Roadmap To a Sentience Stack

original article: http://simpleactually.tumblr.com/post/53897023370/roadmap-to-a-sentience-stack

the talk: http://www.youtube.com/watch?v=a9JY8KbeUQE

Race cars don't look like cheetahs, so why do attempts at machine sentience try to look like brains? An exploration of the unique challenges and immediate options along one possible path to machine sentience.

The "Do Anything Machine" is the first component in the theoretical Sentience Stack, an open source stack of software that when put together can be configured to learn to be a sentient mind. This approach is inspired by the LAMP stack, the collection of disparate decoupled open-source components (Linux, Apache, MySQL and PHP) that were once commonly used together to make it easy to create websites. LAMP reduced the barriers preventing everyone from building great dynamic websites. It also made it possible for individual components to be swapped out or optimized for a given project allowing the needs of individual projects to push the boundaries as needed. All of these things helped to enable the explosion of growth that created the internet as we know it, and enable it to continue improving.

Although currently in its very earliest stages, the Sentience Stack project will rely heavily on Python's extensive meta-programming capabilities and deep integration into the open source community.

Transcript

1. Roadmap To A Sentience Stack
2. Who Is This Guy?
3. AI today is in a similar state to that of CS a Century ago.
4. A Theoretical Machine One Level of Meta Up
5. The highest level pseudo code ever. • Observe a problem • Model the problem • Select algorithm that has a good history with similar problems • Apply the algorithm to the problem • Run the pair • If results are good you win • If results are bad adjust parameters • Eventually abandon the algorithm and try again • Eventually abandon AI entirely in favor of brute force
6. Overview of a Do Anything Machine. • Requires • Generic Problem Statement Class • Problem Comparison Algorithm • Unified Collection of Algorithms • Storage • Lots of CPU Time • First Run • Tries all possible configurations of all possible algorithms • Second Run • Tries all possible configurations of all possible algorithms • Third Run • Compare the previous problems • Can start with best configuration from most similar problem
7. Every time we ask the Do Anything Machine a question it gets better.
8. Lets look at some code now!
9. Introducing The Sentience Stack. Solves any solvable problem given enough resources Can ask for help from or assist other stacks, or just compare notes Uses the Do Anything Machine to improve components of the stack itself Sorts through data to figure out what needs to be figured out Gets the stack unstuck through randomness and rule breaking Uses all layers of the stack to learn how to communicate Models a particular psychological approach to self preservation
10. What I'm working on right now.
11. Questions this project is focusing on. • Generic Problem Statement • How to include other types of machine learning? • Unsupervised • Semi-supervised • Etc. • Problem Comparison • Looping Analysis • CYC • Logistics • How do we store the code to best facilitate future self-improvements? • Non Relational DB • Github • Pickle • Parallelization • Can work be chunked at a sub seeker level? • Many more where these came from.
12. We can do this together. simpleactually.tumblr.com github.com/neuman/do-anything-machine @eric_neuman