies. Next week I will be back in the usual format with a guest speaker,
下周我将以常规形式回归，并邀请一位嘉宾演讲者，
01:15
but in the meantime, I hope you enjoy these three. Thank you very much.
但在此期间，希望你们喜欢这三位。非常感谢。
01:19
I want to introduce our first speaker, Don Barger.
我想介绍我们的第一位演讲者，Don Barger。
01:25
I've known Donny for a long time, and he's taken kind of the whole journey of prediction guard with us.
我认识Donny很久了，他基本上跟我们一起走过了prediction guard的整个历程。
01:35
And I'm so impressed with how he has stewarded the International Mission Board forward
我对他在推动International Mission Board向前发展方面所展现的领导力深感钦佩
01:42
in terms of their AI journey and adoption.
就他们的AI发展历程和采用情况而言。
01:45
He's the Director of Innovation and Artificial Intelligence at the International Mission Board.
他是International Mission Board的创新与人工智能总监。
01:50
He spent a lot of years in Latin America and different parts of the world
他在拉丁美洲和世界不同地方待了很多年
01:55
with really deploying strategies rooted in morality and storytelling.
真正部署植根于道德和叙事的策略。
02:00
He's going to be talking about building AI trust, what we said, yes, too, and what we didn't.
他会谈到建立AI信任，我们说了什么，是的，也是，以及我们没说什么。
02:05
So welcome, Don Barger, from the IMB.
欢迎来自IMB的Don Barger。
02:07
Appreciate it.
谢谢。
02:09
Yes, I've known Don for a while. The first time I met Don was at a much smaller meeting than this
是的，我认识Don有一段时间了。我第一次见到Don是在一个比这次小得多的会议上，
02:14
that he was actually leading, again, another meeting he was leading about AI.
实际上是他主持的，又是另一个他主持的关于AI的会议。
02:19
And at that meeting, I was the dumbest person at that meeting.
在那个会议上，我是那个会议上最笨的人。
02:24
It was hands down, like I walked in, I'm like, I don't think I knew what I was coming to.
毫无疑问，就像我走进去，我就想，我觉得我根本不知道我来的是什么地方。
02:28
And I could say that if it hadn't been for Dan at that meeting,
我可以说，如果不是因为Dan在那个会议上，
02:32
and then there's some other meetings that we met at,
然后还有我们相遇的其他一些会议，
02:34
and we kind of sketched out some stuff in a hotel in London in a lobby.
我们在伦敦一家酒店的大堂里大致勾勒了一些东西。
02:38
And next thing you know, we're really doing a lot of crazy stuff.
然后你猜怎么着，我们真的在做很多疯狂的事情。
02:41
So every organization that's building AI eventually runs into questions about
所以每个在做AI的组织最终都会遇到一些问题，关于
02:48
should we be doing this?
我们应该做这个吗？
02:50
And so that's what I want to share with you today, kind of the journey that I've been on,
所以这就是我今天想和你们分享的，就是我走过的这段旅程，
02:54
who I am, the oddness that I am, the one that's speaking to here,
我是谁，我这个奇怪的人，在这里演讲的这个人，
02:58
I'm not your typical AI presenter.
我不是你们典型的那种AI演讲者。
03:01
Normally you would have somebody that has a degree in computer science or data science or physics.
通常你们会看到有计算机科学或数据科学或物理学学位的人。
03:07
And I will tell you that I have an undergraduate degree in accounting and two degrees in theology.
我告诉你们，我有一个会计学的本科学位和两个神学学位。
03:13
So if that surprises you, I think that it should encourage you
所以如果这让你感到意外，我觉得这反而应该让你受到鼓舞
03:17
because the reality is, for the first time in a long time,
因为现实是，很久以来第一次，
03:20
maybe the first time ever, normal people can build incredible things.
也许这是有史以来第一次，普通人也能做出了不起的东西。
03:25
And I've been so excited to see how our organization has embraced using AI in ways that I was,
我一直特别兴奋地看到我们组织是如何拥抱使用AI的，以一些我原本……
03:33
wouldn't even have dreamed we could do three years ago.
甚至三年前我们连做梦都不敢想能做到的事。
03:37
So the problem is what's responsible.
所以问题在于责任归属。
03:40
And so the question we need to ask a lot of people is, is this the right thing to do?
所以我们需要问很多人的问题是，这样做对吗？
03:44
The responsibility question is the one that builds trust.
责任归属这个问题才是建立信任的关键。
03:47
It's the question in an organization that's 181 years old.
这是一家有着181年历史的组织里正在面对的问题。
03:50
If you show that you're building trust and you're trying to build trust,
如果你展现出你正在建立信任、并且努力去建立信任，
03:54
there are a lot more open to changing things.
那就会有更多人愿意改变。
03:57
So almost four years ago, I'll tell you how we got started,
所以差不多四年前，我跟你讲讲我们是怎么开始的，
04:00
almost four years ago, before I sat down with Dan in London,
差不多四年前，在我和Dan在伦敦坐下来谈之前，
04:04
I had another group of people in London that I was meeting with.
我在伦敦还见了另一群人。
04:07
And we were looking at what we believed about AI.
我们当时在审视我们对AI的认知。
04:10
One of the guys on my team has just finished his PhD in digital ethics and computer ethics.
我团队里有一个人刚读完数字伦理和计算机伦理的博士。
04:16
And AI was not really as popular of a topic then,
那时候AI还不是一个特别热门的话题，
04:19
but we sat down and said, let's talk about what we believe about AI.
但我们坐下来聊，说我们来谈谈我们对AI的信念是什么。
04:23
And what will we do and what will we not do?
我们会做什么，不会做什么？
04:25
Before we ever write one line of code, let's come up with a charter about what we believe AI is.
在我们写下第一行代码之前，先制定一份关于我们对AI信念的章程。
04:31
And we did that.
我们就是这么做的。
04:33
We spent a couple of days with eight people, and we hammered out this agreement
我们八个人花了好几天时间，反复讨论敲定了这份协议，
04:39
to what we would do with AI.
关于我们会用AI做什么。
04:41
And I can look back to that time as something that's impacted everything
现在回头看那段时间，它影响了后来的一切。
04:46
that our team has done since that time.
这是我们团队自那时以来所做的工作。
04:49
Because it allows us to know what is possible and what we're going to say no to
因为它让我们知道什么是可能的，以及我们要拒绝什么，
04:54
without even having to think about it.
甚至不需要思考就能决定。
04:56
And there's been multiple times that we've come across issues that people have asked us to do
有很多次我们遇到人们要求我们做的事情，
05:01
and we've been able to say no, we don't think that's what AI should be doing.
我们能够说不，我们认为那不是AI应该做的事。
05:05
And the conversation that most organizations skip is just because we can do something
而大多数组织跳过的对话就是，仅仅因为我们能做某件事，
05:10
doesn't mean we should do something.
并不意味着我们应该做某件事。
05:12
And so we look at really, really hard what we believe about AI.
所以我们非常非常认真地审视我们对AI的信念。
05:18
And let me just give you some examples of things that we said we would do and what we wouldn't do.
让我给你们举一些例子，说明我们说过会做什么、不会做什么。
05:23
The easy things would be drafting and editing, summarizing, and research.
简单的事情包括起草和编辑、总结以及研究。
05:28
These are things that all of us do.
这些都是我们所有人都会做的事。
05:29
Translation, we use a lot of translation tools.
翻译，我们用很多翻译工具。
05:32
We do a lot of really good work with translation, saving a lot of money in translation,
我们在翻译方面做了很多非常好的工作，节省了大量翻译成本，
05:37
training materials, and learning support, and content.
培训材料、学习支持和内容。
05:41
We create code and we always have humans in the loop, we write code.
我们写代码，而且始终有human in the loop，我们写代码。
05:45
The timing to ship new products has gone from years with lots of contractors
新产品交付的时间从过去需要几年、大量承包商
05:52
that don't really understand what we need to be able to ship it ourselves
那些并不真正理解我们需要什么才能自己把它交付出去
05:56
in prototype in a day and ship within months really good product.
在一天内做出原型，几个月内交付真正好的产品。
06:01
So the things that we said no to, for us we're really important, was AI presenting itself as a human in ministry.
所以我们说不的事情，对我们来说真正重要的，是AI在事工中把自己呈现为人类。
06:08
But I would just say anytime somebody's expecting to be thinking they're talking to a human,
但我想说的是，任何时候有人以为自己在跟人类对话，
06:13
we don't ever want to misrepresent AI as speaking for humans.
我们永远不想把AI错误地呈现为在替人类说话。
06:18
Final decisions where people, about people that would be made solely by AI,
关于人的最终决定完全由AI来做，
06:24
we don't ever want to see that, monitoring people without their knowledge.
我们永远不想看到这种情况，在人们不知情的情况下监控他们。
06:27
That's happening all around us.
这在我们周围到处都在发生。
06:29
We don't think that's something we should be involved in.
我们不认为那是我们应该参与的事情。
06:31
Autonomous actions and high risk and high stakes context, we have people serving all over the world
自主行动以及高风险、高利害的场景，我们在世界各地都有人在服务，
06:37
and we want to make sure that we're not giving AI the autonomy to make decisions about that.
我们希望确保不会把对此做决定的自主权交给AI。
06:43
But then finally, probably the one that we've said no to the most,
但最后，可能我们拒绝得最多的一件事，
06:47
is handing what I would consider sole care over to an algorithm.
就是把我认为的完全照护权交给一个算法。
06:52
And from us, from a Christian standpoint, I don't want AI to be taking on pastoral responsibilities.
从我们的角度、从基督教立场来看，我不希望AI承担牧养责任。
06:59
I don't want it to do counseling because I really think that people need to be interacting with other people.
我不希望它做心理咨询，因为我真的认为人们需要与他人互动。
07:04
We use AI a lot to help screen things and to help encourage people.
我们大量使用AI来帮助筛选内容，也帮助鼓励人。
07:09
But when it comes down to counseling people and sharing with people,
但当涉及到给人做心理咨询和与人分享时，
07:13
we really want that to be a human.
我们真的希望那是人类来做。
07:15
I like to see AI is actually freeing me up to do more of that, not replacing me in doing that.
我喜欢看到AI实际上是在解放我去做更多这样的事，而不是取代我去做这些事。
07:22
So why I lean so hard into this is that we drew the boundaries before we wrote the code.
所以我如此坚定地投入这件事的原因是，我们在写代码之前就先划定了边界。
07:28
And I think that if you've not thought through what you believe about AI, the ontology of AI,
而且我认为，如果你还没有想清楚你对AI的信念、AI的本体论、
07:33
what you would believe about how you treat AI and what you get let it do,
你对如何对待AI的信念以及你允许它做什么，
07:37
I challenge you to stop writing code, stop building,
我挑战你停下写代码，停下构建，
07:41
and at least spend some time with some people and think about what you're going to do and what you believe about AI.
至少花些时间和一些人在一起，想想你打算做什么，以及你对AI的信念是什么。
07:48
I don't think a lot of people are doing that.
我不觉得有很多人在做这件事。
07:50
I don't think Sam Altman spends much time thinking about this question of what we should be doing with AI.
我不觉得Sam Altman会花很多时间去想我们应该拿AI做什么这个问题。
07:57
I don't think Elon Musk thinks about that.
我不觉得Elon Musk会想这个。
07:59
I think we need to be leading out in helping do that.
我觉得我们需要带头帮忙做这件事。
08:03
The sole care question, like the pastoral care, the prayer, the counseling, the discipleship,
那些专属的关怀工作，比如牧养关怀、祷告、辅导、门徒训练，
08:08
deep personal relationships, crises, dealing with people that are dealing with a crisis,
深度的个人关系、危机、跟正在经历危机的人打交道，
08:14
we want people to be dealing with those.
我们希望是由人来处理这些。
08:16
AI can prepare the support, but it can't really replace the person.
AI可以预备支持，但它真的没法取代那个人。
08:20
So we had an ask that come to me and somebody, one of the teams asked me,
所以有人来找我和另一个人，其中一个团队问我，
08:25
hey, we have this urgent need.
嘿，我们有个紧急需求。
08:27
We have people who really need counseling, but we don't have counselors.
我们有些人真的需要心理咨询，但我们没有咨询师。
08:32
There's no way to get them counseling.
没办法给他们提供心理咨询。
08:34
Can you help us build an AI chatbot that would basically become their counselor?
你能帮我们做一个AI chatbot，基本上就是充当他们的咨询师吗？
08:39
I didn't have to think long and hard about it because we had already made that decision before.
我没怎么多想，因为我们之前就已经做过那个决定了。
08:44
No, we can't help you with that.
不行，我们没法帮你做那个。
08:45
Let me share with you the charter, our AI charter, about what we believe about AI.
让我跟你分享一下我们的章程，我们的AI章程，关于我们对AI的信念。
08:50
I shared that with them and they still push back and said, you don't understand.
我把这个跟他们说了，他们还是不认同，说你不理解。
08:53
The need is so great, something is better than nothing.
需求太大了，有总比没有好。
08:56
My response was, something can be worse than nothing if you're doing the wrong things.
我的回应是，如果你在做错误的事情，有可能会比没有更糟。
09:01
We said no, disagreed.
我们说了不，不同意。
09:04
They may still be doing that, but I'm not helping them pull that off.
他们可能还在那么做，但我不会帮他们做成那件事。
09:08
I would just challenge you to think through what it is that you're going to say yes and no to as you're doing that.
我只是想挑战你，在做这件事的时候，想清楚你要对什么说是、对什么说不。
09:14
The no is easy because the no asked long before that person had asked me to help them.
说不是什么难事，因为早在那个人找我帮忙之前，不就已经说出口了。
09:21
I'll just confess to you and Dan knows me.
我跟你坦白，Dan了解我。
09:24
I am famous for hopping on an airplane and building a solution from one place to another.
我以坐飞机到处跑、在一个地方到另一个地方搭建解决方案而闻名。
09:28
The idea of building a chatbot as a counselor was really intriguing to me.
把聊天机器人做成心理咨询师的想法让我觉得很有意思。
09:33
It would be easy for me to go home that night while we're watching Netflix and prototype it.
那天晚上我们看Netflix的时候，我回家就能轻松把它做个原型出来。
09:38
The next day, send in and say hey, we built it.
第二天交上去说，嘿，我们做出来了。
09:41
What do you think?
你觉得怎么样？
09:42
Because we had talked about this beforehand, I was able to say no, this is not something we're going to do.
因为之前我们已经讨论过这件事，所以我能够说不，这不是我们要做的事。
09:48
Second case study.
第二个案例研究。
09:49
One of the surprising tools that we built last year has been out for about a year and a half.
我们去年构建的其中一个令人惊讶的工具已经推出一段时间了，大约一年半。
09:54
It's got around 250.2.5 million users in about a year and a half.
它在大概一年半的时间里拥有了约2.502亿用户。
09:58
We've not had any paid advertising for it.
我们没有为它做过任何付费广告。
10:01
It's just word of mouth.
全靠口碑传播。
10:02
It's something called Faithbot.
它叫Faithbot。
10:04
You can find it at faithbot.tools.
你可以在faithbot.tools找到它。
10:06
We have several tools that we've built for the church on this website.
我们在这个网站上为教会做了好几个工具。
10:11
Faithbot.tools.
Faithbot.tools。
10:12
The Faithbot was my attempt to create a chatbot that would always be grounded in faith.
Faithbot是我尝试创建一个始终扎根于信仰的chatbot。
10:20
It's going to answer your questions from a faith perspective.
它会从信仰的角度回答你的问题。
10:23
It's going to be grounded in things that are important to Christian faith.
它会扎根于对基督教信仰重要的东西。
10:27
We built this chatbot and it's just taken off.
我们构建了这个聊天机器人，然后就火起来了。
10:30
It's amazing to see the growth.
看到增长真是令人惊叹。
10:32
It's in 50 languages.
它支持50种语言。
10:34
We've had a couple of versions of it before.
我们之前有过几个版本。
10:37
The third version I built on a flight from Narita.
第三个版本是我在从成田起飞的航班上构建的。
10:40
No, from Denver to Narita.
不，是从丹佛到成田。
10:42
I got on the plane thinking this is what I'm going to build.
我上了飞机，想着这就是我要做的东西。
10:44
I'm not going to sleep.
我不打算睡觉。
10:45
I landed in Narita and before I landed, it was live.
我降落在成田机场，在我降落之前，它就已经上线了。
10:48
It was pushed to the droplet, live, published, and people were swapped over from the old Faithbot to the new Faithbot.
它被推送到droplet上，上线了，发布了，人们从旧的Faithbot切换到了新的Faithbot。
10:55
The crazy thing is, this thing has the ability to do so much.
疯狂的是，这个东西有能力做那么多事。
11:00
We had to make some decisions early on and one of the examples would be,
我们早期不得不做一些决定，其中一个例子就是，
11:05
I don't want Faithbot to pray for people.
我不想让Faithbot为人们祈祷。
11:08
The first time I was playing with it on the plane, the new version that I built,
我第一次在飞机上玩它的时候，我构建的新版本，
11:12
and by the way, there's 51 languages in each of them use a different LLM.
顺便说一下，每个里面都有51种语言，每一种都使用不同的LLM。
11:16
They all have different mark downs.
它们都有不同的标记方式。
11:18
They all are built uniquely with different worldviews, so it's not just one thing that just translates.
它们都是独特构建的，有着不同的世界观，所以它不只是一个简单翻译的东西。
11:23
It's like 51 tools built into one tool, and it knows where you're entering from,
它就像是51个工具集成在一个工具里，而且它知道你从哪个入口进来，
11:27
and it automatically loads in that language and worldview.
然后它会自动加载对应的语言和世界观。
11:30
So it's really cool.
所以真的很酷。
11:31
But I was playing with it, and it offered to pray for me.
但我在玩它的时候，它主动提出为我祈祷。
11:34
And I thought, well, that's cool, but then I realized that's kind of creepy.
我当时觉得，嗯，挺酷的，但后来意识到这有点瘆人。
11:38
Like, AI can't pray.
比如，AI又不能祈祷。
11:39
Why am I letting AI pray?
我为什么要让AI祈祷呢？
11:41
And so then I spent the next hour trying to unprogram the ability for it to pray.
于是接下来一个小时我都在试图取消它祈祷的功能。
11:46
And making it work couldn't do that.
结果怎么弄都做不到。
11:48
So AI is not really praying.
所以AI并不是真的在祈祷。
11:49
It's just kind of mimicking the training data that we gave it.
它只是在模仿我们给它的训练数据而已。
11:52
The decision point number two was we also don't want people to just to find all of their needs in an algorithm.
第二个决策点是，我们也不希望人们把所有需求都寄托在一个算法上。
11:58
So we want to point people to people.
所以我们想把人们引向人。
12:00
And so we intentionally, when you use it, it will try to connect you to someone else,
所以我们会刻意设计成，当你使用它的时候，它会尝试把你连接到其他人，
12:07
particularly if you're hitting it with a crisis.
尤其是当你带着危机情况来找它的时候。
12:10
So one of the things we learned early on is a lot of people who use this
所以我们早期学到的一件事是，很多使用这个的人
12:14
for people who are struggling, people who are being abused, people who are self-harm thinking about killing themselves.
是那些正在挣扎的人、正在被虐待的人、有自残倾向想自杀的人。
12:20
And it really was shocking to me to read some of the, because they were flagged
读到其中一些内容真的让我很震惊，因为它们被标记了
12:25
and I would go back and make sure the AI was responding well.
我会回去确认AI的回应是否恰当。
12:28
I don't know who they are, but I get the data to make sure the AI is responding well.
我不知道他们是谁，但我会拿到数据来确保AI回应得当。
12:32
So we built it so that when that happens, it's always going to recommend trying to get you in touch with somebody else
所以我们把它设计成，当这种情况发生时，它总是会建议尝试让你联系到其他人
12:39
and it's going to try to connect you to a help.
它会尝试帮你联系求助渠道。
12:41
And if you don't know if somebody, it can recommend services and crisis centers that it can point people to.
如果你不知道找谁，它可以推荐相关服务和危机中心，为人们指明方向。
12:48
I'm also notified of anything that is critical and I can hop into the conversation and take over for the AI.
任何危急情况我也会收到通知，我可以随时介入对话，接管AI的工作。
12:56
That was very possible when we had 5,000 users.
在我们只有5000个用户的时候，这完全可行。
12:59
It is not possible when you have two and a half million users.
但当你有250万用户的时候，这就不可行了。
13:02
So we do a lot of checking.
所以我们做了大量的检查。
13:04
The other things are lines that it won't cross.
另外还有一些它不会跨越的底线。
13:07
So I would say this is for everything we build.
我想说，这对我们构建的所有东西都适用。
13:10
And we let it know that there are certain things that you can't do.
我们让它知道有些事情是你不能做的。
13:14
So we don't want this to ever be hateful or coercive or abusive.
所以我们不希望它变得仇恨、胁迫或滥用。
13:19
So we want to build our tools in ways that are edifying and helping people.
所以我们想以建设性和帮助人们的方式来构建我们的工具。
13:23
It doesn't also replace, it's not going to give you a medical diagnosis.
它也不会替代，它不会给你医疗诊断。
13:28
I'm shocked to the number of people who go there and ask it from financial advice from certain parts of the world.
我很震惊有多少人来自世界某些地方去那里问它财务建议。
13:32
And so we don't give financial advice.
所以我们不提供财务建议。
13:35
We just programmed it to say, stick to what you're good at.
我们只是编程让它说，专注于你擅长的。
13:38
Don't try to do anything you can't or you shouldn't.
不要尝试做任何你不能或不应该做的事情。
13:42
I hate I have to say this, but if your pastor is going to AI to help you write a sermon,
我讨厌我不得不说这个，但如果你牧师要用AI帮你写布道，
13:47
you might want to find a better church.
你也许该找个更好的教会。
13:49
But there's a lot of people as I visit with people and I teach at and in universities,
但有很多人，当我和人们交流、在大学里教书的时候，
13:54
I find that this is more common than I was realizing.
我发现这比我意识到的更常见。
13:58
So it'll actually explicitly not do that.
所以它实际上会明确不这么做。
14:00
It'll actually kind of lecture you on how you shouldn't be doing that.
它其实会有点像在说教你不要那么做。
14:04
And then finally, it's not for sale.
然后最后，它是非卖品。
14:06
So we're not trying to make money off of this.
所以我们并不是想靠这个赚钱。
14:08
We don't charge for these products.
我们不对这些产品收费。
14:10
We do this through people who love it.
我们是通过热爱它的人来做这件事的。
14:12
They like it and they say, hey, I'd like to sponsor it.
他们喜欢它，然后说，嘿，我想赞助它。
14:14
That's what we've been able to do.
这就是我们能够做到的。
14:16
It's also, I mean, we had to keep rebuilding to make it more and more economical.
而且，我的意思是，我们不得不不断重建，让它越来越经济。
14:19
We're careful with the data.
我们对数据很谨慎。
14:21
Originally, I didn't even, I didn't archive anything.
最初，我甚至没有，我什么都没有存档。
14:24
Like it was just, it was gone once you log out.
就像它只是，你一旦登出，它就消失了。
14:27
But a lot of people want to go back because they're accustomed to chat GPT
但很多人想回去，因为他们习惯了ChatGPT，
14:31
and they want to go back and they pick up their old chats.
他们想回去继续之前的对话。
14:34
So we built it out where you can sign in if you want to,
所以我们做了这样的设计，你想登录就登录，
14:37
but you don't have to sign in.
但不是必须登录。
14:38
You can always be anonymous.
你可以一直保持匿名。
14:40
And you just keep that anonymous by choice.
你选择匿名就可以一直匿名。
14:42
So last few things here, I would say that you should evaluate against trust,
最后几点，我想说的是，你应该基于信任来评估，
14:47
not just feasibility.
而不仅仅是可行性。
14:49
So when you're choosing what you're going to do,
所以当你在选择要做什么的时候，
14:51
you're going to evaluate how is this going to be received by the people you're working with.
你会去评估这件事会被和你共事的人怎么接受。
14:55
So a lot of how we've been able to build trust in a very conservative,
所以我们能在这样一个非常保守、
14:59
very old, very risk and change in virtual organization
非常老旧、非常厌恶风险和变化的组织里建立起信任，很大程度上是因为
15:02
is we've proven time and time again to make really good trust worthy decisions.
我们一次又一次地证明了我们能做出非常靠谱、值得信赖的决策。
15:08
And so we've worked hard at building trust within our organization.
所以我们一直在努力在组织内部建立信任。
15:12
So whether or not that's you working within your own organization
所以不管是你自己在自己的组织内部做事，
15:15
or if you're selling AI products to other people,
还是你在向别人销售 AI 产品，
15:17
it's really important to build that trust with people.
建立与人们的信任真的非常重要。
15:20
So I'm going to leave you with two questions.
所以我留给你们两个问题。
15:22
First is, as you're looking at building AI and you're asking yourself
第一个是，当你在构建AI时，你问自己
15:27
what are the things that we're going to do and what we do and what we not do?
我们要做哪些事情，我们做什么、不做什么？
15:30
Always start with how might we?
永远从“我们如何才能”开始。
15:32
So don't start with, there's two types of people.
所以不要一开始就说，有两种人。
15:35
Some people are like 50 reasons why you can't do something.
有些人能给你50个你不能做某事的理由。
15:38
I'm probably the other guy who gives you 52 reasons why you can do something.
我大概是另一种人，能给你52个你能做某事的理由。
15:42
Always be that person, even if 49 or the 52 things don't work.
永远做那个人，即使那49件或52件事都不奏效。
15:46
Be positive and try to do things.
保持积极，去尝试做事。
15:49
And then the last thing I would leave you with is what will we never do?
然后我最后想留给你的问题是，我们永远不会做什么？
15:53
So make that decision early on.
所以要尽早做出那个决定。
15:55
These are the things that we're just not going to do.
这些就是我们绝对不会去做的事情。
15:58
I really don't care if ChatGPT offers this new feature or if a GROC offers this new,
我真的不在乎ChatGPT是否推出了这个新功能，或者GROC是否推出了这个新东西，
16:02
we're not going to go there because we don't believe that this is what we should be using AI for.
我们不会往那个方向走，因为我们认为那不是我们应该使用AI的方式。
16:07
So finally, I've actually found these limits.
所以最终，我确实找到了这些边界。
16:10
These clear limits are actually freeing me up to be able to do way more than I could do before.
这些明确的限制实际上让我能够做到比以前多得多的事情。
16:17
I have a guideline set that says these are the things that we will and we will not do.
我有一套指导原则，规定了我们会做和不会做的事情。
16:22
And as long as I stay within those boundaries, it's actually way more freeing than it was
只要我待在这些边界之内，实际上比以前自由得多，
16:28
whenever I was always wondering, should we do this? Could we do this?
以前我总是想，我们应该做这个吗？我们能做这个吗？
16:31
Thanks.
谢谢。
16:32
There we go. Thank you so much, Donnie.
好了。非常感谢你，Donnie。
16:33
That was a great start to things.
这是一个很好的开始。
16:35
I can attest to the fact that, well, one, Donnie is on a lot of flights.
我可以证实，首先，Donnie经常坐飞机。
16:41
More than I can imagine.
超出我能想象的范围。
16:43
But every time there's a flight, I get a cool message like, hey, check out what I built,
但每次有航班的时候，我都会收到一条很酷的消息，比如，嘿，看看我做了什么，
16:49
which is really exciting.
这真的很令人兴奋。
16:51
So that was a great way to start out today.
所以这是今天一个很好的开始方式。
16:54
I'm thrilled to introduce our next speaker, Mark Lay.
我很高兴介绍我们的下一位演讲者，Mark Lay。
16:57
He's Vice President and CIO at Noblesse, a science and technology organization serving the federal government,
他是Noblesse的副总裁兼CIO，Noblesse是一家为联邦政府提供服务的科学技术机构，
17:05
where he leads IT strategy and delivery and keeps infrastructure secure and resilient.
