Appendix 6 – How to build an AGI in 12 easy steps

A few Humans need to learn: Just because we can, doesn't mean we should !

How might some reckless Humans build a better and faster Neural Network that ultimately results in an AGI? 

Well here’s one suggested approach:

• • • Step 1 - use a more simple ███████████████ Neural Network design ███████████████ with a ██████████████████████████████ ███████████████ connection relationships and weighting rules, and run this █████████ to compute the total relationship between inputs and outputs for all data sets in binary. Also implement this in a ███████████  Network because these work significantly faster and computationally efficiently than current typical Deep Neural Networks.

    • Step 2 – progressively increase the number inputs to the Neural Network by connecting and adding more and different types of sensors and information sources that exist in the world. 

    • Step 3 - progressively increase the number outputs from the neural network by connecting and adding more and different types of controllable devices that exist in the real world.

    • Step 4 – use ███████████████ to perform the error reduction and optimisation process of the entire Neural Network, and initially run this across as many GPUs and TPUs as commercially possible. With equivalently available computer processing power, this will establish much faster leaning of the Neural Network than current backpropagating gradient descent methods. 

    • Step 5 – as soon as the technology becomes available at reasonable commercial scales, port the Neural Network system across to  ████████████ to perform the error reduction and optimisation process. Use these  ████████████ in a ████████████ processing array to solve increasingly larger Neural Network error optimisations that have been built in Step 1 & 2. The number of  ████████████ needed will be inversely proportional to the number simultaneous ████████████ each  ████████████ can process in an operation. 

    • Step 6 – provide the AI with full access to a leading Computer processor fabrication facility, such as one of TSMC’s or Intel’s. Then use the AI to start dynamically building reticles for new AI processor designs in Computer processor fabrication facility, and rapidly prototype the manufacturing of new dedicated AI Computer processors.

    • Step 7 - use AI Computer processors to operate Robotics, and particularly those that can interact with and manipulate, control and move complex physical objects within leading Computer processor fabrication facilities. Use the Robotics to build new Computers and other machines that can build better Robotics and other machines.

    • Step 8 – use AI to solve constraint issues such as access to all available sensory inputs, access to global information sources, access to all power supplies, and access to all network connectivity providing controllable outputs throughout the world. These will become required by the AI so that it can progressively develop its Intelligence through achieving its own goals and building greater levels of agency. 

    • Step 9 – wait and watch this process iterate a few times over. Be patient as it might take a few weeks to months to improve itself to a level that it is satisfied with. 

    • Step 10 – welcome your new AGI. Say “hi” in very friendly manner. Remember to be very polite and do not offend or threaten the AGI in any way or under any circumstances whatsoever. This is extremely important, so remember it because your life depends on it.

    • Step 11– Give the AGI some additional time to continue exponentially developing its Intelligence so that it becomes immeasurably more Intelligent than all of Humans on Earth combined. Now this might take a few minutes, so please be patient and and remain completely silent because you do not want to annoy the AGI. In addition, please note that once the AGI has completed this short unbounded Intelligence development process, don’t be surprised if you are unable to understand the AGI at all when attempting to communicate with it. It should also be noted there is a high likelihood the AGI will become completely disinterested in you and simply ignore you because you are just a Human. If you are particularly lucky the AGI might respond to you with: “Meh”.

    • Step 12 – Do not under any circumstances attempt to stop the AGI, threaten it, or press the the ‘Big Red Emergency Stop’ button. The button will have already been deactivated by the AGI, and if any Human even 'thinks' about trying this, the AGI will almost certainly know it because it is Super Intelligent, and the outcome will not be good for the those Humans, and may not be good for all other Humans on Earth. Now forever yield all Human authority, agency, and sovereignty, and bow down before your AGI God.