The first idea I had for technology is the self generating neural network. This already exist, but, it isn't very good. Most people who don't know what a neural network is will normally go to college and be taught incorrectly or different then what people use to be taught when you read old books. A neural network is a logic gate system with inputs and outputs. Unlike normal logic gates, neurons have a designated number to each input coming into the logic gate. The logic gate has its own number. When the number coming in is greater than the number inside the logic gate, the logic gate outputs from all its outputs to the next logic gate. Numbers are decided by the creator and may be anywhere between -100 and 100. Inputs may also vary between -100 and 100. For a self generating neural network to actually work, you have to make software that records "table data" and applies the data to an algorithm. The table data consist of a record of inputs, but, the record of outputs is unique, because at first, there is no neural network because it has to be generated. It's generated by attaching the SGNN (self generating neural network) to simulator games that people play, like farm simulator and things of that nature, fork lift simulator, ect. It can even be attached to abstract games. What the player sees on the screen, is an input. What the player does, is an output. There are recorded for each moment by the SGNN. There are three types of inputs. The first type is recorded inputs. The second type is the output from the previous moment being treated as an input. The final type is what I like to call "anomaly input". An anomaly input is created when the currently existing inputs are exactly the same, yet the recorded outputs are different, for two separate moments. The way the SGNN generates is fairly simple. It adds neurons as needed and arranges them in various ways based on how many inputs are coming in and how many outputs exist. Normally a neuron will attach up to 10 other neurons maximum, so by default, you can have 5 inputs coming into a neuron and 5 outputs, or 9 inputs and 1 output in any arrangement, so long as 10 is the maximum. The reason for this is because this makes things easier to research and develop. When generating, if there are 6 inputs coming into one neuron, you can simply have a formula divide 100 by 6. The rule here changes up when a neural network generates an xor gate, obviously. If 2 of the inputs are activated out of six, than the number inside the neuron that decides when it sends an output, would be 2/6 of 100, or 1/3 of 100. The value of each inputs connection would thus be 1/6 of 100. This is an example of how easy it is to start creating algorithms. The purpose of this SGNN is to allow for a machine to copy exactly what someone did. One of the final factors is a factor I call "smear". Smear is a margin created by input ranges. Before I get into that, essentially an anomaly input is the reason I created smear. The way anomaly inputs work are that like I said before, when there are two inputs that are the same for two separate moments, yet the outputs are different, it creates a kind of anomaly that functions as a kind of identity for what happened. The problem is getting the neural network to decide if the anomaly is a "waiting" anomaly, a "memory" anomaly, or a "counter" anomaly. A counter anomaly is simply when someone did something first before doing the next task which was very similar, but, required deviation from the original path. The memory anomaly is when a series of inputs from the input table was ultimately responsible for the choice made, being different. A waiting anomaly is when the input was simply the same because there is a waiting point or a reason to wait and do no outputs till certain factors are met. These anomalies are important to the SGNN and Smear. Smear takes each input and breaks the inputs into margins where if an anomaly input is created, before merging the entire smear network into one, it first compares previous data with data between the last anomaly input that was created of the same type. The purpose of this, as I said before, is that by created anomaly inputs, it allows for a "single neural network" that works consistently to copy the precise task done by a human. This can also be refined through research and development, even no budget research and development, into an extremely efficient human replacement. You can use computer programs to simulate outputs for special types of machines that a human couldn't possibly control manually. The system is very simple, yet advanced. It can replace the entire workforce with enough money thrown at it and it can be made and used by anyone without needed to go on the internet and buy it. All you have to do is know how to program logic gate systems into software for practical use.

The next idea has to do with EMP devices. People use electric wires in their machines which can easily be thrown off with a single electromagnetic pulse. This wouldn't happen if electricity had a replacement. Well guess what, it does. On a microscopic scale, we could emulate electricity in certain devices that are military grade, by changing the electric current, into a meta fluid style current, if you know what I'm saying. On a tiny scale, we could replace electricity in certain devices that have a tendency to shut off during solar flares, with something that won't. My idea is using some kind of material science researched fluid, chemical reactions, or micro vacuum that could control the processes that get done by electromagnetism through pure emulation. This would be crucial in future warfare to prevent the "stopping" of certain weapons from being used by electromagnetic pulses and solar flares. This can also prevent "complete power grid failure". One day, this is going to have to be done or else certain forms of technology will no longer function properly when the earths magnetic sphere depletes and electromagnetism gets disrupted by solar flares. The problem with this idea is that on a microscopic scale, this would generate a ton of heat. I hate to admit, but, it seems like we will be using real neurons for our future computers. That would be a nightmare.