Every time a machine catches up to us, we redraw the line around what makes us human — and call whatever's left the "remainder." This week, Jessica and Kimberly (no guest) dig into Language Machines: Cultural AI and the End of Remainder Humanism by NYU professor Leif Weatherby, and ask whether the whole human-vs-machine contest is a trap.

Along the way: why Emily Bender's "it's just intent" argument doesn't hold up as well as it seems to, why cognition and culture can't actually be separated, a 100-year-old linguistics theory (structuralism) that explains why LLMs work at all, and why the body — not the brain — might come first.

In this episode

• The core argument — Remainder humanism: defining "human" as whatever's left over once machines take a skill. Why that's a losing game (the "arm wrestling a forklift" bit).
• Team Bender vs. Team Weatherby — Emily Bender's claim that intent is what separates human language from AI output, and Weatherby's counter: intent doesn't ground meaning, the language system grounds intent.
• Form vs. function — A quick linguistics 101 detour: language isn't just words on a page, it's what those words do in context ("it's hot in here" as a request, not a weather report).
• Cognition vs. culture — The WEIRD psychology problem (Western, Educated, Industrialized, Rich, Democratic) and why decades of "universal" cognitive science findings didn't hold up outside that narrow sample.
• Structuralism, 100 years early — The idea that words get meaning from their relationships to other words, not from pointing at things in the world — and why that theory basically predicted LLMs.
• Meaning without truth — Why hallucinations are what a meaning-making system with no truth-tracking looks like.
• Embodiment — Descartes' "I think therefore I am" flipped: feeling comes before thinking, and what that means for machines that don't have bodies.
• Practical takeaway — How to stop playing defense: quit asking "what can I still do that machines can't," start asking what these systems are trained on, who's represented, and who gets to shape them.

Mentioned in this episode

• Language Machines: Cultural AI and the End of Remainder Humanism — Leif Weatherby
• Emily Bender — linguist, "stochastic parrots" / text extrusion
• Noam Chomsky — universal grammar, cognition as separate from culture
• Maha Bali — "Where Are the Crescents in AI?"
• Michael Pollan & Annika Harris — on consciousness and embodiment

Personal segment

The episode closes with a quick peach-and-pit check-in — home renovation surprises and a therapy update on sitting with feelings in the body instead of just thinking through them.

Leave us a comment or a suggestion!

Support the show

Contact us: https://www.womentalkinboutai.com/

Guest: Dr. Michelle Morkert — gender scholar, leadership coach, founder of the Women's Leadership Collective.

In this repisode (get it? re-episode?), Kimberly and Jessica sit down with Dr. Michelle Morkert to unpack the growing call from AI leaders for "maternal AI," which is the idea that treating AI systems like children we're raising will make them safer, kinder, and less likely to turn on us. Michelle walks through the difference between a gender analysis (counting heads) and a feminist analysis (asking who holds power and why), then the conversation turns to why "maternal" is a loaded, historically fraught word to hand to an industry that has never asked mothers what they actually need or wondered how mothers (or even women in general) might benefit.

Topics covered:

• Gender analysis vs. feminist/intersectional analysis, illustrated through the demographics of the U.S. Senate
• The "maternal AI" proposal from figures like Geoffrey Hinton (computer scientist and cognitive psychologist often referred to as the "Godfather of AI" and Mo Gawdat, former chief business officer at Google X. We talk about why they never get specific about what "maternal" would actually mean in practice.
• Sarah Ruddick's concept of "maternal thinking" as a non-gendered ethical stance, and how it differs from what's being proposed now
• Why Sam Altman's comment that saying "please" and "thank you" to ChatGPT costs OpenAI "tens of millions of dollars," which he called "well spent", is a small but telling data point in this conversation
• Deepfake harm and non-consensual imagery as the more urgent, material issue getting sidelined by the "maternal AI" metaphor
• Radicalization pipelines and the "tradwife" aesthetic as a case study in how "maternal" framing gets co-opted politically
• Donna Haraway's "God trick" and why tech's claim to neutrality keeps women out of the room
• Karen Hao's Empire of AI and the Indigenous-language-model counterexample as a picture of what reciprocal, non-extractive AI development could actually look like

Also referenced in this episode:

• Alison Gopnik, The Scientist in the Crib
• Laura Bates on BBC's Radical with Amal Rajan (dehumanization and algorithmic feeds)
• Allie K Miller's interview on the Mel Robbins Podcast

Leave us a comment or a suggestion!

Support the show

Contact us: https://www.womentalkinboutai.com/