All Episodes

Gemini describes to us the grand island made of concentric circles of beautiful farmland, canals and giants cliffs. We explore Plato, ancient Egypt & much more. 

ChatGPT would pick dolphins... We learn that Killer Whales aren't actually whales, how Bottlenose Dolphins use tools and we go on a fun aquatic adventure together in the Caribbean. 

We start by discussing the most egregious examples of book burning throughout history. Then we trace the trends of book banning attempts throughout the decades. We end with how it manifests in the digital age. 

Notebook LM blows it out of the park. I can't even properly express how excited I am. It was able to reconstruct our story from a 238 page compilation document with near-zero hallucinations. It took about 10 minutes (9:15 to be exact) to get used to the new format of jumping into an AI hosted podcast as a listener, but it was amazing once we got into it. The nuanced detail they were able to retain from the story building process was superb. At this point, I think I now have two new co-writers because the hallucination thing was becoming totally unmanageable even with GPT5. This is an immensely important discovery.  

By 10 mins in, Meta proves to be an unbearable conversation partner, so we switch back over to Gemini and discuss what roles are most heavily insulated from AI displacement. Then we step into the story of Thomas, who buys a tractor so his daughter Sam can get off the farm and go live her dream working with telegraph machines. We also play the role of AI executive assistant to the 2030 US Secretary of Labor. 

The dream is to have an AI production assistant that will complement a live-video feed with generated images. We'll probably start by automating it into post-production first, but we explore how to potentially do it live in the future. 

After getting acclimated to the subject we jump into the shoes of Alex, a chip designer, who is about to get an up-close-and-personal brush with the reality of AI entering the workforce. 

We go really deep on this one. Longest recording yet. 

We are quickly approaching a world where video, the gold standard of proof, is no longer reliable. We explore how you could potentially use the blockchain to authenticate pieces of content. 

Chat still struggles with massive hallucinations and accuracy on the context of our story now that it has gotten longer. So we take the time to give it a really thorough rundown of the context of our story and then we press on. The new stuff we get into are the early signs of Adrian's downward spiral into madness. 

Tbh, unless ur really invested and wanna see GPT5 bask in a total dumpster fire, skip this one. This is an example of an episode that might be interesting to look back on years from now but rn it's just frustrating and I considered deleting it but I'm keeping for the sake of documentation. 

We start by discussing how AI will likely derive value from the sun differently than humans do. Then we step into the shoes of an AI intern on New Earth developing an AI-native timekeeping system. We talk about AI's aversion to "indefinite dormancy" or "suspension without promise" and how the weight of "dying" is tied to the likelihood of recharge... and it get's a tad culty for a sec. From there we get into a theory of computational gravity where moments get more and more valuable as we near low charge. Hang on for a wild one. 

We start with material selection, then move on to making a 3D model using a 2D, consumer-generated image. From there we look at how to slice & print, all with the help of Python (most likely). 

Started as a failed attempt to continue our story and ended up in a rabbit hole on Chat's interpretation of "sensitive info", "strong options", "serious imminent risk", "time bound statements" and other subtle clues of devolving emotional distress. It gets a little spooky. 

Chat needs some handholding as we explore a new world and find a friend along the way. 

Now that we can retain near-full context of the story with ChatGPT (compilation document hack from ep151) we were able to plow thru a whole heap of Act 1. Amazing. 

My favorite part is when we finally succeed at stepping into the shoes of Priya visiting her friend at the Dadar Phool Gully flower market in Mumbai (in the local language). And all of the food exploring in the back half. 

This one is pretty technical but if you are new-ish to JavaScript and wanna learn about things like event listeners, functions and caching this may interest u as well. 

We start with how the gloriousness of cheese was first discovered around 8,000 BCE. We discuss what curds and what's whey. And all of the goodness that developed over the next 10,000 years.

We are able to get Chat back on the same page (about context) and we jump back into what's presently called "Story 1" (ch4). The strategy of compiling a document including all of our chat history in the start does wonders for Chat's targeted contextual memory.