OpenAI 案例：Codex 如何赋能 loveholidays 全员开发
What Codex Unlocks for loveholidays
The more you can hand off a lot of the work to AI to do for you, your job elevates. Your job isn't anymore to be handed a solution and implement it. You've got something else to do that for you now. You have to get involved in the business problem. Scene 1A, Take 1, OpenAI, loveholidays. I'm Mike Jones, I'm CTO at loveholidays. I've also looked after the contact center here for the last four years too. loveholidays is a leading online travel agent.
你越能把大量工作交给AI去做，你的职位就越提升。你的工作不再是接受解决方案并实施它。现在有别的工具为你做这些。你必须参与到业务问题中。场景1A，第一场，OpenAI，loveholidays。我是Mike Jones，loveholidays的首席技术官。过去四年我也一直负责这里的客服中心。loveholidays是一家领先的在线旅行社。
We operate in eight different countries across Europe. We use our technology to process trillions, well, 60 trillion package combinations each day to help people find their perfect holiday. At loveholidays, our platform vision is to build the general intelligence for travel. Bringing together the great technology that we have with the people expertise, and Codex is basically the means that we're democratizing all of that work across the organization.
我们在欧洲八个不同的国家运营。我们利用技术每天处理数万亿，嗯，60万亿种套餐组合，帮助人们找到完美的假期。在loveholidays，我们的平台愿景是构建旅行的通用智能。将我们拥有的优秀技术与人员专业知识结合起来，而Codex基本上是我们将所有这些工作在整个组织中民主化的手段。
At loveholidays, Codex is becoming a single control plane, a single interface that is shared by engineers, by data scientists and by business. It has a lot of power through that because you no longer need to teach everyone a different tool. We codify all the best practices. We codify validations. We continuously approve it each time we notice that actually we had the wrong assumptions, or the reality has changed. This means that expertise of data engineers or cloud infrastructure engineers is fully codified.
在loveholidays，Codex正在成为一个单一的控制平面，一个由工程师、数据科学家和业务人员共享的单一界面。这带来了很大的力量，因为你不再需要教每个人不同的工具。我们将所有最佳实践编码化。我们将验证编码化。每当我们注意到我们的假设有误或现实发生变化时，我们都会持续改进。这意味着数据工程师或云基础设施工程师的专业知识被完全编码化。
So the best they know is now also available to Codex and therefore anybody self-serving their infrastructure data needs, get that expertise on tap 24-7. We want to decouple our ability to trial new ideas from actual engineering time. Our engineers developed a portal called Search Playground, which is using our design systems, it uses our front-end technologies and it has an interface to Codex. We have developed more than 10 different novel search experiences.
所以他们最好的知识现在也适用于Codex，因此任何自助服务基础设施数据需求的人，都能全天候获得这些专业知识。我们希望将尝试新想法的能力与实际的工程时间解耦。我们的工程师开发了一个名为Search Playground的门户，它使用我们的设计系统，使用我们的前端技术，并且有与Codex的接口。我们已经开发了超过10种不同的新颖搜索体验。
Majority of those were developed by non-engineers. What they provide is an ability to have an idea, materialize it into a real website that's running, share the same websites with external customers, gather their feedback, or actually launch real experiments on our website. Over the last year, AI-assisted coding has gone from essentially zero to about 80% of all of our code and that's helped us actually go a lot faster than we've done previously.
其中大部分是由非工程师开发的。他们提供的能力是：有一个想法，将其实现为一个正在运行的真正网站，与外部客户分享这些网站，收集他们的反馈，或者实际上在我们的网站上启动真实的实验。在过去的一年里，AI辅助编码从基本上为零增长到我们所有代码的约80%，这帮助我们实际上比以前快了很多。
In that time, we've increased deploys by about 73% with the headcount remaining broadly the same. We doubled the amount of changes done to our data, yet we have half as many support requests coming in. Users of our data platform can make a lot more changes autonomously, whilst data infrastructure engineers can spend more time actually improving the platform, making it smoother again or solving novel problems that we couldn't address previously.
在那段时间里，我们的部署量增加了约73%，而员工人数基本保持不变。我们对数据所做的变更量翻了一番，但收到的支持请求却减少了一半。我们的数据平台用户能够自主进行更多变更，而数据基础设施工程师则可以花更多时间真正改进平台，使其更加流畅，或解决我们以前无法解决的新问题。
Technology is just a means to an end. It's not about the technology itself, it's about the impact that it has, making sure that we're really intentional about not just giving people access to the tools but helping them solve business problems and measuring the impact that they're having. Everybody is a builder. Making changes to our applications, infrastructure and deploying code, is no longer engineering-only activity.
技术只是达到目的的手段。关键不在于技术本身，而在于它产生的影响，确保我们不仅有意让人们接触工具，还要帮助他们解决业务问题并衡量他们产生的影响。每个人都是建设者。对我们的应用程序、基础设施进行更改以及部署代码，不再仅仅是工程活动。
We have product managers, designers and commercial business stakeholders driving value and making deployments. Amazing, nice one. Thank you so much.
我们有产品经理、设计师和商业利益相关者推动价值并执行部署。太棒了，干得好。非常感谢。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力