AI and I Dan Shipper

An interview with best-selling sci-fi novelist Robin SloanOne of my favorite fiction writers, New York Times best-selling author Robin Sloan, just wrote the first novel I’ve seen that’s inspired by LLMs.

The book is called Moonbound, and Robin originally wanted to write it with language models. He tried doing this in 2016 with a rudimentary model he built himself, and more recently with commercially available LLMs. Both times Robin found himself unsatisfied with the creative output generated by the models. AI couldn’t quite generate the fiction he was looking for—the kind that pushes the boundaries of literature.

He did, however, find himself fascinated by the inner workings of LLMs

Robin was particularly interested in how LLMs map language into math—the notion that each letter is represented by a unique series of numbers, allowing the model to understand human language in a computational way. He thinks LLMs are language personified, given its first heady dose of autonomy.

Robin’s body of work reflects his deep understanding of technology, language, and storytelling. He’s the author of the novels Mr. Penumbra’s 24-hour Bookstore and Sourdough, and has also written for publications like the New York Times, the Atlantic, and MIT Technology Review. Before going full-time on fiction writing, he worked at Twitter and in traditional media institutions.

In Moonbound, Robin puts LLMs into perspective as part of a broader human story. I sat down with Robin to unpack his fascination with LLMs, their nearly sentient nature, and what they reveal about language and our own selves. It was a wide-ranging discussion about technology, philosophy, ethics, and biology—and I came away more excited than ever about the possibilities that the future holds.

This is a must-watch for science-fiction enthusiasts, and anyone interested in the deep philosophical questions raised by LLMs and the way they function.

If you found this episode interesting, please like, subscribe, comment, and share! Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:


Subscribe to Every: https://every.to/subscribe 
Follow him on X: https://twitter.com/danshipper 

Links to resources mentioned in the episode:


Robin Sloan: https://www.robinsloan.com/ 
Robin’s books: Mr. Penumbra's 24-Hour Bookstore, Sourdough, Moonbound
Dan’s first interview with Robin four years ago: https://every.to/superorganizers/tasting-notes-with-robin-sloan-25629085 
Anthropic AI’s paper about how concepts are represented inside LLMs: https://www.anthropic.com/news/mapping-mind-language-model 
Dan’s interview with Notion engineer Linus Lee: https://www.youtube.com/watch?v=OeKEXnNP2yA 
Big Biology, the podcast that Robin enjoys listening to: https://www.bigbiology.org/ 

We use it to find bestselling author Steven Berlin Johnson’s next project.

I sat down with bestselling author Steven Johnson to see if we could come up with a concept for his next project—using AI. The results were amazing.

We loaded 200,000 words of NASA transcripts and all of Steven’s reading notes since 1999 into NotebookLM, Google’s personalized research assistant. We wanted to see if it could help us explore the Apollo 1 fire and find relevant and surprising ideas from history that could work to explain it.


NotebookLM condensed disparate 200,000 words of NASA transcripts into readable formats like FAQs and chronological timelines.
It sifted through the material to identify the catalyst for the fire.
The model even went through Steven’s Readwise notes to find a relevant, and unexpected, story from history that we could use to explain the history and origins of the fire

If you’re a fan of Steven Johnson’s work or you’re interested in AI as a creative tool, you need to watch this episode. 

All of this happens as a live exploration of NotebookLM, and it’s a seriously wild ride.

If you found this episode interesting, please like, subscribe, comment, and share! 

Want even more?
Sign up for Every to unlock our ultimate guide to prompting ChatGPT. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:
Subscribe to Every
Follow him on X



Links to resources mentioned in the episode:
Follow Steven JohnsonNotebookLM
Steven’s newsletter, Adjacent Possible
Steven’s latest book about the rise of the modern detective: The Infernal Machine
A few of Steven’s other books:
How We Got to NowWhere Good Ideas Come FromThe Ghost MapEmergenceThe Invention of Air

Learn how the smartest people in the world are using AI to think, create, and relate. Each week I interview founders, filmmakers, writers, investors, and others about how they use AI tools like ChatGPT, Claude, and Midjourney in their work and in their lives. We screen-share through their historical chats and then experiment with AI live on the show. Join us to discover how AI is changing how we think about our world—and ourselves. 

For more essays, interviews, and experiments at the forefront of AI: https://every.to/chain-of-thought?sort=newest

The NYT’s Kevin Roose has 18 new friends—none of whom are human.

His new friends are AI personas that he made with Noma, Kindroid, and other AI companion apps. There’s fitness guru Jared, therapist Peter, trial lawyer Anna, and over a dozen more.

Kevin talked to them every day for a month, sharing his feelings, asking for parenting advice, and even using them for “fit” checks.

This isn’t the first time Kevin has had an…unusual interaction with an AI persona. A year ago, he was the target of Bing’s chatbot Sydney’s unhinged romantic affections.

Kevin has gone deeper into the world of AI companions than anyone I know. He is a tech columnist at the New York Times, cohost of the Hard Fork podcast, and the author of three books. In this episode, I sat down with Kevin to learn more about his interactions with AI. We dive into:


Why AI companions aren’t just for lonely people or shy teenagers
Why AI personas are better friends than ChatGPT
How AI companions can be used to safely explore different social contexts
The risk of young people relying on AI for friendship
The icks of AI dating and intimacy
How to use AI to articulate what you value in your relationships

This is a must-watch for anyone curious about how AI is changing the way we form relationships.

