So there's no sort of behind-the-scenes manipulation, either algorithmically or manually.

因此，没有任何形式的幕后操纵，无论是算法操作还是手动操作。

But without that human touch a lot of flaws show up in algorithmically generated deep fake videos.

但没有了人工修饰，算法生成的这种深度伪造视频便会出现大量缺陷。

Infinite scrolling, algorithmically selected content, cuts every two seconds.

无限滚动，通过算法选择内容，每两秒剪切一次。

The books that aren't algorithmically favored will disappear.

不受算法青睐的书籍将会消失。

It cannot algorithmically clean themselves, apparently they need someone else to do so.

它无法通过算法清理自己，显然他们需要其他人来这样做。

But algorithmically, I just looked at the left half of the book next.

但在算法上，我只是看了下书的左半部分。

(copy machine whirring) So by 2011, the majority of the YouTube homepage was generated algorithmically.

（复印机呼呼作响）因此，到 2011 年，YouTube 主页的大部分内容都是通过算法生成的。

In less than six years TikTok has weaned the world off old-fashioned social-networking and got it hooked on algorithmically selected short videos.

在不到六年的时间里，TikTok让世界脱离了老式社交网络，转而沉迷于由算法筛选的短视频。

There's no reason why Facebook, for example, can't proposethat, you know, algorithmically, that you go to local investigative reporting.

例如，Facebook 没有理由不能建议，你知道，在算法上，你去当地的调查报告。

That's 2008. But by 2011, the majority, if not all of the YouTube homepage had become algorithmically generated.

那是2008年。 但到2011年，大多数（如果不是全部）YouTube主页都是通过算法生成的。

But as long as there is an algorithmically determined feed on the site, Meta controls how much of that content we see.

但只要网站上存在通过算法确定的提要，Meta 就会控制我们看到的内容量。

If they cannot algorithmically clean themselves or conduct proper hygiene, in the U.S., concern about facial recognition and AI seems to have bipartisan support.

如果他们不能通过算法清洁自己或进行适当的卫生， 在美国，对面部识别和人工智能的担忧似乎得到了两党的支持。

Meta, which owns Facebook and Instagram, has turned both apps' main feeds into algorithmically sorted " discovery engines" and launched Reels, a TikTok clone bolted onto Facebook and Instagram.

拥有Facebook和Instagram的Meta，已将这两款应用的主要推送转变为由算法整理的" 发现引擎" ，并推出了Reels，这是绑定在Facebook和Instagram上的TikTok的翻版。

Some creators suggested using dislikes as a metric, but that comes with the obvious drawback of algorithmically deprioritizing unpopular opinions and leaving creators vulnerable to dislike harassment.

一些创作者建议使用不喜欢作为衡量标准，但这有一个明显的缺点， 即算法会降低不受欢迎意见的优先级， 并使创作者容易受到不喜欢的骚扰。

And in doing so, we also generate, algorithmically, human-inspectable, symbolic, commonsense knowledge representation, so that people can inspect and make corrections and even use it to train other neural commonsense models.

与此同时，我们通过算法生成人类可以检视、以符号表达的常识知识表示，让人们可以检查、修正，甚至用其训练其他神经常识模型。