Thursday, January 11, 2018

I Have a Dream

Yaskawa Electric’s industrial robot Motoman.
Photograph by Yoshikazu Rsuno — AFP/Getty Images

I Would Like to Build a Robot Cook

I need money to finance a robotics project. I would like to build an intelligent robot cook, one that is smart enough to walk into any equipped kitchen and fix a meal of scrambled eggs with bacon, toast and coffee, and clean up afterwards. Mainstream artificial intelligence practitioners enjoy bragging about their achievements using deep learning and other AI techniques. They want us to believe that they are making progress toward true intelligence by making machines that can beat a human expert at board games like chess or GO. Don't let them fool you. None of it has anything to do with artificial general intelligence (AGI). In spite of the hype, their brand of AI could not be used to perform the simplest tasks that humans have no trouble with. The robot cook I am talking about is hopelessly beyond their capabilities.

Raising the Money

I want to obtain the necessary funds for my project without raising suspicions from ill-intentioned parties. My current plan is to offer a device for hearing impaired individuals that would enhance foreground speech while muting all background noises, including other voices. Many hearing impaired people have trouble tuning out unwanted noises from a conversation. They don't so much have a hearing problem as a problem with their attention mechanism. This is not a problem that can be solved with ear plugs at this time. It requires significant computing resources which can only be provided by a more powerful device such as a smartphone.

Smartphone App

My plan is to market this product as a smartphone app. The worldwide hearing aids market is in the billions of dollars. Such an app could potentially help me raise all the funds I need for my robotics project. I refuse to accept any investment from third parties. Wish me luck.

2 comments:

Peter (stn1986@hotmail.com) said...

Good luck Louis.

Be aware that software is relatively easy to reverse-engineer. Especially if you write the app in Java, but C++ is not immune to the same problem. Any interested and competent actor will be able to extract the workings of your algorithm. Be careful.

Louis Savain said...

Peter, thanks for the advice. Yes, you are right but I am not concerned about it. The reason is that the released app will not include the most important part of the system, which is the learning module. Once the network has been adequately trained, the learning module can be cut out. The app will be written in C# and possibly CUDA or OpenCL to take advantage of GPU parallel processing available in many smartphones.