Kyligence技术合伙人祝海林：希望通过语言这种底层技术让AI和大数据更加普世化 | DIVE 基础软件大会专访

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在当前数字化转型的大背景下，数据和AI已然成为了企业生存和发展的必备生产要素。作为支撑所有上层技术的根基，编程语言近年来越来越趋向领域化，而大数据和 AI 的蓬勃发展，必然需要一个更加为之量身定制的语言，以 SQLFlow、MLSQL 为代表的这类语言的诞生正是顺应了这个趋势。本文，InfoQ 有幸采访到了"},{"type":"link","attrs":{"href":"https:\/\/dive.infoq.cn\/2021\/beijing\/","title":null,"type":null},"content":[{"type":"text","text":"【DIVE全球基础软件创新大会2021】"}],"marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}]},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"编程语言新风向专场讲师、Kyligence技术合伙人\/资深数据架构师、MLSQL作者祝海林，听他分享当前大背景下我们为什么需要一门面向大数据和AI的编程语言。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"MLSQL从无到有，都经历了什么？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"MLSQL语言是人工智能、大数据、云计算三者融合之下的时代需求产物，传统语言难以胜任这些新的需求。 "}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：首先请您介绍下现在所负责的工作吧？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"我现在在Kyligence 主要负责MLSQL开源相关的工作，目前已经构建起了一支专职的开源团队。随着团队的全职人员不断扩充，也说明了我们公司对开源投入很大，也很有信心。另外，这个团队我们还在继续招人。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：作为MLSQL的作者，您设计这门语言的初衷是什么？它诞生的背景是怎样的？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"祝海林："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"天下武功，唯“快”不破。 这个快又包含了两个层面。第一个是，我的软件够快，比如Kyligence旗下的开源项目Kylin就在OLAP层面做到了极致的快，能对大数据做到亚秒级的查询速度。第二个是，软件使用者效率快，也就是基于这个系统，使用者能够高效率地完成自己的任务和工作，原本要一周能完成的事情，现在可以两小时就完成，从而让生活变得更加简单。 "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"MLSQL的出发点是第二个，我们希望随着人工智能、大数据、云计算的发展，在这个大的技术发展环境下，用户可以用一个语言，一个引擎就能在云上、云下涵盖AI\/BigData的所有任务，比如OLAP查询、ETL处理、流批一体、机器学习等等。这个语言必须足够简单，能够让大多数人接受，这个引擎也必须足够简单，能够轻易被维护。而在MLSQL之前，用户往往需要众多的系统和语言进行组合，才能完成复杂的大数据和AI任务的，这极大地降低了使用者的效率，提高了维护成本，使得大数据和AI难以真正走入寻常百姓家。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：MLSQL的定位是一门"},{"type":"text","marks":[{"type":"color","attrs":{"color":"#40485b","name":"user"}},{"type":"strong"}],"text":"面向大数据和AI的语言，那它和其他语言有什么区别？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"祝海林："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"MLSQL语言是人工智能、大数据、云计算三者融合之下的时代需求产物，传统语言难以胜任这些新的需求。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"首先，MLSQL这个语言天然是分布式执行的，大部分传统语言比如Python、Java必须依赖框架以及新的编程范式才能完成分布式。典型的如Java的虚拟机（执行引擎）就是单机的，语言也是单机的，你必须依赖Spark 这种框架，学习它的分布式编程范式，你才能进行分布式数据处理。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"第二，传统语言对于用户来说入门太难了，原因在于它还是太“底层了”。真正效率高的是什么语言？ 是声明式语言，比如SQL，还有比如命令行。声明式语言，不用你在编程的时候关注和业务逻辑没关系的东西，你只要描述你的任务是什么，其他的比如任务如何生成，如何拆解和执行，都是底层语言引擎需要做的事情。这样才能让更多的人使用起来。毕竟，会SQL的人肯定比会Python的人多，对不对？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"第三，新时代，我们不再要求语言必须有那种纳秒、微秒级别的响应时间了。比如在做大数据分析、ETL、亦或是AI的时候，我们大体都是毫秒级，秒级或者以上其实都是可以接受的。这个时候，分布式的解释型语言就很有价值了，虽然没办法做到纳秒、微秒级别的执行速度，但是在可接受范围内，它比传统语言更易用。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"第四，传统语言都没有在语言层面提供安全能力。比如哪些语法，哪些特性可以针对指定用户开启或关闭。MLSQL在设计之初就提供了这种能力，比如用户可不可以使用MLSQL 里的load语句加载数据，能加载什么数据，加载的数据的哪些列，哪些行，要不要执行脱敏策略等等，都被内置在语言层面。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"还有就是扩展性问题，传统语言语法以及对语言引擎（VM）做扩展，是有很高的门槛的。MLSQL 执行引擎是使用Scala\/Java开发的，而且设计的时候就是插件内核，我们可以很轻易的通过插件机制扩展MLSQL的语法，或者增强执行引擎的能力。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"MLSQL就是这么一门面向大数据和AI设计的声明式云原生语言，以SQL为主体，保留了其分布式，声明式特性，同时弥补了原生SQL可编程性的不足。