Add post about disallowing AI robots

This commit is contained in:
Devon Campbell 2023-08-14 18:00:14 -04:00
commit 1f98d50d2d
4 changed files with 66 additions and 11 deletions

View file

@ -8,6 +8,7 @@
:root {
--accent: #890b63;
--accent-alt: #6b2957;
--code-color: #632d55;
--text-over-background: #dcdcdc;
--fancy-font-stack: BitBold, monospace;
--accent-over-background: rgb(62, 249, 138);
@ -17,6 +18,7 @@
:root {
--accent: rgb(62, 249, 138);
--accent-alt: #06cb56;
--code-color: #ddfde3;
}
}
@ -78,6 +80,10 @@ a:visited:hover {
text-shadow: 0 0 2px #fff, 0 0 8px var(--accent);
}
code {
color: var(--code-color);
}
body > nav {
align-self: start;
grid-row: 2;

View file

@ -0,0 +1,30 @@
---
title: How to Stop Feeding AI
layout: article.njk
date: git Last Modified
tags: post
excerpt: It's now possible to tell OpenAI's bot not to crawl your web site, so I did it.
---
OpenAI is a 30 billion dollar company whose value comes not just from the language model that allows it to write convincingly but also from the limitless text it was trained on. You and I wrote that text. I've kept watch on the mailbox, but I haven't yet received my stock options. As a result, I'm taking the step to ask OpenAI not to train on this web site.
You can too! All it takes is adding a file to the root of your web site called `robots.txt`. As the extension implies, this is just a plain text file. Add this to it:
```
User-agent: GPTBot
Disallow: /
```
Heck, while we're in here, why not just go ahead and block Google and Microsoft too. I doubt I'll ever rank since I'm not doing any SEO, but I'll be proactive on the off chance they decide they will offer any of my content as an answer directly on their results page.
```
User-agent: Googlebot
Disallow: /
User-agent: Bingbot
Disallow: /
```
I'm not depending on these mechanisms to be discovered anyway, so there's really nothing in it for me.
Let these robots feed their insatiable hunger on someone else's content.

View file

@ -0,0 +1,29 @@
---
title: How to Stop Feeding AI
relativeUrl: /blog/how-to-stop-feeding-ai
date: git Created
tags: feed
---
OpenAI is a 30 billion dollar company whose value comes not just from the language model that allows it to write convincingly but also from the limitless text it was trained on. You and I wrote that text. I've kept watch on the mailbox, but I haven't yet received my stock options. As a result, I'm taking the step to ask OpenAI not to train on this web site.
You can too! All it takes is adding a file to the root of your web site called `robots.txt`. As the extension implies, this is just a plain text file. Add this to it:
```
User-agent: GPTBot
Disallow: /
```
Heck, while we're in here, why not just go ahead and block Google and Microsoft too. I doubt I'll ever rank since I'm not doing any SEO, but I'll be proactive on the off chance they decide they will offer any of my content as an answer directly on their results page.
```
User-agent: Googlebot
Disallow: /
User-agent: Bingbot
Disallow: /
```
I'm not depending on these mechanisms to be discovered anyway, so there's really nothing in it for me.
Let these robots feed their insatiable hunger on someone else's content.

View file

@ -7,14 +7,4 @@ date: git Last Modified
**Last updated {{ page.date | formatDatetimeForHumans }}**
Added an [about page](/about) with some information about why and how this site was built and some cool buttons you can use if you want to link to me. Someday, this may be the home to my links out to other sites and potentially some web rings? Not sure yet.
Just shipped minor updates to the YouTube channel list and the internet video list. All the lists are relatively new, so check those out if you haven't already:
- [better alternatives to popular things](/lists/better-alternatives-to-popular-things) (warning: very opinionated; feel free to make your own list if you disagree)
- [my YouTube subscriptions](/lists/youtube-channels) (warning: massive)
- [my favorite internet videos](/lists/favorite-internet-videos)
We've watched Reddit and Twitter fall over the last few months. OK, they haven't really fallen, but they've taken steps that the nerds among us who pay attention to these sorts of things don't especially care for. Up until now, I think it would have been easy to write these off as one-off occurrences. Truth is, I think we are in the era of social media deciding it's time to pay up.
In my latest post, I share [some of the next services that will probably come to collect and alternatives you can switch to right now](/blog/which-dominoes-fall-after-reddit). Get ahead of the game!
I've asked OpenAI to stop ingesting this site to train AI. [Try it on your own site!](/blog/how-to-stop-feeding-ai)