他负责领导IT战略和交付，并保持基础设施的安全和韧性。
17:12
And I've had the privilege of interacting with a bunch of people at Noblesse and partnering with them in certain respects
我有幸与Noblesse的许多人交流过，并在某些方面与他们合作过
17:19
and it's great to be in a room with Mark at a whiteboard.
能和Mark一起站在白板前真是太棒了。
17:24
That's my favorite thing.
这是我最喜欢的事。
17:26
Ask amazing questions and he's going to talk to us about how you can orient yourself as AI is moving fast.
问一些精彩的问题，他会跟我们讲讲在AI快速发展的当下，你该如何找准自己的方向。
17:33
And what are the actual decisions you need to make as we're all in this environment where AI is moving faster than we can keep up?
以及在这个AI发展速度快到我们都跟不上的环境里，你实际需要做哪些决策？
17:41
I'm thrilled that Mark's here to talk about that. So welcome, Mark.
我非常高兴Mark能来这里聊这个话题。欢迎Mark。
17:46
Thank you.
谢谢。
17:47
I really love the leash concept. It's amazing.
我真的很喜欢leash这个概念。太棒了。
17:51
And as a CIO, I want to take it up on the level.
作为CIO，我想把它提升一个层次。
17:55
I have four decisions that I'd like to go through.
我有四个决策想逐一过一遍。
17:59
And it really starts with, when I talk to fellow CIOs and CTOs,
而这一切真的要从头说起，当我跟其他CIO和CTO交流的时候，
18:06
we all say AI is moving at the speed of light.
我们都会说AI正以光速在发展。
18:09
Every time we turn around, it's different, it's evolving.
每次一转身，它就不一样了，它在不断演进。
18:13
I meet with a vendor one month, their pitches different the next month, the features.
我这个月见一个vendor，下个月他们的pitch就变了，功能也不一样了。
18:17
They're all better, but they're different.
它们都更好了，但就是不一样了。
18:19
And so from a CIO perspective, that's really difficult to manage.
所以从CIO的角度来看，这真的很难管理。
18:24
And I think it's difficult because traditionally, we've had a business problem, process functions,
我觉得难的原因是，传统上，我们有一个business problem，process functions，
18:32
and codified how to address it.
并且把它如何应对给规范下来了。
18:35
And now we're in a situation where we are turning over to technology,
而现在我们处于这样一种情况：我们把一个目标交给技术，
18:40
a objective, we're putting guardrails or a leash around it,
我们给它加上护栏或者说拴上绳子，
18:45
and having it go off and produce an outcome.
然后让它自己去产生一个结果。
18:48
And that's a scary thing.
这是一件很可怕的事。
18:50
And it totally changes.
而且它完全改变了。
18:52
I believe the way you need to look at it.
我相信你需要看待它的方式。
18:54
And so what I'd like to go through is four questions where,
所以我想梳理的是四个问题，其中，
18:59
or decisions that I think are really critical to sort of,
或者说是一些我认为非常关键的决策，在某种程度上，
19:03
at a nuts level above some of the talks earlier,
在比之前一些演讲更基础的层面上，
19:06
just how you should look at AI across the organization.
就是你应该如何在整个组织中看待AI。
19:09
The first decision is really about what should I delegate to AI.
第一个决策实际上是关于我应该把什么委托给AI。
19:14
And my premise here is this idea of turning over activities to an automated system
我在这里的前提是，把活动交给自动化系统这个想法
19:25
is very similar to what we've done for many, many years.
和我们多年来一直在做的事情非常相似。
19:30
And that is manage people.
那就是管理人。
19:32
We have to delegate activities to people.
我们必须把活动委托给人。
19:34
And my premise is when you're in front of chat GPT or another tool
我的前提是，当你面对ChatGPT或其他工具时
19:39
or you're trying to decide to build a system,
或者你正在决定要构建一个系统，
19:42
you need to decide what you're going to delegate to that AI.
你需要决定你要把什么委托给那个AI。
19:46
And how much of a leash you're going to put on it.
以及你打算给它多大的自由度。
19:48
And I really think there are two dimensions of that.
我真的觉得这有两个维度。
19:50
There's what reward you're going to get out of it and what risk.
这里面有你能从中得到什么回报，以及有什么风险。
19:54
On the reward side, you're trying to get something better or something more efficient.
在奖励方面，你是在试图获得更好的东西或更高效的东西。
19:59
And if you don't get either of those, you really shouldn't put it in AI.
如果你这两样都做不到，你真的不应该把它放进AI里。
20:03
And then the question is on the risk side, is there a risk of an error?
然后问题就在于风险方面，是否存在出错的风险？
20:09
And what's that risk of an error? Or is there a low risk of an error?
那个出错的风险是什么？还是说出错的风险很低？
20:12
And what's the consequence of that error?
那个错误的后果又是什么？
20:15
And so really, if you take this a step further, you can look at it and say,
所以真的，如果你再进一步看这个问题，你可以说，
20:20
if there's better outcomes and less burden to do it, that's a clear gain.
如果有更好的结果而且做起来的负担更小，那就是明显的收益。
20:27
But then the next question is, is there risk in getting that gain?
但接下来的问题是，获得这个收益有没有风险？
20:31
And how do you address it?
你又要怎么应对？
20:33
If there's no risk, like when I go online and explore things, that's down in the lower left side.
如果没有风险，就像我上网探索东西的时候，那就在左下角那个位置。
20:39
The upper right side is an area where you may not, you might have to put guardrails very tightly around something.
右上方是一个你可能不能、你可能必须把 guardrails 非常紧地围绕在某事物周围的区域。
20:46
And earlier Heidi made a point about some guardrails are deterministic guardrails.
之前 Heidi 提出了一个观点，有些 guardrails 是确定性的 guardrails。
20:53
The deterministic guardrails fall up in or in that left lower quadrant,
确定性的 guardrails 落在左下象限，
20:59
where you have high risk of something happening.
在那里你有某事发生的高风险。
21:04
So the first idea is look at everything as you're delegating,
所以第一个想法是，当你委派时审视一切，
21:10
just like you would delegate to one of your workers.
就像你会委派给你的一名员工一样。
21:13
They can make mistakes, but you want to control those mistakes and you want to keep them in balance.
他们可能犯错，但你想要控制那些错误，你想要让它们保持平衡。
21:18
You want to put a leash on what you're delegating.
你想要给你所委派的事情拴上一根牵引绳。
21:21
And I find this an easy framework to sort of look at how much of a leash you have to put on
而且我觉得这是个挺容易上手的框架，可以用来看看你得给AI多大的约束空间，
21:27
and really where you should put the AI and where you're going to get the most value.
以及到底该把AI放在哪儿、在哪儿能获得最大的价值。
21:32
The second, the next area is really the idea of who's accountable when AI acts.
第二个、下一个领域其实是关于当AI采取行动时，谁来负责。
21:40
And this is an area where I think a lot of organizations have oversimplified this.
我觉得很多组织在这个问题上把事情过度简化了。
21:45
And the first premise is that accountability can't be delegated.
第一个前提是，问责不能被委托。
21:50
I mean, humans can delegate accountability to other humans.
我的意思是，人类可以把问责委托给其他人类。
21:54
But with AI, we should not be delegating accountability.
但面对AI，我们不应该把问责委托出去。
21:59
So somewhere up the chain in the racey chart for the A,
所以在RACI图里，往上追溯的某个环节，对于这个A，
22:03
there always needs to be a person.
总得有人来做这件事。
22:05
And there are four areas where I think this is important.
我认为有四个方面很重要。
22:08
It sort of covers the landscape.
这大致涵盖了整个领域。
22:10
The first is protecting the data.
第一个是保护数据。
22:12
And one of the things that we've done in our organization,
我们在组织里做的一件事，
22:15
we have, we deal with a lot of compliance from the federal government, different agencies.
我们处理很多来自联邦政府、不同机构的合规要求。
22:21
And security, we actually spent the last year identifying over a thousand data sets that we thought were relevant.
还有安全方面，我们实际上花了去年一整年时间，识别出了超过一千个我们认为相关的数据集。
22:29
And we, I don't think we got them all, but we started.
而且，我不认为我们全都找齐了，但我们已经开始了。
22:32
And then we identified a data store, a person who owns that.
然后我们确定了一个数据存储，以及一个拥有它的人。
22:36
Okay. And it's responsible for understanding what the data is, what the quality is, what the timeliness of it.
好的。它负责理解数据是什么、质量如何、时效性怎样。
22:43
And also what, what, how does it have to be protected?
还有，它需要以什么方式受到保护？
22:46
Who can see it? Who can't see it?
谁能看到它？谁不能看到它？
22:48
If you don't have that, I mean, and one of the things that came out of this is even people that you know,
如果你没有这些，我的意思是，这件事带来的一个发现是，即使是你认识的人，
22:54
they know they own the data, our HR department, are their finance department.
他们知道自己拥有数据，我们的人力资源部门、他们的财务部门。
22:59
They were so traditionally programmed to say, well, the data's in the accounting system,
他们传统上被训练成这样说话：好吧，数据在会计系统里，
23:04
or the data's in the HR system, it protects it.
或者数据在人力资源系统里，系统会保护它。
23:08
Okay. As you go to a GenTik AI, and these tools you can't live in that type of world anymore,
好。当你使用GenTik AI这类工具时，你不能再生活在那种模式里了，
23:14
you have to have somebody who owns the data and knows the data.
你必须有一个拥有数据、了解数据的人。
23:18
In the end, after a year of work there, we got down to about 100 and 115, 120 data sets.