在保证足够简单，同时又可以高度灵活的情况下，让用户更好的玩转数据，所见即所得。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：MLSQL发展至今，经过了哪些迭代？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"祝海林："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"最早MLSQL是为了使用SQL解决ETL而诞生的，后面我们覆盖到了流，到后来，我们发现还可以做OLAP查询服务，再后来，我们发现我们还可以完成机器学习的工作，很惊喜的是，不仅可以做到，效果还相当的好。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在这期间，我们给SQL添加变量，分支语句，函数，模块等标准语言的支持，同时还拥抱了Python 语言（语言内置了Python脚本执行引擎）从而具备拥抱更大的AI生态的能力。到这个时候，我们就发现，MLSQL 已经有一个语言的完整形态了，可以解决大数据和AI领域的大多数问题。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"MLSQL的技术挑战及未来规划"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"开源还是挺费钱的，尤其是人力方面。"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：您一门编程语言并不容易，您在做MLSQL时是否遇到了技术上的挑战？这期间有没有让您印象深刻的事件？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"祝海林："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"挑战肯定是有的。一个比较典型的例子是，MLSQL涉及的技术点太多了，作为开发者来说要把这些技术点都做好，这就非常有挑战了。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"比如在开发MLSQL的过程中，为了展示MLSQL的能力，我们需要一个好的Web端，这个时候我就重新去学JavaScript\/TypeScript，使用ReactJs去开发了一套Web Console，为了能够在桌面上使用，我学习VSCode扩展开发，开发了VSCode MLSQL Lang 插件，为了能够把它更好地部署到K8s上，我学习Go语言，开发了mlsql-deploy, 为了能够支持分支语句，我用Scala手写了一套表达式引擎，从词法分析，到语法分析，到最后生成目标代码全部自己实现了一遍。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"除此之外，各种大数据技术你都要有一定的掌控，融入进来，这才能做到让大家开箱即得，这包括诸如CDC（我们支持MySQL\/HBase）、数据湖（Delta Lake）、物化视图、聚合下推等等大数据相关技术。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"印象中，倒是没有遇到技术上比较深刻的事情，但是有一个很深的感觉就是，现代软件的复杂度太高了，一个人的时间是怎么着都不够的，无论你的精力和速度有多快。除非你是做那种小而美的软件，解决一个非常领域化的问题，才可能持续单兵作战。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：那MLSQL又是如何做推广的？单纯依靠社区的力量是不够的吧？是不是在资金上也投入巨大？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"祝海林："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"我们的推广其实非常之少，开源后，如果这门语言确实解决了一些问题，并且能够给用户带来收益，那么就会有更多的用户愿意去尝试。很多用户都用了几年后，突然某天在群里提了一嘴，我才知道，竟然有用这么久的用户了。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"在我加入Kyligence前的这些年，主要时间都花在了架构设计和代码开发上，用于宣传的时间非常少，偶尔去参加一些meetup做个分享，或者给一些公司做个布道，这期间也得到了很多国内媒体的帮助，他们也愿意支持像我这样的项目，很是感谢。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"MLSQL的推广单纯依靠社区的自然生长确实是不够的，而且宣传这个东西，是需要有一定的财力支持的。比如我做官网，自己掏钱买了云主机，买了域名，然后我发现我买不起https认证，尽管我一直想搭建一个线上的体验服务，让用户可以直接体验MLSQL语言，但是考虑到昂贵的费用（大数据嘛，机器配置要求略高）一直没有买，这个时候虽然时不时会有用户会提供机器支援，但很不稳定，这对于宣传其实是很负面的。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"所以我个人认为，开源项目最好还是后面能有一家公司做支持。国外的很多开源项目虽然没有公司支持，但是有基金会支持。"},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"总之，开源还是挺费钱的，尤其是人力方面。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：您对MLSQL的未来规划是怎样的？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"祝海林："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"远景目标是，MLSQL能够成为大数据和AI的标准语言，同时实现数据的平民化，帮助中小企业非常低成本的落地大数据和AI。我们现在正在努力将其打造成一个可信赖的产品，这包括易用性、性能、稳定性、安全等方方面面。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"加入Kyligence，从朋友圈好友走向事业合伙人"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"blockquote","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"我们希望从语言这种更底层的方式，去更普世化地解决一些问题，让大数据和 AI 变得更简单，更加平民化。"}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：前些年您一直在埋头苦干搞MLSQL，那又是什么样的契机让您选择加入Kyligence？