The Wordsmith's Forge

Latest new exhibits in the LLM-Generated Garbage hall of shame

Featuring Santa Claus and reindeer.

Warning: Do not read with mouth full!

The anti-spam function on Dreamwidth has been glitchy recently. If you have that enabled on your blog, and it is out of order, then nobody will be able to comment. You might want to turn it off until the problem gets fixed.

Intro to the Web Revival #1: What is the Web Revival?

The Web Revival is about reclaiming the technology in our lives and asking what we really want from the tools we use, and the digital experiences we share. The Web Revival often references the early Internet, but it's not about recreating a bygone web; the Web Revival is about reviving the spirit of openness and fresh excitement that surrounded the Web in its earliest days. The Web Revival is not one single movement, but a loose collection of ideas and groups that fall under many names.


I heartily approve of this movement. I can't code, but I can boost the signal. So if you're involved in Web Revival, feel free to share your favorite links or other resources. Because we deserve better than enshittification.

Do you want to code your own website, or support others who do? Check out the FujoGuide.

For a while now, I've been making posts that list active communities on Dreamwidth. The weekly posts cover different topics, mostly active communities plus a few dormant that could be revived. The seasonal gather posts list all the active communities from the past three months, without the dormant ones, but adding a list of newly created communities from the same timespan. These also include the parameters for the listings.

This seems like a good time to make a masterlist of these resources. Please share widely to help people find communities they'll enjoy on Dreamwidth.

Introduction -- February 4, 2022

( Read more... )

Amazon is disabling the ability to download ebooks you pay for.  So you won't really own them, even though you pay for them.  This makes it a poor choice to find ebooks, and for authors, a poor place to offer them to readers.  Remember, if you can't save a copy to your hard drive, you don't really own it and the company can take it away at any time, for any reason or no reason.  This has already happened with other services that promise you will "always" have access to materials.  If it can't be saved to equipment you own, under your control alone, it's not secure.

MARCH META MATTERS CHALLENGE, CHECK-IN NO. 1 on marchmetamatterschallenge 

Welcome to year 6 of the Meta Matters Challenge! I'm your moderator, yourlibrarian.

These posts will be used for us all to check-in with one another, offer encouragement and answers to questions, and maybe tips we come up with as we copy our work to other locations.


This challenge runs through the month of March.  Visit the above post to introduce yourself and your goals.

A new research program is Indigenizing artificial intelligence

A new initiative is challenging the conversation around the direction of artificial intelligence (AI). It charges that the current trajectory is inherently biased against non-Western modes of thinking about intelligence -- especially those originating from Indigenous cultures. Abundant Intelligences is an international, multi-institutional and interdisciplinary program that seeks to rethink how we conceive of AI. The driving concept behind it is the incorporation of Indigenous knowledge systems to create an inclusive, robust concept of intelligence and intelligent action, and how that can be embedded into existing and future technologies.

( Read more... )

Thanks to work by nsfwords, the series page for Quixotic Ideas is now up to date.  Go check it out to find anything you might have missed. 

Thinking about Hapsburg AI and the problems caused when new AI trains on older AI-generated content made me realize a new use for AI: sabotage.  So far, it has been common for companies to train their AI using online content, and they don't give a flying fuck about consent or boundaries.

So a creative person sharing their work online could mix in some AI-generated entries, such that any AI scraping their site without permission would take damage from the inbreeding effect.  *cackle*  Just let your audience know what you're doing and how to distinguish your work from the boobytraps.  I'm sure people would have fun thinking of creative and amusing ways to do that.  And what if everyone did that?  It would rapidly become very unsafe for programmers to steal content for training purposes.

Remember this article about AI inbreeding?

There's a name for it:
Again, doubling down, the "Tech Bros" are using AI generated training data. Data science literally calls this "Hapsburg AI," after the 19th Century Austrian royal family. It's data science's anser to inbreeding.