diff options
Diffstat (limited to 'src/blog/one-author.html')
-rw-r--r-- | src/blog/one-author.html | 89 |
1 files changed, 89 insertions, 0 deletions
diff --git a/src/blog/one-author.html b/src/blog/one-author.html new file mode 100644 index 0000000..374ad88 --- /dev/null +++ b/src/blog/one-author.html @@ -0,0 +1,89 @@ +title: One Author +date: February 10, 2023 11:51 +--- + +<p> +Microsoft just <a href='https://news.microsoft.com/the-new-Bing/'>announced</a> +their ChatGPT-powered Bing experience. While Google is worried and scrambling to +catch up on the new space race, we as humans have even more important things to +worry about than our bottom line. +</p> + +<p> +From a purely technological perspective, these AI developments look incredible. +We're getting to the point where computer assistants we know from sci-fi films +are becoming an every day reality. +</p> + +<p> +But what are the wider implications of this? +</p> + +<p> +There's tons of potential societal outcomes this could lead to, starting with +kids a few years from now wondering why Tony Stark is the only character with an +all-knowing AI assistant in Iron Man, all the way to a full singularity event. +For this article, I'll focus on just one aspect: Internet content. +</p> + +<h3>AI content on the web</h3> + +<p> +For some time already, there's been a ton of AI-composed or -aided "content" on +the web. You may have heard others (e.g. +<a href="https://www.youtube.com/watch?v=N8P6MTOQlyk"> + Luke Smith</a>, <a href="https://jacobwsmith.xyz/stories/human_writing.html"> + Jacob Smith</a>; no relation between the two as far as I'm aware) complaining +about the generic SEO garbage sites that just produce tons of generic, +search-engine friendly articles on commonly searched topics, just to farm clicks +and ad views. +</p> + +<p> +More recently, as various advanced language models started being published, +people started speculating that eventually AI will start replacing even the +higher-class content. YouTuber penguinz0 +<a href="https://www.youtube.com/watch?v=3iQ8RXhkNwQ"> claims </a> that, in his +opinion, a bit of playing around with ChatGPT produced a better game review than +a popular game journalism website. BuzzFeed +<a href="https://www.buzzfeed.com/jonah/our-way-forward"> + openly stated +</a> that they will start publishing AI generated articles. +</p> + +<h3>One author</h3> + +<p> +While people have been suspecting that <i>eventually</i> more and more content will +be generated, Microsoft is now straight up recommending you do it <i>now</i> (see +minute 37:30 of the press conference, linked above). +</p> + +<p> +Composing individual social media posts might not seem like that big of a +deal. But it is if you take into account how media builds upon itself. Much +content is created by citing, commenting on, being inspired by other content. +And AI is itself specifically suited to quickly generating this sort of +derivative material. As the density of AI generated content increases, its +<i>rate of increase</i> will start to grow non-linearly. +</p> + +<p> +Just imagine. A Wikipedia article cites that so-and-so said something on +Twitter. That The Guardian reported that this or that happened. A leaked email +from so-and-so revealed that... But all those sources were written by AI. And +maybe the article itself was composed with the help of Bing's compose +functionality? +</p> + +<p> +Then you search for something relevant and the Bing bot reads the article and +provides you an answer based on it. Maybe you're smart enough to double check +there are "real" sources to corroborate the bot's answers and it's not something +it made up, so you find the Wikipedia article yourself, full of external +citations, and are satisfied. +</p> + +<p> +But fool you are, the internet has one author. +</p> |