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.

Appreciate Josh’s eye for detail here.

If this is so great for the community, why won’t the company even identify itself publicly?

Helpful reminder that data center problems are not just hypothetical—they’re potentially local.

Closing digital divides is good, and increasing diversity in tech fields as good, but I’ve been complaining for years about computer science ed that we stop at the nobility of those goals and don’t ask ourselves about the deeper motivations behind those initiatives. So it is with AI: A more diverse field more available to all is better than what we have, but we also have to ask whether AI education is actually a social good.

I genuinely think it’s useful to remember that non-AI datacenters are also contributing to the climate crisis, but that doesn’t let AI off the hook. It’s like saying “sure, we’re spending far beyond our means, but have you considered that we’re already in debt?

This emphasis on “objective truth” further underscores the need to talk epistemology when we talk AI.

Yes, but AI will also save us time writing emails, so this seems like a fair tradeoff.

Je compte bientôt préparer un document professionnel en français. Il y a cinq ans, j’aurais demandé à quelqu’un de m’aider à corriger des erreurs. En 2025, par contre, je me demande si mes petites erreurs éventuelles feront preuve que j’ai tout écrit moi-même au lieu de me fier à l’IA.

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).

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.

Ça fait plusieurs mois que je m’intéresse aux écrits de Jacques Ellul comme base théorique pour comprendre les techniques et technologies de nos jours. En fait, j’ai déjà écrit en février au sujet de l’intelligence artificielle générative et combien l’œuvre d’Ellul semble utile pour les critiques de l’IA malgré le fait qu’Ellul a vécu et écrit bien avant l’ère de l’IA comme nous la connaissons aujourd’hui.

Je suis en train de lire son livre posthume Théologie et technique (bien lentement, il faut l’avouer—j’avais commencé le livre en mai avant de devoir recommencer il y a quelques jours), et je trouve qu’il y a plusieurs passages qui me semblent utile lors des débats actuels au sujet de l’IA générative.

Voici des rappels importants.

Voilà pourquoi il faut parler de la théorie de connaissance quand on parle de IA:

On est passé d’un moteur de recherche à un moteur de réponse. C’est-à-dire que les algorithmes proposent des versions rédigées à partir des données qu’ils auront collectées sur Internet, puis reformulées sans que vous ayez rendu visite aux sites contenant ces éléments de réponse à votre requête.

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.

The sheer hubris of this attitude! The AI industry must exist, even if it means that it will put others out of business, and therefore any moral standard that would put the AI industry out of business must be abandoned. Very Ellulian.

I somehow missed a lot of this history, so I’m bookmarking this for future reference.

Yesterday afternoon, I was explaining (poorly) to some friends that I had been thinking about what the story of the Liahona in the Book of Mormon has to teach readers of that volume of scripture about (generative) AI. So, that connection was naturally on my mind when I was reading more of Jacques Ellul’s Presence in the Modern World over breakfast.

I continue to be pleasantly surprised by how relevant Ellul’s writing feels for today. Presence in the Modern World was first written in 1948, and even if you consider that the translation I’m reading is based on a 1988 second edition, that’s still enough time to earn the description “prescient.” (I’ve been reading Ellul in a mix of translated English and original French, depending on what’s more practical for the book in question.) Here’s a passage that particularly stood out to me this morning:

I hate everything in this article.

Data centers are coming to Kentucky, and that has me worried.

Interesting point from Masnick.

That this is coming out of Kentucky only makes me more upset.

Chilling read. gift link

One thing this article misses is how often CAPTCHA has been used to train AI. It’s always been playing both sides against each other.