最终，在那里工作了一年之后，我们把数据集缩减到了大约100到115、120个。
23:27
And these are not data files. These are data sets of information, tabular, or even reports that were critical to our operations of our business.
这些不是数据文件。这些是信息数据集，是表格，甚至是对我们业务运营至关重要的报告。
23:36
And we had about 30 data storage of those, but that was the first line.
我们大约有30个数据存储，但那只是第一层。
23:41
The second is this whole concept of validating the output.
第二是整个验证输出的概念。
23:46
I mean, you can put guardrails on it, but depending on what you've delegated and the risk of error,
我的意思是，你可以给它加上护栏，但取决于你委托了什么以及出错的风险，
23:51
there's a whole activity of data output quality.
数据输出质量本身就是一个完整的活动。
23:55
And I think when you look at, is it a platform, or it's a centralized tool, who's checking that data quality?
我认为，当你审视它到底是一个平台，还是一个集中化的工具时，谁来检查数据质量呢？
24:04
Going back to the platform versus the enterprise system, I like to use the analogy of,
回到平台与企业系统的对比，我喜欢用这样一个类比，
24:12
we've been living with this for many, many years. Somebody earlier said Lotus 123, which I'm old enough to remember,
我们已经和这个问题共存了很多很多年。之前有人提到Lotus 123，我年纪够大还记得它，
24:18
but Excel is a platform that's been deployed to all our employees and finance teams.
但Excel是一个已经部署到我们所有员工和财务团队的平台。
24:25
Really, to me, it's the equivalent of AI.
对我来说，它真的就相当于AI。
24:29
It allowed before Lotus 123 and Excel, there weren't tools to manipulate numbers or math the way we can.
在Lotus 123和Excel之前，没有工具能像我们现在这样操作数字或进行数学运算。
24:38
That changed the world. I see AI doing the same thing in terms of pushing it out.
那改变了世界。我看到AI在推广普及方面也在做同样的事情。
24:44
But if you think about Excel as opposed to a centralized power BI tool where you have structured reports,
但如果你把Excel和一个集中化的Power BI工具对比，后者有结构化的报表，
24:52
when you give the finance team an Excel spreadsheet, they can now build macros.
当你给财务团队一个Excel表格时，他们现在可以构建宏。
24:59
They can build traditional things that they only know and only their teams know how to run.
他们可以构建一些传统的东西，只有他们自己、只有他们的团队知道怎么运行。
25:06
But so who controls that? Who manages the output?
但那么谁来控制这些？谁来管理输出？
25:10
You can't centrally manage that decentralized resource.
你无法集中管理那种去中心化的资源。
25:16
Third, which I think Ivan also mentioned, is monitor the behavior of the agents.
第三点，我觉得Ivan也提到过，就是监控agent的行为。
25:21
You've got to watch them, which is a totally new concept.
你得盯着它们，这是一个全新的概念。
25:24
But from managing employees' work, we already know how to do that.
但从管理员工的工作来看，我们已经知道怎么做了。
25:28
It's like, how do you take that principle that we've used for years and think of it now in an automated fashion and where it would apply?
就好比，你如何把用了多年的原则，现在以自动化的方式去思考它，以及它会在哪里适用？
25:35
And then the last area which sort of comes back to my Excel example is owning the outcomes.
然后最后一个领域，某种程度上又回到了我那个Excel的例子，就是拥有结果。
25:43
Who's accountable for the outcomes?
谁对结果负责？
25:45
We have teams where we have a contracts team where we have 20 contracts administrators.
我们有团队，有一个合同团队，里面有20个合同管理员。
25:53
One of the early career contact administrators built an app to pre-process all.
其中一个早期的合同管理员做了一个应用来预处理所有……
25:59
He went and got everybody's knowledge from the 2030 year contracts administrator and built this system that would do a first pass on the contracts.
他去把20、30年合同管理员的所有知识都收集起来，建了这个系统，可以对合同做第一遍处理。
26:10
Okay, we didn't do that in the IT department. They did it in their group.
好吧，我们IT部门没做这个。是他们自己团队做的。
26:15
That has to be managed and maintained. How does that? How does that?
这需要被管理和维护。那怎么……怎么……
26:20
How do you keep that going? So there is that business ownership side of it and keeping track of that.
你怎么让这个持续运转？所以就有那个业务归属的一面，以及持续跟踪它。
26:26
Again, I like to use the Excel analogy, at least with the finance department, they understand the only macros.
再说一次，我喜欢用Excel来打比方，至少财务部门的人能理解，他们只懂宏。
26:33
The last area, which is another area that I don't think a lot of people are talking about, but CIOs worry about it a lot because it's costly.
最后一个领域，也是我认为很多人没有在谈论的另一个领域，但CIO们非常担心，因为成本很高。
26:43
The question is, what technical debt are you creating? And there's a whole new different type of technical debt.
问题是，你在制造什么样的技术债务？而且有一种全新的、不同类型的技术债务。
26:49
And the first layer of it comes from centralized or distributed platform and what I was already talking about with regards to departments building their own application.
它的第一层来自集中式或分布式平台，以及我之前提到的关于各部门构建自己应用的问题。
27:00
That's the first piece of it. But the second piece of it, and I've had experience this, we rolled out certain tools where we let teams go off and build apps and share apps.
这是第一部分。但第二部分，我自己有过这样的经历，我们推出了一些工具，让团队去构建应用并共享应用。
27:12
Well, guess what? The vendor changed the tool such that we went to the next version of it and none of the old apps converted.
结果呢？供应商改变了工具，我们升级到了下一个版本，所有旧的应用都没有被转换过来。
27:21
Okay, so there's a whole level of when you're rolling out a technology, whether it's whether it's a control plane layer tool, whether it's an end, you know, just a user interface app or whatever in the technical stack, the technical stack is still evolving.
所以，当你推出一项技术时，无论是控制平面层的工具，还是终端用户界面应用，或者技术栈中的任何东西，技术栈仍然在演进。
27:40
And the data you're using is evolving. So there's a lot more to worry about in terms of what dependencies related to data or other systems or other things in the value chain of delivering that.
你使用的数据也在演进。所以在交付的价值链中，关于数据或其他系统或其他事物的依赖关系，有更多需要担心的事情。
27:54
And then the second piece of it is once you know that, can you reverse it? What's the cost of reversing it that actual cost time, whatever. And then the last piece is, which is, I think every organization is starting to hit, is do you have the capacity to evolve?
然后第二点是，一旦你知道了这一点，你能逆转它吗？逆转它的成本是什么，实际成本、时间等等。然后最后一点，我认为每个组织都开始碰到的，就是你有能力去演进吗？
28:12
I mean, a lot of people have talked about the idea of or the cost of tokens, if I build something, is it going to be too expensive for the value I'm getting out of it.
我是说，很多人讨论过token的想法或成本，如果我构建了某个东西，它会不会太贵，跟我从中获得的价值不匹配。
28:23
But the other point is, let's say the technology changes and I need to put something else in place, either that centralized IT team or that decentralized group that's leveraged, do they have the capacity to keep running what they deployed?
但另一点是，假设技术变了，我需要换成别的东西，不管是那个集中式IT团队还是那个被赋能的去中心化小组，他们有能力持续运行他们所部署的东西吗？
28:39
These are probably the hardest questions to answer in a fast-moving environment.
这些可能是在快速变化的环境中最难回答的问题。
28:45
So those are the four I have. I just want to close on this. It's sort of an adaptation of Drucker's management theme, which is, hey, I can do a lot of stuff.
这就是我有的四个问题。我想用这个来收尾。这算是对Drucker管理主题的一种改编，就是，嘿，我能做很多东西。
28:56
There's nothing there and it can produce a lot of output. But it's really how you choose the and address the four decisions before of whether it's worth producing and the cost of that. So thank you.
那里什么都没有，它能产出很多输出。但真正关键的是你如何选择和处理之前的四个决策，即它是否值得产出以及产出的成本。所以谢谢。
29:14
Thank you so much, Mark. Appreciate it.
非常感谢你，Mark。很感激。
29:18
If you're doing an AI model or creating your first agent to do a basic task, can be really easy with modern tools. But it's a whole nother thing to create an agent that's tied into all of the different systems that it needs to access in order to deliver real ROI over time in your organization.
如果你在做一个AI模型或者创建你的第一个agent来做基本任务，用现代工具可以非常简单。但创建一个agent，让它接入它需要访问的所有不同系统，以便随着时间在你的组织中交付真正的ROI，那完全是另一回事。
29:36
That's why I love what SERP API is doing and why I've used it in the past to connect valuable resources into my agents.
这就是为什么我喜欢SERP API在做的事情，也是为什么我过去一直用它把有价值的资源接入到我的agents里。
29:45
SERP API handles the complexity and delivers search results as structured data that your application can use, whether that's from Google, being Amazon, Walmart, YouTube, or hundreds of other APIs.
SERP API处理了复杂性，把搜索结果以结构化数据的形式交付给你的应用，不管是来自Google、Amazon、Walmart、YouTube还是数百个其他API。
29:57
They also have built-in enterprise security and compliance. They're SOC 2, Type 2, SOC 3, and ISO 27001 certified, making it a strong option for organizations where security and compliance matter.
他们还有内置的企业级安全和合规。他们通过了SOC 2 Type 2、SOC 3和ISO 27001认证，对于安全和合规至关重要的组织来说，这是一个强有力的选择。
30:14
I encourage you to visit SERP API.com slash practical AI to get started with SERP API for free and get 250 searches every month. That's SERP API.com slash practical AI.
我鼓励你访问SERP API.com slash practical AI，免费开始使用SERP API，每月获得250次搜索。网址是SERP API.com slash practical AI。
30:30
Next we have, I'm excited to hear this story and journey from Darren at Arrowcore.
接下来，我很期待听到来自Arrowcore的Darren分享的这个故事和历程。
30:42
Darren is a senior data analytics manager at Arrowcore Technologies, which is local here to our surroundings.
Darren是Arrowcore Technologies的高级数据分析经理，这家公司就在我们周边本地。
30:50
But has this incredible technology to perform washes on jet engines all around the world.
但他们拥有这项令人惊叹的技术，可以在世界各地对喷气发动机进行清洗。
30:58
So they do this all around the world. And so what's been interesting to see is how Darren has been able to practically roll out agents into their organization, which a lot of the organization isn't building these agents, but certainly benefits from the insights that these agents are able to put together from this wash data and work orders and other things from all over the world.
他们在全球范围内做这件事。所以一直很有意思的是看到Darren如何切实地将agents推广到他们的组织中，组织中很多人并不是在构建这些agents，但 certainly 从这些agents能够从全球各地的清洗数据、工单和其他信息中整合出的洞察中受益。
31:27
So this is super practical and I would encourage you both to, of course, listen but connect with Darren after his talk as well to follow up on that really practical journey. So welcome, Darren.
所以这个非常实用，我当然鼓励你们两位去听，同时在Darren演讲之后也跟他联系，跟进那段非常实用的历程。那么欢迎Darren。
31:41
There we go. So my name is Darren Talley. I am a senior data analytics manager for a company called Arrowcore Technologies. As Daniel said, we wash jet engines.
好的。我叫Darren Talley。我是一家叫Arrowcore Technologies的公司的高级数据分析经理。正如Daniel所说，我们清洗喷气发动机。
31:52
We are an incredibly analog company in a digital world. There are no digital components on the wash trucks whatsoever. So my job is to bring some of the digital world to our company.
在一个数字化的世界里，我们是一家极其模拟化的公司。清洗车上完全没有数字组件。所以我的工作就是把一些数字世界的东西带到我们公司来。
32:03
And as Daniel so eloquently said, we are all over the world, right? So we have about 11 domestic sites at the major airports here in the US started deploying internationally and overseas sites in Japan in Riyadh in in Brazil.
正如Daniel如此雄辩地说的，我们遍布全球，对吧？我们在美国主要机场大约有11个国内站点，已经开始在日本、利雅得、巴西部署国际和海外站点。
32:18
That expansion is growing every year as our technology really grows to help the big airlines in it, you know, especially in a time where you feel prices are going up. No one can deny that. And so that means your airline tickets are going up to reflect that we're trying to counteract that as best as we can.
随着我们的技术真正发展，帮助大型航空公司，这种扩张每年都在增长，你知道，特别是在你觉得价格在上涨的时候。没人能否认这一点。所以这意味着你的机票价格也在上涨以反映这一点，我们正在尽最大努力抵消这种影响。
32:36
Give a quick bit since, you know, this is a very niche thing that we do. Again, we wash jet engines. We use a patented nucleated foam technology to penetrate and clean the insides of jet engines, right?
简单说一下背景，因为你知道，这是我们做的一件非常细分的事情。再说一次，我们清洗喷气发动机。我们使用一种获得专利的成核泡沫技术来渗透并清洁喷气发动机的内部，对吧？
32:49
So you got a lot of carbon build up, a lot of fouling that happens inside these engines. Our job is to go in multiple cycles go in and clean that stuff out where that extends to what I do.
所以这些发动机内部会有大量积碳，大量污垢堆积。我们的工作是进行多个循环，进去把这些东西清理掉，这就延伸到了我所做的事情。
32:59
I look at the before and after engine performance data determine the efficacy of the wash itself, but looking also into long term trends, looking into predictive maintenance type things trying to find the bad apples that are out there for fleets of aircraft and determine how we can help the major operators.
我查看清洗前后的发动机性能数据，确定清洗本身的效果，同时也研究长期趋势，研究预测性维护之类的东西，试图找出机队中的那些问题个体，并确定我们如何帮助主要运营商。
33:16
Like Daniel said, this is a very practical presentation. This is what I have done over the last year. I think these kind of presentations fit in this kind of format very well.
就像Daniel说的，这是一个非常实用的分享。这是我过去一年里做的事情。我觉得这类分享非常适合这种场合。
33:26
You have a ton of people who talk about how in the safeguards and what how important all of those things are. But I think sometimes you get left coming out of here being like, well, what can I do?
有很多人都在讲safeguards有多重要、这些事情有多关键。但我觉得有时候你听完之后会想，那我能做什么呢？
33:36
How do I apply this to my company? What do I take home? So I'm going to give you two very, very clear examples of what myself and my team did in order to bring using AI to bring better information and better data to the rest of our team for decision making.
我怎么把这个应用到我的公司？我能带走什么？所以我要给你们两个非常非常清晰的例子，是我和我团队做的事情，用AI把更好的信息和更好的数据带给团队其他人来做决策。
33:52
So the first one we're going to talk about. So the biggest challenge we have is access to data. As I said, we are an analog company in the digital world. How do we get the information to the key decision makers in time for them to make those decisions?
第一个我们要聊的。我们最大的挑战是数据访问。就像我说的，我们是一家在数字世界里的模拟公司。我们怎么把信息及时送到关键决策者手里，让他们能做出决策？
34:05
The other one being work order verification. I'll talk about that one a lot more. That's a very nuanced problem. But I think it's probably a little more interesting in the first part. So access to data. Does anybody still know right SQL and like a daily basis?
另一个是工单验证。这个我会讲得更多。那是一个非常微妙的问题。但我觉得第一个可能更有意思一些。所以，数据访问。还有人知道SQL并且每天都在用吗？
34:19
Yeah, that's actually more than I expected. So again, analog company digital world, not a lot of people know how to write SQL. Most people don't have access to the database intentionally as deletes happen a little too easily these days.
嗯，这比我预期的要多。再说一次，模拟公司在数字世界，没多少人会写SQL。大多数人没有数据库的访问权限，这是故意的，因为现在删库太容易了。
34:32
But then that leads to the analyst bottleneck as we like to call it, which is the request comes to me. I have to go in, pull the data, merge the data, analyze the data, send the results. And we do that for 40 hours a week.
但这就导致了我们说的分析师瓶颈，就是请求到我这里来。我得进去拉数据、合并数据、分析数据、发结果。我们每周40个小时都在干这个。
34:46
And we've done no long-term growth. We've done no extra tasks. We've done no research. All we've done is fill out how many washes did we do in Atlanta last night?
我们没有做任何长期增长。我们没有做任何额外的任务。我们没有做任何研究。我们做的全部就是填一下昨晚在Atlanta洗了多少次？
34:55
55 times for different employees. And then that also leads to another lovely coin term, which is work around sprawl, which essentially is if I, because again, as you didn't hear in my title, not developer, not SQL person.
55次，针对不同的员工。然后这也引出了另一个可爱的硬币术语，叫做work around sprawl，本质上就是，因为我，再说一次，正如你从我的头衔里没听到的，不是developer，不是SQL person。
35:15
So all of this turns into pandas.reads CSV with group buys and location filters, right. So those get very complicated. You look into my Jupyter notebooks and you see cells and cells and cells are just every random pull that I had to do that week, go through another week, determine what's important, delete everything else.
所以所有这些都变成了pandas.reads CSV，带着group buys和location filters，对吧。所以这些变得非常复杂。你翻我的Jupyter notebooks，你会看到一格一格一格，全是我那周不得不做的各种随机拉取，再过一周，确定什么重要，把其他全删掉。
35:34
That just gets spread around. No one else has access to that. It's not useful to anybody.
这些东西就散落各处。没有其他人能访问。对任何人都没用。
35:39
So those are kind of all the challenges that we particularly faced in our data to mention also our data analysis group consists of me and one other employee.
所以这些就是我们特别在数据方面面临的所有挑战，还要提一下我们的数据分析组就我和另一个员工。
35:47
So here comes Daniel with his big toy. And the first one, so Agent Forge was our was, you know, really a big key for us to be able to deploy some of these tools out there. Right.
所以Daniel带着他的大玩具来了。第一个，Agent Forge是我们能够把其中一些工具部署出去的关键。对吧。
35:58
So the first one, all right. Well, if I just give it this file to the agent, it probably can tell me all the answers about all the work orders from the last week, month year.
所以第一个，好吧。如果我直接把这个文件给agent，它大概能告诉我关于上周、上个月、去年所有work orders的所有答案。
36:07
Okay. Well, then there's MCP servers. Okay. Cool. So like now I can just query the database directly. Well, you still have to go into Agent Forge. You still have to go in and pull the information out.
好。然后还有MCP servers。好。酷。所以现在我可以直接查询数据库了。但你仍然得进Agent Forge。你仍然得进去把信息拉出来。
36:17
No one likes an extra login. You know, that's really hard. Boom. Now you can put it in teams. And I think my phone's blown up three or four times already with guys using the agent that we put into teams in order to do.
没人喜欢多一次登录。你知道，那真的很难。砰。现在你可以把它放进Teams里。而且我觉得我的手机已经炸了三四次了，都是那帮人在用我们放进Teams里的agent来干活。
36:27
The teams in order to access the information. So that ability to just ask very simple text query. You know, how many after you, which one I put up there. Oh, yeah, return me all the work orders for this particular tail number. There was an issue, something happened.
团队为了获取信息。所以那种能够直接用非常简单的文本查询的能力。你知道，比如在你之后有多少个，我放在上面的是哪一个。哦，对，把某个尾号的所有工单都返回给我。出了问题，发生了某件事。
36:41
All right. Here's all the information. Here's the actual query. Here's the results. You know, fantastic for our team to be able to go and find that history, right. History is always the most important part.
好的。这是所有信息。这是实际的查询。这是结果。你知道，对我们团队来说能够去找到那些历史记录，太棒了。历史记录永远是最重要的部分。
36:51
We've gone through so many different systems to be able to see all the work orders that are on that aircraft when they were who did them where they were.
我们经历了那么多不同的系统，才能看到那架飞机上的所有工单，什么时候做的、谁做的、在哪里做的。
37:00
And then obviously all of the technical data that we gather from every single wash once they got moved into teams very high adoption. Everybody's using it all the way through.
然后显然还有我们从每一次清洗中收集的所有技术数据，一旦它们被迁移到Teams里，采用率非常高。所有人从头到尾都在用。
37:10
And it's getting it's getting a lot, a lot of hits and seeing that in the analytics. And it's been great. It's it's truly relieved to, you know, burden off of our analysis team to do more things that are beneficial to the company.
而且它得到了很多、很多的点击量，在分析数据里能看到。这非常棒。它真的减轻了，你知道，减轻了我们分析团队的负担，让他们能去做更多对公司有益的事情。
37:20
You know, things that used to take 30 minutes an hour a week for complex stuff. The system now doesn't a couple of minutes. If you gave it a really complicated one.
你知道，以前那些复杂的东西要花30分钟、一个小时、一周的时间。现在这个系统只需要几分钟。如果你给它一个非常复杂的。
37:29
You know, this came back in about three seconds. So it does a really, really good job of getting the decision makers, the information they need.
你知道，这个大概三秒钟就返回了。所以它在把信息传递给决策者方面做得非常、非常好。
37:38
So yeah, the big one, self service, right, getting our decision makers that information that they particularly needed getting leadership to not have to ping me every time they need an answer to a question.
所以是的，最重要的一点，自助服务，对吧，让我们的决策者获得他们特别需要的信息，让领导层不必每次有问题就来找我。
37:49
And also kind of freeing up myself and freeing up my other analysts to be able to do these bigger projects.
这也把我自己和我其他的分析师解放出来，能够去做这些更大的项目。
37:55
Different types, you know, bigger reporting, bigger research projects, bigger experiments through our data and how important all of that is.
不同类型的，你知道，更大的报告、更大的研究项目、通过我们的数据做更大的实验，以及所有这些有多重要。
38:03
But we weren't really able to find the time to do it. And now we're able to go do that.
但我们之前真的找不到时间去做这些。现在我们可以去做了。
38:08
So that's kind of on the simpler end. I know that there's a lot of the ability for people just to kind of connect.
所以这算是比较简单的一端。我知道有很多能力让人们直接连接。
38:13
Clawed directly to your data and all those different ways we talk a lot about safe guards and guardrails and like that.
直接连接到你的数据，以及我们经常谈论的各种方式，关于安全防护和护栏之类的。
38:19
We thought that building it through Agent Forge with the MCP servers provides that was guardrails also giving you the access that is wholly required.
我们认为通过 Agent Forge 和 MCP servers 来构建它，提供了那些护栏，同时也给了你完全所需的访问权限。
38:27
And now kind of the more interesting one.
现在说更有趣的那个。
38:29
So this will call work order verification, right? So hopefully most of you have not had to spend a night in an airport.
这个叫做工单验证，对吧？希望你们大多数人都不用在机场过夜。
38:36
But if you ever have that's when all the maintenance happens.
但如果你有过那种情况，那正是所有维护进行的时候。
38:39
So guys are out there on every plane doing a whole ton of work, including our team who work shifts from about 7 p.m. to about 5 a.m.
所以伙计们在每架飞机上做着大量的工作，包括我们团队，他们从晚上7点左右工作到凌晨5点左右。
38:47
Doing these engine washes out on an aircraft.
在飞机上做发动机清洗。
38:50
We do anywhere between let's say 15 to 30 washes in a night across the very different sites. You can get about six washes in a night at a particular site.
我们一晚上在不同的站点大概做15到30次清洗。在某个特定站点，一晚上大概能做6次清洗。
38:59
Typically it is one technician out there doing the wash on the particular engine.
通常是一个技术员在那里对特定发动机进行清洗。
39:04
He's out there with his phone trying to get all the data that we need him to gather writing down all the information that he used to collect.
他拿着手机，试图收集我们需要他收集的所有数据，写下他过去收集的所有信息。
39:10
Also making sure that he pumps the right amount of fluid into the engine to wash it.
还要确保他往发动机里泵入正确量的液体来清洗它。
39:14
There's no one out there to validate his inputs his or her inputs.
现场没有其他人来验证他或她的输入。
39:18
Sorry. So errors happen that fingering happens in the middle of the night 5087 looks a lot like 5078.
抱歉。所以错误会发生，打错字的情况会在半夜出现，5087看起来很像5078。
39:27
It happens. So how can but when that comes back at the end of the month.
这种事难免。那怎么能……但当月底那个数据回来的时候。
39:31
Well, we don't have the right billing counts. We don't have the right aircraft types.
嗯，我们没有正确的计费数量。我们没有正确的机型。
39:35
We don't have the right numbers. What happened there?
我们没有正确的数字。那里发生了什么？
39:40
We have had a long system of using these pictures that the operators have to submit every night on which are the first three kind of shown there on the top picture on the right, which would be a picture of a whiteboard that we've built on the side that acts as a pseudo form that the operators have to fill out.
我们长期以来一直使用一套系统，让操作员每晚提交这些照片，也就是右上角那张图里最上面显示的前三张，那会是一张白板的照片，我们在一侧建了这块白板，充当操作员必须填写的伪表单。
39:54
Plus taking a picture of the aircraft and then taking some other pictures of things of the we call effluent, but it's the the crud that comes out the bottom of the engine is that last picture.
再加上拍一张飞机的照片，然后再拍一些其他东西的照片，我们称之为effluent，也就是从发动机底部排出来的那些污垢，就是最后那张照片。
40:04
I'll love to talk about that with you afterwards if you're interested.
如果你感兴趣的话，我很想之后跟你聊聊那个。
40:08
But our real key is, you know, verifying remotely the work order and the wash gut every night. So right, our challenge.
但我们真正的关键，你知道，是远程核实工单和每晚的wash gut。所以，对，我们的挑战。
40:16
And here's what we came up with for that solution. The other side of this would be that we do have a quality control plan in place for someone else to verify the work orders based on the pictures, looking at them remotely.
下面是我们针对那个解决方案想出的办法。另一面是，我们确实有一套质量控制计划，让其他人根据照片远程查看来核实工单。
40:28
Other site leads from other spots, other professionals. We kind of kept noticing that the same mistakes kept being made over and over.
来自其他地点的其他现场负责人、其他专业人士。我们一直注意到同样的错误在反复出现。
40:34
The scenario I just described where somebody looks at a 5078 and thinks that it's 5087 logs it in.
就是我刚才描述的场景，有人看到一个5078，以为它是5087，然后把它录进去。
40:40
That would be flagged when 5087 comes in the next night and you're like, well, we just we just did this one.
等到第二天晚上5087进来的时候就会被标记出来，你就会想，我们刚做过这个啊。
40:46
Why is it coming in again? Oh, that's a repeat work order. Okay, delete come the end of the month. No, no, that one was real.
为什么又来了？哦，那是个重复工单。好，月底删掉。不不不，那个是真的。
40:52
There was the other one was wrong. How do we fix that? So this is where a vision model came into place.
另一个才是错的。我们怎么解决这个问题？这就是视觉模型派上用场的地方。
40:58
Luckily, because of our use in Agent Forge and some of the work with Daniel and big thanks to Sharon, who really helped me build this and deploy it.
幸运的是，因为我们在Agent Forge中的使用以及与Daniel的一些合作，还要特别感谢Sharon，她真的帮我构建并部署了这个。
41:07
So the crews take those photos. They upload them in there in the power app that we use for them to generate the particular work order every day at 815 in the morning.
所以工人们拍那些照片。他们在我们用来让他们每天早上8点15分生成特定工单的Power App里上传这些照片。
41:17
Agent goes in scans all the pictures verifies it against the information put in into the app.
Agent会进去扫描所有图片，跟录入app里的信息做核对。
41:23
You know, the best part of this would be those end numbers are, you know, are pretty big written on the side of the aircraft.
你知道，这个最棒的地方在于那些尾号，就是写在飞机侧面挺大的那些数字。
41:28
Vision model does pretty great at returning them. It also actually does pretty great at returning that little tiny number in the top of the tail when we needed to do that as well.
Vision model在识别这些数字方面表现相当好。它其实在识别尾翼顶部那个小数字方面也做得相当不错，当时我们需要做这个的时候。
41:36
So vision model goes in looks at the aircraft identifies it returns it if it aligns with what is input in the app checks it off.
所以vision model进去看飞机、识别它、返回结果，如果跟app里输入的一致就打个勾。
41:45
If it doesn't returns a flag that flag gets sent to me and says, Hey, something's wrong with this work order. Please go have somebody a human go look at it and see what's going on here.
如果不一致就返回一个flag，那个flag会发给我，说，嘿，这个工单有问题。请让人去人工看一下，看看是怎么回事。
41:54
Sometimes it's the wrong picture. Sometimes it's a cloud or a dark night a bad picture or something like that. Compare and flag.
