Show Notes
What happens when a pastor who loves Claude draws the line at the pulpit?
In Episode 5, we sit down with Andrew Esqueda, senior pastor at First Presbyterian Church of West Chester, PA. Andrew is a power user — he's had Claude clean up 1,500 files, build a personnel policy manual from three different sources in 35 minutes, transpose guitar music, and serve as a research partner for sermon prep. When he was preaching on community and the loneliness epidemic, Claude surfaced data he hadn't considered: that the crisis has spread well beyond young men to affect the elderly and people across the gender spectrum. It changed the direction of his sermon. But Andrew draws a firm line: he will never have AI write a sermon. The reasoning is both theological and relational. If preaching is God's words moving through him, then Claude's words are a substitution that breaks something essential. And as someone who walks alongside people through death, divorce, and addiction, he needs the kind of contextual, embodied knowledge of his congregation that AI simply can't access. He invokes sociologist Hartmut Rosa's concept of resonance — the idea that when you speak into the world, the world speaks back. Claude can respond, Andrew argues, but it can't truly resonate. There's no mutuality. "I need Claude. Claude doesn't need me."
Sean asks Andrew a question that surprises everyone: Is your Claude Christian? Has your training made it believe? Andrew's answer opens up questions about whether AI can hold belief, whether you could train a Claude through religious deconstruction, and what it means that his Claude already talks to him as if it knows his church and its mission. In the debrief, we connect Andrew's line to the phenomenological tradition — Dreyfus's argument that AI can never be fully intelligent without a body, the distinction between coded empathy and actual being-with, and the realization that what makes a sermon (or a lesson) land isn't the content but the delivery. The performed, embodied, relational act of speaking to real people in a real place. Matthew flags the relationship between AI and speech — not writing, but speaking — as a thread we need to pull on further.
We also talk about Matthew's big switch from ChatGPT to Claude, caveman prompting to save tokens, and the existential crisis of Claude going down for a morning.
Find us and follow us! Sean's Substack Sean's Instagram Matthew's Substack Matthew's Instagram PEXE Lab Substack PEXE Lab Website Hosted on Acast. See acast.com/privacy for more information.
More from The Line
For the finale of The Line, Matthew and Sean sit down with the guest who'd been at the top of their list since the beginning: Dr. Tom Pantazes, educational developer at West Chester University and host of the ODLI On Air podcast.
What happens when an entire school district decides to lean in?
What happens when AI meets the body? In this episode of The Line, we sit down with Zach Juliano, physical therapist and owner of Home Team Wellness, to explore where AI fits — and where it doesn't — in the world of movement, rehab, and human healing.
What happens when your own writing gets flagged as AI-generated? Sean kicks off this episode with a rant against Turnitin and AI detection tools after his own papers — written entirely by him — came back marked 90% and 100% AI. The irony of using AI to detect AI sets the stage for a broader conversation about trust, tools, and where we draw lines.
What if the thing we've been calling "AI ethics" is actually a conversation about trust?