//------------------------------// // The Lab: Conceptualization // Story: The Storm War // by Mint_Keyphase //------------------------------// Date of creation: Nova Aetas 3, Jan 24 I trotted over to a computer, where several blueprints were shown. They are all firearms, just of different sizes: An automatic rifle, a mortar, and an artillery piece. I printed out the first one, took it in my wings, and pinned it in front of the workbench. Of course I could have just used another screen, but I just feel more comfortable with a physical one. I used some pipes as the barrel of the prototype, as those are the closest thing I can get without waiting for a shipment. Holding the pieces by hoof, and a welding torch by wing, I started putting the design together. I could have just used telekinesis and make this a million times easier, but for some reason I get a mild migraine after I use magic, probably because that I've never learned the "official" spells. As I put the pieces together, my mind started to wander. I was driven out of society because of my wings, a genetic fluke, some ponies who saw me ran away, but most started attacking me, throwing stones, garbage, even filth and in one case a blacksmithing hammer. I eventually settled down in the Everfree Forest, built my first outpost, and began expanding. The process was slow at first, but when I eventually invented a utility robot, PoniBots, progress took off. Each step previously requires its own specialized equipment, these equipment are now replaced by PoniBots which can assume one job, switch to another by just changing tools, and move when done. Before I knew it I've already built a network of secret bases throughout Equestria, outputting a huge amount of material for me to use to my heart's content. I snapped out of my flashbacks and looked down on the, now completed, gun. It wasn't elegant or intricate, but that doesn't matter as A) It's a prototype; B) Aesthetics does not impact performance; and C) I'm not putting them on display anyways, and those who see it aren't going to be talking much about it's ugliness. I went to the testing range and fired a few blanks, it worked as expected, and would hopefully remain so when mass-issued. Time to move on to the other blueprints. Date of creation: Nova Aetas 3, Jan 25 I am sick of this. I need to automate defense to at least some degree, I'm getting jumpscared by the alarms every few hours by some tiny little skirmish that just barely managed to trigger the autopanic. I need to get a system in place or I'm going to lose my mind. My first thought would be to just hardcode a response to common situations, it's the easiest to do, but I shot that down because it was too easy to outsmart once the enemy worked out the standard responses. If they have competent officers that can see the direct relation between their actions and the resulting counteraction, I would be toast if they found out an exploitable weakness. Instead, I need something that can adapt. Since I've already had an experience in AI, it isn't really too bad, all I needed to do was to put down a few thousand lines of code and "teach" it to fight. The coding is easy, but the training is much more annoying than I thought. Not only do I need to have it work consistently, but also tell it to work in less predictable way. I need to find a balance between these contradicting requirements that would (hopefully) work out the best. The first hundred sessions aren't that bad, it's even somewhat fun to see the unexpected actions of the AI; At the 200s you would be able to see the AI steadily improving over each sesion; But after 300 sessions it becomes more and more boring. At least the training after that can be done via real combat without too much risk. Even though it is still much dumber than an actual commander, it should hold its ground against smaller nuisance attacks. And alas, I can finally work in peace. Date of creation: Nova Aetas 3, Jan 26 The new artillery was finished, and this time, I can test it properly. I have thought of a new way to test new inventions: instead of building and testing it myself, I can simply send the schematics to the custom-construction factories and have them produce a few prototypes, then give instructions on test procedure and just look at the video and data gathered. This also allowed me to do multiple projects at once as I don't need to actually build the prototype myself, so I am free to move on once the schematic is done. The testing range has already been prepared, I watched the video feed and hoped for the best. The artillery roared and sent off a shell into the distance, I switched to the camera near the target and waited for the shell to land. A few anxious moments later, "Boom!", a fireball engulfed the camera, I flinched instinctively when the fireball filled up the screen. When the dust settled, the target painted on the ground is all but obliterated, leaving a charred crater on the ground. The results had lined up perfectly with my expectations, and even over delivered in some areas. I approved the artillery for production, and with a satisfying click on the keyboard, factories will be chugging out these new equipment to the frontline soon. I went on to review the footage of the mortar's testing, which has been completed yesterday. After seeing the performance and power of the huge artillery cannons, my feelings were numbed down slightly especially ad the mortar couldn't even make a decent-sized hole in the ground. But I had to remember that the entire mortar is small enough to be stuffed down an artillery cannon and fired off, so it is all but understandable that they are much weaker. THe test was conducted without any defects found, and mortars are sent off for production. The nuisance attacks seems to be increasing in frequency, just in time for the first batches of new rifles to be issued, they should be much better than what's in use right now, gotta like the efficiency of mass production. With all that settled, I started to design a new type of PoniBots, geared towards combat, and combat only.