Monthly Archives: February 2019

The beauty of different systems coming together

说一说最近我看到的Google Maps AR里所发生的美妙化学作用。这种不同的系统(产品技术、设计哲学、以及商业战略)都完美结合的例子并不常见。

Google Maps AR的视频:

应用层面是AR增强现实,稍用了点机器学习来识别建筑物,同时谷歌街景有世界各地城市的建筑物图像。虽然这三者(AR + ML + Street View)的配合缺一不可,但私以为其中最savvy之处在于街景,因为这并不容易被其他公司复制:过去数年数个街景汽车在地球各地所采集到的数据。以往的努力完全pave the way for今天的应用。

这种有默契的配合,让我想到我之前的一篇博客Cutting a hole to make a handle里面提到的”不重造轮子“之美。利用已有资源来建造商业moat优于重新去添加或开发yet to happen的“轮子”。这种天衣无缝的感觉,就像Facebook主产品(带有Social Graph的Feeds)和其商业模式(based on Social Graph的highly targeted的广告)那样的一拍即合。对于这个,Benedict Evans或许形容得更加形象:“sharing和feed browsing是revenue-generating behaviors。”

Search for China

Interesting comments on Baidu’s downturn in China:

While most discussions about Project Dragonfly (Google Search for China) focus on ethics or privacy, this perspective seems to be missing – Search as product might no longer be relevant to China users. No?

74% of Chinese online payment is on mobile vs 15% in the U.S. (source)
98% of Internet Users in China are mobile users (source)

From the user behaviors standpoint, unlike US, it’s a totally different climate when the dominant of population are mobile-only users who have never used PC (or web). Search by nature arguably isn’t a strong vehicle in mobile; Apps are. On mobile, people don’t like searching stuff to get jobs done, but instead would browse feeds to consume whatever algorithm throw to them.

From the data aggregation standpoint, Search was born in the era of World Wide Web, and is largely beneficial from the open web. But in the mobile world, each App is an isolated island, and it became difficult for Google to harvest data (I doubt if PWA / Instant Apps only work best with Google’s own ecosystem like Android and Play but struggle to work with iOS Apps).

That said, even Search might not be the best product-market fit for China users, Assistant might have a chance to get it through. Because Assistant is a native form to mobile (or all kinds of devices). I have a hard time to imagine what a success will be looking like for Assistant though. If Assistant can succeed, I think it’s a better product than Search for China’s population.

(Disclaimer – views in the article are my own and have nothing to do with my company.)