Bill Tozier is creating software & tutorials for working with complex systems
7

patrons

$28
per month
I've worked as an engineering consultant for more than 25 years, helping software engineers and scientific teams in industry and academia plan and release adaptive systems. For a decade before that, I worked in the academic world, right at the interdisciplinary bridge between Systems Biology, Complexity Theory, and Artificial Intelligence.

Along the way, I’ve watched countless teams of enthusiastic beginners and thoughtful professionals propose, plan and launch great projects... only to freak out when things start to “go emergent”. No matter how smart you are, traditional project management and technical skills don't always prepare you for building systems that are intentionally surprising

And in the most exciting and fruitful projects, being surprised should actually be your goal. Whether you're hoping to apply genetic programming to inspire human inventors, or just want to think reasonably about self-organizing social systems—and whether you know it or not—when you undertake these projects, you're establishing a novel, active relationship between yourself and the thing you’ve made. Not only can this be technically challenging, but it can feel like an affront to a lot of folks’ habits and expectations. What does it mean to “plan” a project that's exploratory? How can you develop automated tests, when the thing you're testing is supposed to self-assemble?

I've learned a lot in my professional career. Now that most of the NDAs I (foolishly) signed as a kid are elapsed, I want to share what I know more widely. In the next few years I'll be:

  • writing essays and tutorials
  • releasing open-source software systems 
  • running workshops and tutorials
to help more people better understand the payoffs and hazards of building autonomous stuff.

With your support, I'll be able to commit to doing this full time.
Tiers
Half-a-Byte is Better Than None!
$1.28 or more per month
To thank you for being one of my Half-Byte Supporters, I'll give you access to my Patrons-only feed. I'll be posting pre-release development notes, special tutorials, puzzles and answers to Patron-submitted questions.
Bytepenny Supporter
$2.56 or more per month
To thank you for being one of my Bytepenny Supporters, you'll also have an opportunity to vote on forthcoming exercises and explorations from a list of possibilities I'll distribute every few weeks.
Kilopenny Supporter
$10.24 or more per month
I'll display your name in the Kilopenny Supporters list, and you'll have access to pre-release drafts of complex systems and genetic programming tutorials and walk-throughs, plus the opportunity to pair program with me as I work though a coding session, building one of the demos of complex systems and machine learning.
Goals
$28 of $100 per month
For the last year I've been developing a new interpreter for Lee Spector's Push language for genetic programming. It's called Klapaucius, and it's almost done now.

Unlike any other purpose-built Genetic Programming language I know, in Klapaucius you'll be able to evolve systems with an expressive programming language with many of the features you'd expect from a normal human-writeable language: collections (including ordered vectors, sets, and key-value stores), iteration and recursion, numeric types including complex numbers and interval arithmetic, abstract and higher-order functions, and more. The Klapaucius dialect of Push is extensible, and designed from the start to support modeling your domain in the same high-level objects and functions you'd use to talk about it.

In other words, this isn't just another way of doing Symbolic Regression. It's intended as a basis for evolving pure mathematics and novel computational algorithms.

The code is written and comprehensively tested, but the documentation will be tremendous. I'd like the opportunity to work on it full-time in the coming weeks.

With your support, I'll be able to devote most of my attention to that task until it's complete
1 of 1
I've worked as an engineering consultant for more than 25 years, helping software engineers and scientific teams in industry and academia plan and release adaptive systems. For a decade before that, I worked in the academic world, right at the interdisciplinary bridge between Systems Biology, Complexity Theory, and Artificial Intelligence.

Along the way, I’ve watched countless teams of enthusiastic beginners and thoughtful professionals propose, plan and launch great projects... only to freak out when things start to “go emergent”. No matter how smart you are, traditional project management and technical skills don't always prepare you for building systems that are intentionally surprising

And in the most exciting and fruitful projects, being surprised should actually be your goal. Whether you're hoping to apply genetic programming to inspire human inventors, or just want to think reasonably about self-organizing social systems—and whether you know it or not—when you undertake these projects, you're establishing a novel, active relationship between yourself and the thing you’ve made. Not only can this be technically challenging, but it can feel like an affront to a lot of folks’ habits and expectations. What does it mean to “plan” a project that's exploratory? How can you develop automated tests, when the thing you're testing is supposed to self-assemble?

I've learned a lot in my professional career. Now that most of the NDAs I (foolishly) signed as a kid are elapsed, I want to share what I know more widely. In the next few years I'll be:

  • writing essays and tutorials
  • releasing open-source software systems 
  • running workshops and tutorials
to help more people better understand the payoffs and hazards of building autonomous stuff.

With your support, I'll be able to commit to doing this full time.

Recent posts by Bill Tozier

Tiers
Half-a-Byte is Better Than None!
$1.28 or more per month
To thank you for being one of my Half-Byte Supporters, I'll give you access to my Patrons-only feed. I'll be posting pre-release development notes, special tutorials, puzzles and answers to Patron-submitted questions.
Bytepenny Supporter
$2.56 or more per month
To thank you for being one of my Bytepenny Supporters, you'll also have an opportunity to vote on forthcoming exercises and explorations from a list of possibilities I'll distribute every few weeks.
Kilopenny Supporter
$10.24 or more per month
I'll display your name in the Kilopenny Supporters list, and you'll have access to pre-release drafts of complex systems and genetic programming tutorials and walk-throughs, plus the opportunity to pair program with me as I work though a coding session, building one of the demos of complex systems and machine learning.