Look, I know I’m predisposed to not like any new AI product, but this seems horrifying. Gift link.

Ugh, this article makes it sound even worse.

This is horrifying. Reading the headline is one thing, but reading some of the details is stomach-churning. I’m not a lawyer, and as disgusted as I am with this, I don’t know what legal liability should look like here. I feel more comfortable describing this as ethically bankrupt, though. I think I would have many fewer concerns about generative AI if it weren’t a platformized consumer product. Whatever the right legal response to this is, OpenAI has some moral responsibility for this sort of thing.

À mon avis, l’expertise « au niveau doctorat » est surtout lié au processus de connaître et non à la connaissance elle-même, et l’IA générative ne respecte pas du tout ce processus.

This seems performative to me, and this paragraph gets at why I think so: AI companies are increasingly pushing into education — perhaps in part to try and fight the reputation that AI tools have acquired that they help students cheat. Features like Gemini’s guided learning mode and ChatGPT’s similar study mode, which was announced last week, could theoretically help with actual learning, but the question is whether students will want to use these modes instead of just using the AI chatbots for easy answers.

There are, of course, social benefits to open sourcing powerful tools like these ones. However, I’m reminded of “open source” Android, which is a deliberate business decision that benefits Google—and of how many NCII-generating tools are based on open weight/open source models. gift link

Dammit, am I going to have to stop using Canvas?

Karen Hao’s Empire of AI really emphasized for me how much stock is being put in AGI—especially as a motivator for AI companies. I am fine wirh concepts being hard to define, but I do think things get tricky when you can’t articulate how you’ll know when you’ve met the goal that serves as your raison d’être.

It feels like it’s Big Tech’s world and schools are just living in it.

This is a good book, with a powerful thesis and a great epilogue that ties things together. It isn’t perfect, but I think most of my quibbles are related to the subject matter and the genre. It’s hard to write a book about a contemporary subject of such importance, and I think it’s tricky to write a book that combines history with more of a critical take on the AI ecosystem.

I was already planning to voice skepticism about Apple partnerships with universities in a manuscript I’m writing, but now I’ve got this to cite as well.

Generative AI products make me mad, I don’t like them, and I’m not going to defend them. That said, if this gets framed as a copyright problem, is there any way to give Studio Ghibli (or Pixar or the Seuss estate) power to cry foul here that doesn’t also shut down fan art, parodies, and the like? I’m skeptical, and that’s why I think “labor” is the more productive—if more legally ambiguous—framing here.

I believe that scraping the internet to profit off of generative AI is ethically problematic BUT I concede that it should be fair use BUT this is still a soulless and terrible argument.

Yeah, it’s really hard to have any sympathy here at all.

Interesting stuff here. I think most complaints about OpenAI “censorship” are hogwash, but it’s still fascinating—and worrying—to see how much control the company exercises over its product.

This is a darker version of some of the thoughts I had when I first heard about the “PhD comparison.” Before you click through to the article, I also want to use this short post as a complaint that I don’t think “intelligence” is a thing—and that PhDs certainly wouldn’t be a measure of it if it were.

In a roundabout way, I think this helps demonstrate why scraping data for generative AI isn’t a question of copyright. Even when there is a legal agreement, it can still be exploitative—it’s a question of digital labor.

Oh, please no no no. I usually read a whole article before posting it, but just the first few paragraphs are giving me such a visceral reaction that I don’t know if I’ll make it through the rest. The existing tech giants already have such a hold on us, let’s please not let OpenAI in the door.

Look, it shouldn’t take this story for people to realize that OpenAI exploits others’ contributions to make its products, but if it does the trick, I’ll take it. (And this is admittedly creepier than its base-level exploitation.)

Some better, broader coverage of complaints I made in a blog post earlier this week.

I don’t think intellectual property is the way to fight back against generative AI, but it is wildly out of line for a company who profits off using other’s intellectual property to be this petty.

Today, Stack Overflow announced that it was entering into a partnership with OpenAI to provide data from the former to the latter for the purposes of training ChatGPT, etc. I’ve used Stack Overflow a fair amount over the years, and there have also been times where I tried to get into some of the other Stack Exchange sites, contributing both questions and answers. I haven’t really been active on any of these sites in recent times, but I still decided to take a couple of minutes this afternoon and follow the advice of one outraged Mastodon post: delete my contributions and shut down my accounts.

Hmm. Unsurprising but all the more frustrating for it.

Fascinating read on web crawlers and robots.txt

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

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.

I haven’t been following this debate, but Doctorow and White’s points resonate with me.

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.

Look, if an automated process could save human moderators from the awful work they have to do, I’d be all for it. I’m unconvinced that GPT-4 could do it, though.

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 a welcome step, but I’m concerned it’s an empty, distracting gesture—it certainly doesn’t solve the deeper issue.

Good focus on the digital labor aspects of this whole thing. I sympathize with Reddit for not wanting to provide free value for generative AI (this is one of the trickiest parts of that conversation), but Reddit’s users are right to balk at providing free value for the platform.

Last paragraph here is an important one: I’ve seen a lot of headlines about OpenAI calling for regulation, but it’s noteworthy that it’s hypothetical future regulation.

Lots of helpful stuff in here.