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"祝海林："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"MLSQL是我16年在北京的时候开源出来的，后面来到了杭州，经过五年的努力，MLSQL已经有一定的成熟度，也有不少小创新，在很多公司得到了应用实践，不少互联网公司也借鉴了MLSQL的很多理念。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"这个时候我觉得MLSQL如果一直作为一个个人开发者的作品，会影响它的发展，我希望它能踏入一个新的快车道。我和Kyligence CEO 韩卿先生是很多年的“微信朋友圈好友”，这个时候他经过长时间的观察，也颇看好MLSQL, 而在我看来，Kyligence 也是一家非常优秀的拥有开源社区运营经验的公司，所以我们很快一拍即合，最后2021年的时候，我以技术合伙人的身份加入了Kyligence。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：您加入Kyligence后，MLSQL在Kyligence的业务板块里发挥了怎样的作用？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"祝海林："},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"Kyligence 一直都对开源项目有着较大的投入，Kylin 已经进入 Apache 很多年，而且也获得很大的成功。总体来看，Kylin 以强大的分析能力和丰富的周边生态而备受欢迎，Kylin 其实是非常好地覆盖了 BI 这一部分，部分覆盖 ETL\/ML。那 MLSQL 则是很好地覆盖了 ETL\/ML，部分覆盖了 BI ，所以两者具有很好的互补性。第二点，我们公司也尝试从语言层面去解决大数据和 AI 领域的问题。可能以前大部分还是从应用产品或者组件的方式去切入，这种方式的缺点是解决问题更像打地鼠，出现一个问题，用一个框架或者产品去解决。我们现在是希望能够通过这种更底层的方式，去更普世化地解决一些问题，让大数据和 AI 变得更简单，更加平民化。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"InfoQ：目前MLSQL的市场接受情况是怎样的？是否达到了您的预期？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"祝海林：对于MLSQL，"},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"用过的都说好。好些用户都是三年以上老用户了，唯一不足的是，一线互联网公司很少有用的。不过因为之前在社区方面没有做太多工作，甚至从来没有组织过任何活动，属于完全放养状态，所以用户数也在预料之中。今年（2021年）开始，在公司的支持下，我已经在规划这方面的工作了。在十月、十一月，我们很快会有很多新的社区工作落地，这包括新的官网、10月底线下的MLSQL meetup、线上体验版，还有专门的公众号，知乎专栏等等。我对未来还是充满信心的，这五年技术的发展也让我愈发相信MLSQL的技术路线是走对的。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}},{"type":"strong"}],"text":"嘉宾介绍："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"link","attrs":{"href":"https:\/\/dive.infoq.cn\/2021\/beijing\/presentation\/4003","title":null,"type":null},"content":[{"type":"text","text":"祝海林"}],"marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}]},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"，Kyligence技术合伙人\/资深数据架构师、MLSQL作者，拥有13+年研发经验。最近几年专注在Data + AI 融合方向上，致力于帮助企业更好落地Data\/AI。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"祝海林老师将在InfoQ举办的首届"},{"type":"link","attrs":{"href":"https:\/\/dive.infoq.cn\/2021\/beijing\/","title":null,"type":null},"content":[{"type":"text","text":"【DIVE全球基础软件创新大会2021】"}],"marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}]},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"上分享《面向Data和AI的语言MLSQL》的主题演讲。无论是企业还是一线研发人员，都可以通过MLSQL更好地去落地大数据和AI。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"当前，数据库、操作系统、中间件、编程语言……各个行业不断变化，但时势造英雄。基础软件发展的“黄金时代”中，怎么成为行业“英雄”？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":" "}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"阿里云李三红、PingCAP黄东旭、MegaEase陈皓，还有信通院、统信软件、极狐、清华大学、红杉资本等各个行业的专家学者，一起讲述行业变化，更要明析时代大势下蕴含的无限机会！点击"},{"type":"link","attrs":{"href":"https:\/\/dive.infoq.cn\/2021\/beijing\/","title":"xxx","type":null},"content":[{"type":"text","text":"此次"}]},{"type":"text","marks":[{"type":"color","attrs":{"color":"#494949","name":"user"}}],"text":"，看看还有谁在等着你吧！"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.geekbang.org\/infoq\/3f\/3ff4b0120ac0aeeeb094cd139b4fb29e.png","alt":null,"title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}}]}