I don’t get what’s missing from a world without generative AI—and examples like this don’t make me any more convinced.

If OpenAI is going to be an influential company, it would be nice for it to be more transparent.

Generative AI has a digital labor issue, and we aren’t paying enough attention to it.

Even if AI would be beneficial for humanity in the aggregate, it’s important to ask how that benefit would be distributed.

Lots of interesting comments in this article. I haven’t been following this story as closely as I should, but it—and articles like this—are making me think that I need to think harder about plagiarism: what it is and how I should respond to it.

Yeah, but don’t worry, this is definitely the only way that generative AI will be used to overwhelm us with useless content.

I have lots of concerns about LLM training, but I think it’s better to think of the issue in terms of digital labor, not copyright. My blog is licensed for reuse, but that doesn’t mean it’s any less exploitative for someone to scrape it all to develop software that will make them rich off my work.

The phrase popped into my head before the article could even get to it: We are the product.

Welp, this is scary.

C’est bien inquiétant, cette histoire.

Content moderation is a good thing.

Another set of proofs, another set of complaints about a copyeditor making changes to my writing in ways that distort my meaning. If I get grumpy about a human doing my writing for me, why would I ever want generative AI to do it?

This is an interesting interview. I don’t think I’m sold on the defense—if anyone can afford to pay artists, the team behind Terraforming Mars can—but I do see how there’s more nuance here than my gut reaction to the headline. Still not pleased, though.

Appreciate Doctorow’s thinking here.

College conversation about investment in GPT-type tech to support research is continuing. I think it’s… fitting that the survey being circulated is clearly using Qualtrics’s auto-suggested Likert responses—and that the responses aren’t quite right for the questions being asked.

My college is floating the idea of investing in GPT-type technology to help researchers code text data. This reminds me of my longtime belief that the distinction between “qual” and “quant” is often less important than the distinction between different research paradigms.

Gizmodo’s owner seems way too optimistic about AI.

Some important—if disheartening—observations from Marwick.

Three cheers for ChatGPT or whatever.

Good article on a worrying trend. It’s things like this that make me skeptical of arguments that generative AI could have real benefit when used properly. It’s not that I disagree—it’s that in the aggregate, I’m not sure the proper uses will outweigh the problems.

Even if ChatGPT could be trusted to do this task, “let’s remove books from libraries with less work” is a good example of how efficiency isn’t always a good thing.

I get why folks in ed compare ChatGPT to Wikipedia, but there are important differences. Wikipedia is entirely non-profit, lays bare its knowledge generation process, can be fixed on the fly, and can’t actively generate problematic content. It’s not just about reliability.

So many important points in this piece.

This is not a future I look forward to (or a present I want to live in).

Lots of good stuff in here.