If you found this episode interesting, please like, subscribe, comment, and share!

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:


Subscribe to Every
Follow him on X

Links to resources mentioned in the episode:


Kevin Roose
Hardfork, the podcast that Kevin cohosts
Kevin’s latest book about being human in a world designed for machines
Kevin’s piece in the New York Times about his experience making AI friends
Two of the apps that Kevin used to create AI companions: Kindroid and Nomi
Dan’s piece that explains why AI writing will feel real through psychologist D.W. Winnicott’s theory
Every’s piece that explores AI companion app Replika

Nick Dobos, maker of the #1 programming GPT, on prompt-gramming with AI

Nick Dobos showed me how to ship a website with two words and a single click. 

He’s the creator of Grimoire, the #1 custom GPT for programming that has been used for over 1 million chats. 

All he gave Grimoire was two words: “coffee website.” Just a minute later, Grimoire built the website and pushed it live to the internet. It was wild.

Grimoire can do a lot more than create websites—it’s a coding assistant with 75+ built-in hotkey commands and sample projects, a guide to learning how to code from scratch, and a tool for programmers to find answers to their questions in real-time.

Before he created Grimoire, Nick was an iOS developer at Twitter. When ChatGPT came out, Nick started experimenting with it—and ended up building Grimoire. Today, he’s at the leading edge of experimenting and building with AI. 

I sat down with Nick to explore how people are using Grimoire and what it tells us about the age of programming by prompting. We dive into:


How AI is massively lowering the barriers to code


Why it’s important to solve the “blank canvas problem” that people experience while creating with AI


How AI tools can streamline your creative process


Why Grimoire has an edge over ordinary ChatGPT


The best ways to use Grimoire to code smarter and faster



This is a must-watch for coders, creative people, and anyone curious about how AI is changing the way we interact with computers.

If you found this episode interesting, please like, subscribe, comment, and share! 

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:


Subscribe to Every: https://every.to/subscribe 


Follow him on X: https://twitter.com/danshipper 



Timestamps:


Introduction: 00:00:31


How Nick built Grimoire, the top-ranked GPT for programming: 00:05:20


Ship a website with two words and a single click: 00:10:25


How Grimoire is solving the “blank canvas problem” in AI creation: 00:14:57


The coding curriculum that can take you from zero to full programmer: 00:16:30


Why Grimoire has an edge over ordinary ChatGPT: 00:23:29


Nick’s thoughts on building the system prompt for a GPT: 00:34:10


The utility of AI as a new layer on top of existing apps: 00:40:04


How Nick uses a custom GPT to unpack his emotions: 00:43:11


How to use AI to break down tasks—from programming to daily to-do lists: 00:50:35



Links to resources mentioned in the episode:


Nick Dobos: @NickADobos


Grimoire: https://chat.openai.com/g/g-n7Rs0IK86-grimoire 


Nick’s website for his experiments with AI: https://mindgoblinstudios.com/ 


AI-first code editor Cursor: https://cursor.sh/ 


Open Interpreter: https://www.openinterpreter.com/ 


Lisa Feldman Barrett’s book: How Emotions Are Made


Demo Hume, the empathetic AI voice: https://demo.hume.ai/ 

This AI can read emotions better than you can.

It was created by Alan Cowen, the cofounder and CEO of Hume, an AI research lab developing models that can read your face and your voice with uncanny accuracy. Before starting Hume, Alan helped set up Google’s research into affective computing and has a Ph.D. in computational psychology from Berkely.

Hume’s ultimate goal is to build AI models that can optimize for human well-being, and in this episode I sat down with Alan to understand how that might be possible. 

We get into:


What an emotion actually is


Why traditional psychological theories of emotion are inadequate


How Hume is able to model human emotions


How Hume's API enables developers to build empathetic voice interfaces


Applications of the model in customer service, gaming, and therapy


Why Hume is designed to optimize for human well-being instead of engagement


The ethical concerns around creating an AI that can interpret human emotions


The future of psychology as a science 



This is a must-watch for anyone interested in the science of emotion and the future of human-AI interactions.

If you found this episode interesting, please like, subscribe, comment, and share! 

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:


Subscribe to Every: https://every.to/subscribe 


Follow him on X: https://twitter.com/danshipper 



Timestamps:


Dan tells Hume’s empathetic AI model a secret: 00:00:00


Introduction: 00:01:13


What traditional psychology tells us about emotions: 00:10:17


Alan’s radical approach to studying human emotion: 00:13:46 


Methods that Hume’s AI model uses to understand emotion: 00:16:46 


How the model accounts for individual differences: 00:21:08


Dan’s pet theory on why it’s been hard to make progress in psychology: 00:27:19


The ways in which Alan thinks Hume can be used: 00:38:12


How Alan is thinking about the API v. consumer product question: 00:41:22


Ethical concerns around developing AI that can interpret human emotion: 00:44:42



Links to resources mentioned in the episode:


Alan Cowen: @AlanCowen
Hume: @hume_AI; hume.ai
If you want to demo Hume: demo.hume.ai
The nonprofit associated with Hume: Hume Initiative
Lisa Feldman Barrett’s book: How Emotions Are Made
The TV series based on Paul Ekman’s theory of emotion: Lie to Me