Futurists like Ray Kurzwiel and Elon Musk predict that in the future, AI and humans will interface so seamlessly with digital computational devices that computer algorithms will feel as if they an inate part of the human mind.
Daemon.life’s goal is to develop the algorithms that humans will integrate with and to embody those algorithms within a robotic avitar.
This concept of a daemon was taken from the book series His Dark Materials by Phillip Pullman
Before daemon.life, I built several robots including this incomplete humanoid model
The concept of daemon.life started when I began building an AI personality for my BB8 robot.
Eventually, the AI personality became divorced from the BB8 robot itself and the project took on a life of its own and I had launched a kickstarter for the project.
I tried to collect my thoughts on why the kickstarter failed. At first I thought it was because I was using a lot of APIs and was not doing speech synthesis and speech recognition using my own neural nets. I had learned neural nets in college and used them in small gardening robot simulations, but those used simple fully connected archetectures and standard backpropgation. I sat down and went though Stanfords Conv Net course and did most of their assignments to gain a more in depth understanding of machine learning.
In the end, I ultimately decided that the reason my kickstarter failed was not that I didn’t have enough home made deeplearning, but was rather because I did not embody the AI into a physical product. If I was going to develop open source AI, I needed to put it into a physical computing device so that I have a package to sell it in.
I developed a very simple robot with a single axis of rotation that can make a laptop with a webcam follow a person when they move out of frame (left or right)
I put that project up on my website and I actually made me a few dollars, just by selling open source hardware files.
At this point, I pretty much knew what my business model would be: build a simple robot to install my AI software into and then sell the robot to tech enthusiasts who are interested in having a personal AI.
I began development on my next robot.
At first it was going to be a simple evolution of my swivel bot
However, I ended up taking my design skills as far as I could take them and built a robotic dragon with a cellphone for the CPU and face.
The dragon robot was simply too expensive. I needed something cheaper
I began work on a spider robot, but it was too complex and became a dead end.
Added a structure sensor to the spider. pic.twitter.com/TJKRoZoDmp— Collin Bell ⓥ/⚛/Ƀ (@SlightlyCyborg) June 15, 2017
On July 14, 2017 I had my first significant success. I built motor mount that can move around the popular Structure Sensor on a 2d axis.
The launch netted me $228 in revenue, $96 of which is in donations for open sourcing the project.
2 people bought the product.
Again, it was simply too expensive to build the whole product.
The structure sensor costs > $300. This means that I could only market the device to people who already owned the structure sensor.
Too damn cute pic.twitter.com/w47LxJc6CQ— Collin Bell ⓥ/⚛/Ƀ (@SlightlyCyborg) July 17, 2017
This robot got >3000 upvotes on reddit
There are several things going for this robot as a medium for distributing daemon.life to users.
I have not launched this product yet, but plan on doing so via a YouTube video and ,in the spirit of the 90’s, an EBAY listing.
I recently moved to NYC without a plan, except to meet interesting people. I posted an ad on craigslist as a mechetronics engineer and found BJ Henrici who is the great uncle of Alexis Ohanian (the founder of reddit). I promised to do 80 hours/week of work for him as a product manager for the next 6 months. I am developing a laser inspection & microscope inspection system of synthetic fiber producing machinery.
The future of my goals are as follows: