登录/注册
下载豆瓣客户端
豆瓣 6.0 全新发布 ×

豆瓣

扫码直接下载

iPhone · Android
  • 豆瓣
  • 读书
  • 电影
  • 音乐
  • 同城
  • 小组
  • 阅读
  • FM
  • 时间
  • 豆品
豆瓣读书
搜索:
  • 购书单
  • 电子图书
  • 2024年度榜单
  • 2024年度报告

Be what AI cannot do

Ivan 2020-01-04 16:29:45

Remember that you are looking for subtle, human details that are impossible to capture with charts and numbers, so try to rely on your ability to smell and feel a situation. Be what AI cannot do.

作者强调人工智能如何发展,都不能完全替代人类。提出了thick data这种数据分类的说法,与之对应的是thin data,那种隐藏在屏幕背后,用表格和数字来揣测人类行为的数据。

一家食品商利用专家系统(类似AI)改进自身生产线,用算法作出理论上最完美的汤,结果呢?汤尝起来糟糕极了。尽管是早期案例,在今天仍然适用。Complex systems have many intangible aspects that are easily dismissed with even the most computationally advanced techniques.

Being human is still pretty cool, so fuck the AIs!

作者特意提示,不要陷入以下三个思维误区:

1. Thinking like a classical engineer

and believing there's got to be only one way to build it right

2. Thinking like a classical desinger

and believing your solution is the one that all will bow down to and adapt

3. Thinking like a senior leader

and believing that what worked well in the past is obviously applicable yet again

科班出身、身居高位也许确实会令人丧失吐故纳新的勇气。


赞
转发
回应 只看楼主

> 我来回应

> 去How to Speak Machine的论坛

© 2005-2025 douban.com, all rights reserved 北京豆网科技有限公司 关于豆瓣 · 在豆瓣工作 · 联系我们 · 法律声明 · 帮助中心 · 图书馆合作 · 移动应用