有时候是图片错了。有时候是云或者黑夜、图片质量差之类的。比对然后标记。
42:02
That's the exact prompt that I wrote. I ran did help me with that as well. But just kind of back to the, you know, how many times I had to go through and simplify that prompt to make sure that it didn't say,
那就是我写的原话prompt。我跑了，did帮了我一把。但就是回到那个，你知道，我花了多少次去简化那个prompt，确保它不会说，
42:14
Well, I think this aircraft is blah, blah, blah. So no, you need to return N and these numbers and that's it.
嗯，我觉得这架飞机是blah blah blah。所以不行，你需要返回N和这些数字，就这些。
42:21
How many times we went back and forth fighting about that to make sure that it was correct. But the results. So this was wildly successful.
我们来回反复争论了多少次，就为了确保它是正确的。但结果呢。所以这个项目非常成功。
42:27
So successful that it has been, it goes out to leadership every day at noon with a list of all the work orders done in the previous 24 hours across every site, both commercial and in the military.
成功到什么程度呢，它每天中午都会发给管理层，列出过去24小时内在所有站点完成的所有工单，包括商业和军事的。
42:38
It also then flag was expanded to flag other abnormal procedures, other text query or text inputs that the operators put into the app to say, OK, well, I had this issue.
然后它还被扩展了，用来标记其他异常流程、其他文本查询或操作员输入到app里的文本内容，比如说，好吧，我遇到了这个问题。
42:50
We worked around it. We fixed it. But just so that everybody knows we had this issue during the wash or things like we did this for a smoke and fume resolution that the operator asked us to do or we missed a cycle or we missed a wash.
我们绕过了它。我们修复了它。但只是让大家都知道我们在清洗过程中遇到了这个问题，或者类似的情况，比如我们为了烟雾和废气解决方案做了这个，是操作员要求我们做的，或者我们漏了一个周期，或者我们漏了一次清洗。
43:05
All that information gathered agent also now goes through those notes flags and for anything related to smoke fume events flags and anything related to abnormal procedures.
所有收集到的信息，agent现在也会去检查那些备注标记，检查任何与烟雾废气事件相关的内容，标记任何与异常流程相关的内容。
43:14
And then goes through and does a whole bunch of other checks right we found that the same things were being missed over and over again by the human QC going through and checking these work orders.
然后还会做一大堆其他检查，对吧，我们发现同样的事情被人工QC一遍又一遍地漏掉，他们在检查这些工单的时候。
43:24
You know, we work across the middle of the night. There's weird things with what happens when work orders cross midnight and all of a sudden work order start taking 36 hours and you're like, well, no, that's probably not true. It was probably four hours. It just got crossed over midnight. How did that happen? So checking all of those type things.
你知道，我们在半夜工作。工单跨午夜的时候会出现一些奇怪的情况，突然工单开始变成36小时，你就会想，不对，那可能不是真的。大概是4小时。只是跨了午夜而已。怎么会这样呢？所以要检查所有这些类型的事情。
43:42
So again, like I said, wildly successful expanded to all work orders flags are currently just set and tracked and the team goes through and resolves them as needed.
所以，就像我说的，非常成功，扩展到了所有工单，标记目前只是设置和追踪，团队会根据需要去处理它们。
43:52
We are building in a resolution system as we speak. So those will actually be flags that need resolution by a human to go through and check and either identify that they are real fix the problem or say that they weren't really a problem at all.
我们正在构建一个resolution系统。所以这些实际上会是需要人工来解决的flag，需要人去检查，要么确认它们确实是问题并修复，要么说它们根本不是问题。
44:05
The biggest results of this I think would be a an op as we talked about in the last presentation, which was fantastic.
我认为最大的成果，就像我们在上一个presentation中谈到的，那个presentation非常棒。
44:11
Where your AI agents fit, you know, and this was a big workload reduction and a better outcome in both of these cases.
你的AI agents适合放在哪里，你知道，在这两种情况下，这都是大幅的工作量减少和更好的结果。
44:18
Typically, at the end of the month, we will have a big meeting between finance data operations to say, okay, let's make sure we got all the wash counts right.
通常，在月底，我们会在finance和data operations之间开一个大会，说，好，让我们确保所有wash counts都是对的。
44:26
Just to make sure that all of the every work order is what we thought it was every, you know, all the billing goes out to our customers correctly.
只是为了确保所有的每张work order都是我们以为的那样，你知道，所有billing都正确地发给了我们的客户。
44:34
We canceled that meeting this month. There were no incorrect work orders. There were no reason to have to sit and go through everything.
我们这个月取消了那个会议。没有错误的work order。没有任何理由需要坐下来把所有东西都过一遍。
44:41
We were talking about five, six person hours to go through that meeting and instead kind of just went off and did what we should be doing, development, strategy, looking into the future.
我们之前说那个会议要花五、六个person hours来过一遍，而这次我们直接去做了我们应该做的事，开发、战略、展望未来。
44:53
So in summary, again, a very practical intention of what this presentation was we built a text to SQL agent in order for leadership to have better access to our data.
所以总结一下，再次说明，这个presentation的一个非常实际的目的就是，我们构建了一个text to SQL agent，以便leadership能更好地访问我们的数据。
45:02
Also to alleviate the load off of our analysts of answering every simple question that came through, that also allowed for, you know, better context for our leadership.
同时也减轻了我们分析师的负担，不用回答每一个进来的简单问题，这也让领导层有了更好的上下文。
45:12
Also better time for the analysts to go do more strategic tasks.
也让分析师有更多时间去做更有战略性的任务。
45:16
Our vision verification has been exactly wildly successful in reducing the amount of work required for everybody to go in and fix work orders.
我们的视觉验证在减少每个人去修复工单所需的工作量方面非常成功。
45:26
And then the daily digest that comes from that, right? Here's the track. Here's every work order. Here's every issue we had last night. What needs to be fixed?
然后由此产生的每日摘要，对吧？这是轨迹。这是每个工单。这是昨晚我们遇到的每个问题。哪些需要修复？
45:34
What doesn't need to be fixed? What can we do better? Upcoming projects. This is fun.
哪些不需要修复？我们哪里可以做得更好？即将到来的项目。这很有趣。
45:38
So obviously making the operators making sure that all the information that they put in is correct. So looking into more visual models of gathering that data by itself, maybe not requiring the operator to put as many things into the app is required.
所以显然要让操作员确保他们输入的所有信息都是正确的。所以研究更多视觉模型来自动收集数据，也许不需要操作员在应用里输入那么多东西。
45:52
Or not doing it at all and say, can I gather all the information from the truck from the pictures from other methods?
或者干脆完全不做，说，我能不能从卡车的图片、从其他方法收集所有信息？
45:59
And then the other one being the mission of truck valve temperature pressure settings. So again, like I said, analog in a digital world.
然后另一个是卡车阀门温度压力设置的任务。所以再次，就像我说的，模拟世界中的数字化。
46:08
There's no computer on that truck that washes the engines. It is a guy out there mechanical ball valves and switches. And that's it.
那辆洗发动机的卡车上没有电脑。就是一个人在那儿操作机械球阀和开关。就这样。
46:16
So we want to make sure that the right settings are getting put in, but again, don't want to complicate the truck. So can we use AI vision in order to ensure that all the right settings are put into the truck every night?
所以我们想确保正确的设置被输入进去，但同样不想把卡车搞得太复杂。那么我们能不能用AI视觉来确保每天晚上所有正确的设置都被输入到卡车里？
46:27
Thank you. Thanks Daniel. Thanks to everybody. We've had a wonderful journey in the AI space specifically with the prediction our team and for me and my team. So just a thanks to you.
谢谢。谢谢Daniel。谢谢大家。我们在AI领域经历了一段美妙的旅程，特别是预测方面，我们的团队和我个人都是如此。所以 just 感谢你。
46:38
Thanks, Darren. Yeah, that was great. So so wonderful for all of you. I mean, we've got a lot of practical folks here. You know what it takes to get something from idea into daily use.
谢谢Darren。是的，那太棒了。对大家来说都非常好。我是说，我们这里有很多务实的人。你们知道把某个东西从想法变成日常使用需要什么。
46:52
And Darren's done that and got that into daily use in their team's environment. It's being utilized. The value is realized that this is this is the anti, you know, the MIT study is like 95% of AI pilots fail turns out we just need more darrens, I think.
Darren做到了这一点，并且把它带入了他们团队的日常使用中。它正在被使用。价值正在被实现，这就是那个反例，你知道，MIT的研究说95%的AI试点都失败了，结果发现我们只是需要更多的Darren，我想。
47:12
All right, that's our show for this week. If you haven't checked out our website, head to practical AI dot FM and be sure to connect with us on LinkedIn X or blue sky.
好的，这就是我们本周的节目。如果你还没有看过我们的网站，请前往practical AI点FM，并一定要在LinkedIn、X或blue sky上与我们联系。
47:26
You'll see us posting insights related to the latest AI developments and we would love for you to join the conversation. Thanks to our partner prediction guard for providing operational support for the show.
你会看到我们发布与最新AI发展相关的见解，我们很希望你能加入对话。感谢我们的合作伙伴prediction guard为节目提供运营支持。
47:36
Check them out at prediction guard dot com. Also, thanks to break master cylinder for the beats and you for listening. That's all for now. But you'll hear from us again next week.
请在prediction guard点com查看他们。另外，感谢break master cylinder提供的节拍，也感谢你的收听。现在就到这里。但下周你会再次听到我们的消息。