In his imagination he saw the first machine absorbing matter from the planet, converting it, working it, and sending it around that half-twisted loop that no human mind could comprehend. Soon there would be two identical machines standing n the rock, their legs waving in an explorative fashion. And then four of them, then eight… ― Tony Ballantyne Recursion

What if that is where we are going wrong? We search for signs of other intelligent life, yet maybe none exists. Maybe there is no other intelligent life. There are other ways of evolving. ― Tony Ballantyne Capacity

Until now, the densest single unit of information processing on this planet was the human brain ― Greg Bear, Blood Music

Opponents replied that when you modeled a hurricane, nobody got wet. When you modeled a fusion power plant, no energy was produced. When you modeled digestion and metabolism, no nutrients were consumed – no real digestion took place. So, when you modeled the human brain, why should you expect real thought to occur? ― Greg Egan, Permutation City

He tried to dredge up the familiar, comforting truths: The Copy would survive, it would live his life for him. This body was always destined to perish; he’d accepted that long ago. Death was the irreversible dissolution of the personality; this wasn’t death, it was a shedding of skin. There was nothing to fear. ― Greg Egan, Permutation City

How did I ever live before I had an AI in my brain? ― Elizabeth Bear, Scardown

He’s not Richard Feynman-just an artificial persona, a program meant to mimic the original. ― Elizabeth Bear, Scardown

zimboe—a zombie with no self-awareness, but programmed to claim to be aware in an attempt to deceive the truly conscious beings around it. ― Charles Stross, Accelerando

This story is not all mine, nor told by me alone. Indeed, I am not sure whose story it is; you can judge better. But it is all one, and if at moments the facts seem to alter with an altered voice, why then you can choose the fact you like best; yet none of them is false, and it is all one story. ― Ursula K. Le Guin, The Left Hand of Darkness

In fact, while we read a novel, we are insane – bonkers. We believe in the existence of people who aren’t there, we hear their voices, we watch the battle of Borodino with them, we may even become Napoleon. Sanity returns (in most cases) when the book is closed. ― Ursula K. Le Guin, The Left Hand of Darkness

“The cat’s A.I. was realistic, right down to the total lack of loyalty.”― Dennis E. Taylor, For We Are Many

“Some problems simply do not yield to reductionist techniques, particularly those that are dependent on emergent phenomena.”― Dennis E. Taylor, Heaven's River

“At one time, we thought that the way life came together was almost completely random, only needing an energy gradient to get going. But as we’ve moved into the information age, we’ve come to realize that life is more about information than energy. Fire has most of the characteristics of life. It eats, it grows, it reproduces. But fire retains no information. It doesn’t learn; it doesn’t adapt. The five millionth fire started by lightning will behave just like the first. But the five hundredth bacterial division will not be like the first one, especially if there is environmental pressure. That’s DNA. And RNA. That’s life. …”― Dennis E. Taylor, We Are Legion (We Are Bob)

Generative Storytelling

Exploring Storytelling with AI

Arcadiademo source

An AI-powered experiment in story telling that throws you into a Game Of Thrones-esque world filled with political intrigue, backstabbing, and violence, as a family fights to the death to be next in line for the throne.

This was written before GPT tool calls and JSON support even existed, so I first taught GPT a custom DSL, and then I wrote a parser for that DSL. This ended up saving a lot of context tokens compared to JSON, and it also allowed GPT to extend the DSL in some cool and creative ways on its own!

LLM Escape Room Helperdemo source

LLM Escape Room Helper generates REST tool/action endpoints to help create LLM powered escape room experiences. By hiding context from the LLM until it is needed, users cannot prompt engineer their way to a solution. It also provides endpoints for checking if a puzzle solution is correct. There is a hosted version at this site you can use, or you can deploy your own from the GitHub source. The only dependencies are Node and a local instance of Redis.

AI Agent Timersource

Many AI Agent platforms don't have the concept of how long they have been interacting with a user within a given session. AgentTimer solves this by creating a toolcall endpoint that LLMs can request the current conversation duration from. It is pretty bare bones right now, it only returns the current conversation in seconds. You need to provide your own unique ID for conversations, though eventually I'll get around to adding a second endpoint that can also generate IDs.

About Me

I started my career in C/C++ compilers, moved on to household robotics, then I lead the UI dev team for Microsoft Band. I founded a startup, worked on HBO Max during their big launch, and now I'm working in ML/AI. I believe that LLMs are the first paradigm shift in computing since the smart phone revolution, and I'm excited that I've had a chance to be part of two revolutions in personal computing during my career.

As a possible result of my time in Microsoft Health, all my hobbies are physical in nature, including BJJ, Kickboxing, and (previously) MMA training.

You can get in touch with me on Twitter or Telegram at TheDevlinB.

On my blog I write my thoughts on software engineering and I recount stories from throughout my career.