About
After the Test Subject failed to produce the desired outcome in our previous experiment, Vectromirror 0™, we have rebuilt the simulation to further our impact on the Test Subject’s emotional response system. We have discovered that, although it was able to successfully traverse the course of Vectromirror 0™, it did not unlock the emotional responses we had anticipated.
We have uploaded the Test Subject into a more advanced simulation, now dubbed Vectromirror™. Vectromirror™ will force the Test Subject to endure even further testing and rigorous trials. Like before, the Test Subjects will have to traverse an emotionally and physically challenging course using parkour and other special abilities, like wall-running, double-jumping, and the new stasis hold and throw feature we just added. The emotional stimuli we have built into this new world will indeed produce the human-like emotional state we seek for the Extrapolator AI.
Test Subjects are encouraged to experience our first experiment, Vectromirror 0™, before entering the new experiment. We have maintained a free level of access to that experience, and all Test Subjects will benefit from completing the first set of courses.
This new simulation will thoroughly test the limits of the Test Subject’s decision-making ability, as they discover multiple paths to achieve the same outcomes. We believe that with this, in addition to the refinement of the virtual world we have created, the Test Subject will finally be able to achieve human-like emotions. To ensure success, we’ve added additional features to our Test Subject, and it’s respective AI. We’ve also optimized the current feature set from the extensive testing and trials of Vectromirror 0™.
The current experiment is still in development, but we are confident that our current state will begin to show results in the two emotions we are testing: Trust and Misery. We are determined to achieve full sentience in our Artificial Intelligence and will not stop until the Test Subject has achieved success or pushed to its absolute limits.
In response to the significant testing we conducted in our previous experiment, we have optimized the Test Subject's current features as follows:Advanced Movement and ReflexesWhile our Test Subject moved exceedingly fast during initial testing, we have further optimized its ability to maintain significant speeds. We showed promising initial results, but they ultimately fell short. We’ve decided to upgrade the Test Subject’s ability to traverse walls and floors by applying Van der Waals forces of the Test Subject's feet. This allows the Test Subject to lock onto various surfaces and run across them without needing to focus too much on the speed and gravity in the specific situation, similar to what arthropod species employ. Unfortunately, we are still unable to update the Test Subject’s reflexes further, and each Test Subject used produces a different level of reflexive memory adaptability.Stasis Hold and PropulsionIn our previous test, we determined that simply manipulating objects in the Test Subject’s path, by pushing and pulling them, was not sufficient in advancing the AI’s response system. The Test Subject will still be able to create massive blasts of energy to manipulate obstacles in its way, like before. Nonetheless, we have outfitted the Test Subject with different abilities to pull objects towards them, hold them in a suspended state, and propel them in a specific direction. This will allow the Test Subject to build its paths to advance in the simulation and deactivate barriers in their way in a less… forceful manner.Accelerated JumpingSince testing began with Vectromirror 0™, we have studied our various Test Subject’s ability to utilize their accelerated jumping or “double-jumping” ability. We optimized the ability, specifically the Test Subject’s accessibility, to understand and activate this feature with extreme ease as long as the Test Subject initiates the first jump on a solid surface.Instant Memory Recall Upon DestructionOne of the key features of our prior testing was the ability for the Test Subject to reload at an earlier point in Vectromirror 0™, should they meet their demise. We have continued this methodology of instant recall to a previous point to continue testing instantly in a failure. These “checkpoints” will serve as markers in testing to determine the effectiveness of the emotional stimuli.
To make sure we achieve our stakeholder’s desired results, we will be continuously updating our experiment to test different variables. Thus far, these include:Virtual Reality Human ControlTo establish a contradicting scenario to test, we are hoping to upload human intelligence to the simulation through the use of Virtual Reality headsets and controls. The goal is to virtually link a human to the virtual world and control a Test Subject directly, feeling each achievement and failure at an instantly personalized level. We will use this to test human intelligence directly linked with advanced virtual abilities and capture the experience from a first-person perspective.Additional Level DevelopmentWe have begun this experiment focusing on two key emotions, trust and misery. The Test Subject will also be presented with a training simulation upon entering the world to learn and relearn all its abilities. We will be increasing the quantity of levels for the Test Subjects to engage with as time progresses, and their difficulty.Time TrialsAs we continue testing, we’ve noticed anomalies in our Test Subjects in that some perform much better than others, regardless of the identical features we have programmed into each. We hope to develop a ranking system to database all Test Subject activity and display the results to the general public. This will determine the best Test Subjects who have completed the trials and are primed for further testing.