Musk leans into the bro in tech bro.

Look, I really will stop posting about Grok and epistemology, but the news stories keep coming.

Okay, really don’t want to spend any more time writing about Grok, but let’s talk about this passage:

“With respect to academic questions, Grok 4 is better than PhD level in every subject, no exceptions,” Musk claimed during the livestream. We’ve previously covered nebulous claims about “PhD-level” AI, finding them to be generally specious marketing talk.

To return to my thoughts on AI and epistemology, I don’t think having a PhD is (or should be) a benchmark for content knowledge. Rather, I think it demonstrates (or should demonstrate) a commitment to the process of knowledge production, and LLMs cannot truly compete with humans there.

Yesterday, I wrote my thoughts on how Grok’s “Nazi meltdown” helps illustrate some of my concerns about AI and epistemology.

This coverage of Grok’s latest demo only reinforces that—Musk’s tinkering with the LLM to get the results he wants is at odds with his states naïve epistemology that an LLM can be “maximally truth-seeking,” as though there is a self-evident truth that an LLM can deliver in a straightforward way (that is, without all that mucking about behind the scenes).

Yesterday, I blogged (en français) on Jacques Ellul’s emphasis on the need for a technology-responsive ethic that emphasizes (among other values) tension and conflict. Ellul explores this ethic—one of non-power—in a few different writings that feel like different drafts of the same thing, and so I’ve seen that emphasis come up a few times as I’ve tried reading his work. Every time, it surprises me a little bit. Why, in articulating an ethical framework, would you emphasize tension and conflict?

Disgusting and deliberate.

More on why we need to talk epistemology when we talk generative AI:

Musk tweaking his AI model to be more aligned with right-wing edgelords was inevitable, but there’s a broader point to be made: each AI model is a black box that supposedly gives objective answers but in reality is shaped by its owners. As more people look to AI to learn about the world, the people who control how it’s trained and how it responds will control our prevailing narratives.

When I was in grad school, a faculty member in my program told me a story about his then-quite-young son, who was having a grand old time climbing on top of the kitchen table and then leaping off of it to the floor below. (Truth be told, my memories of this conversation are fuzzy, and the son might have been engaged in some otherwise dangerous behavior.) The father tried to tell the son to stop doing this, warning: “You could have hurt yourself!” The son’s response? “But I didn’t!” Sure, the action had been potentially dangerous, but the landing had been a success, and the son didn’t see what the big deal was.

The “You do not blindly defer to mainstream authority or media” system prompt is raising questions already answered by the system prompt. Also, lol that they have to explicitly tell Grok not to call it “Twitter.”

Aside from the headline-grabbing parts of Grok’s recent freakout, this story does a really good job of emphasizing that AIs don’t “think”… and that “truth” isn’t really a valid concept either, no matter Musk’s marketing.

None of this is good, and I think there are dangers in using copyright as the bulwark against AI. Conversely, I will take a bit of pleasure in administration infighting, especially if it gets in the way of the AI companies.

That Elon Musk would abuse his government position for his own benefit isn’t at all surprising. That he would use it to argue for an exemption to public records laws to hush up a neighborhood dispute is petty enough to catch my attention. [gift link]

Oh look, it’s all my least favorite things about tech right now, combined in a single, enraging story.

Since early 2019(!), I’ve been slowly but surely orienting my online presence around my Hugo blog. This doesn’t mean that I’ve given up on social media platforms, but that those are merely appendages to a website that I have more control over. In fact, I’m really pleased with the POSSE—Post to Own Site, Syndicate Elsewhere—setup that I’ve developed over the past couple of years. It currently works like this: All of my posts start on this website, and then I use the EchoFeed service to send posts to my Mastodon accounts and a Bluesky account (Micro.blog also imports my posts via RSS).

Every fake ad in this month’s print issue of The Onion is making fun of Elon Musk. All of them. What a great news source.

I haven’t paid much attention to gestures at all this before now, but this critique convinces me.

Creepy as hell. Gift link.

My Day One journaling app told me this morning that today marks one year since I read Cory Doctorow’s For the Win after picking it (and many other of his books) in a Humble Bundle. That means that it’s finally time to write out some thoughts that I’ve been having over the past several weeks, all centered around wanting to reread the book. I have a lot of other books on my “to read” list right now, so it’s kind of ridiculous to want to get back to this one, but there are two things that have been really pushing this idea.

This is a mess of a post blending thoughts on tech policy with religious ideas and lacking the kind of obvious throughline or structure that I’d like it to have. It’s also been in my head for a couple of weeks, and it’s time to release it into the world rather than wait for it to be something better. So, here it is:

I am frustrated with generative AI technology for many reasons, but one of the things at the top of that list is the knowledge that today’s kids are growing up in a world where it is possible—even likely—that their middle and high school experiences are going to involve someone using generative AI tools to produce deepfake nudes (or other non-consensual intimate imagery—NCII) of them. See, for example, this horrifying story from the New York Times last April.

Acknowledging the ambiguity in Musk’s salute is important: not to let him off the hook, but to recognize how much more dangerous things are when they’re ambiguous than when they’re straightforward.

Honestly, I get the original call, but I’m glad Bluesky backtracked. Content moderation is hard.

I have qualms about the U.S. providing military support to any cause, but if there is such a thing as good U.S. military support, Ukraine is a clear example of it, and either way, this is an embarassment.

Some bonkers details in here.

DOGE continues to insist on the dumbest way to carry out its dumb goals.

One more thing right out of the Twitter playbook from DOGE: Firing people without understanding how important they are.