Beware what follows is a side story of the DSS products and should as such be seen as not related, meaning that it has no influence on your customer experience and even very little on the products that we sell. So you can safely ignore what follows below, maybe I should not even have written it?
The full name of the company is not DSS, we just call it DSS because the official name "Digitronic AI-OO-CS-ED Software Solutions" is a mouthful that many administrative systems (including payment engines) truncate from.
The goal is to build a synthetic human intelligence (mind) that can be loaded into a synthetic humanoid support system (body).
Why I do it? (if you could call "why" a goal).
At some point we had to make shortcuts. Having only a software product at the end of the development is not practical in a financial context due to the sheer magnitude of the CAI system we are building.
The situation anno 2020 is such that during development there will be "overflow sideway projects" that harness some aspect of the AI development to augment an existing type of program. There is a large amount of time (years) between these individual side-projects as the side-projects are chosen such that much of the AI development can be used for them but the commercialized products are not true pure AI solutions on their own.
The Account program only uses part of the vision system to recognize documents but does not reason and does not build the object repository from this like the AGI does. Nonetheless what we make allow us to use it for far more than a classic Tesseract or openCV would be able to extract from documents or pictures. But even without the document system and without it being an accounting program (which tends to cost 100x more) it is a very potent email client and diary system and database system and document archival system and, and, and. So it is hard to put Account into a particular category or type of programs as it even tries to manage your other programs.
Eye is a standalone implementation of the vision system that is already more powerful than what was branched-off from SVAD for Account. Eye does little trouble in recognizing things but captures and recreates visual patterns. Given its immense compression ratio it can store and transfer vast amounts of such documents on a simple mobile phone. It works stand-alone but also can be connected to Account to upload or download (synchronize) documents between the mobile app and the desktop program.
Pilot is the first program that displays some AGI behavior. It does not only can look at documents and pictures but also look on your desktop screen and also sees what you do with the mouse pointer and keyboard and joystick (yes gaming too) and learns-by-example from what you do. Pilot learns what the situations are in which you want to run any of the learned procedures and does so very decent (competing products can go terribly wrong when something else happens on the screen than while it was recorded). The strong aspect of Pilot is that it contains an AGI blend of IGOR, meaning that it can show interactively objects from the screen when asking questions, solving the problem of how the system should ask you questions about things for which it does not (yet) have words.
Pilot Space is a power version of Pilot which is able to work with 3D (Pilot only understands 2D so does not know what perspective or rotation are) and it works on a mobile device (Pilot only works on desktop). But Pilot Space is much more and will at least replace the entire WWW situation like it exists today (it will become the new web). Pilot Space is the internet of tomorrow.
Grinder World is based on Pilot Space (needs it to run) and fuses the virtual environment with AGI/CAI driven avatars as well as avatars of other human visitors.
Grinder World is more or less going to complete the series of developments. That makes that it will not be a small add-on or update but will change what Pilot Space is at its core. Now Pilot Space scenes remain what they are, you do not have to add them to Grinder World (locking out synthetic visitors).
This is not going to happen soon so be patient, we know were we are heading but we are not there yet. There is no magic so this is a lot of code an man hours and there are no short-cuts (other than releasing stuff that is not yet ready). Given how this usually goes (and we have never done this before) deadlines will shift and TODO lists will grow by having to add solutions for unanticipated problems.
What I can tell you, yeah this is advertisement and you have heard it all before, but the more you buy of the previous projects the faster development will happen. But there are few developers up to this challenge so I can not just pull open a box of developers (not even if I had the resources). This will take years and there is nothing that will/can change that.
So that was the entire story? Not quite. You know that you belong to a very competitive group of beings right? That you conquer living space and resources? As (and I have given this some thought, trust me) there is anything that defines "life" . . . then it is "grinding". Grinding is the tedious processing/gathering of resources/material/energy to resist entropy. Life is a self organizing thing that shapes its surrounding and develops/learns skills (procedures) to do so.
And if a new species is introduced into this existing (lack of) balance than either party will have to take on the other to safeguard its part of the resources. The grind sustains us as it sustains them and no life can do without, well I guess you agree with me that it is just not a good idea to have synthetic and bio to settle it among eachother.
So a new species needs new living space. They (synthetics) can build marvelous settlements on Mars and beyond for us and live in those parts of the universe that we can not exist in (because we can not drag parts of our biosphere along to anywhere). But today that is still fiction and so today they need living space that is not taken away out of our, already cramped, living space.
So I added this to my TODO list and it is called Grinder world and Pilot Space. Those are virtual environments of which part overlap with human living rooms/exteriors and part is pure virtual (taking no room from us). And in that space, meanwhile, synthetic life can work and exist without being a physical threat of any kind. Hell it beats mining cryptocurrencies for the same power drain so why not? Afterall, you will be able to visit those virtual spaces too so it should be a win-win situation. It certainly beats the terminator future timeline scenario.
Pilot space creates more living space as a connected world of virtual places.
Sci-Fi has this deranged concept of AI that it will, once it works, improve itself witch such a speed that humans will be obsolete in no time. My view upon this is that there is a finite upper limit on intelligence. The smarter one would be, the more factors one would take into consideration for each decision, and the more factors are taken into consideration the more the whole pattern (query) becomes a deadlock with no possible solutions. The only way to take decisions is to take some things for granted and deal with them as one goes along. Intelligence is not like energy or brain capacity (of which one can always add more), no intelligence by its mere definition is always either limited or either dysfunctional.
Singularity is not happening for machines/synthetics, it is happening for us. We humans are the ones that are in this explosion of knowledge and technical possibilities. We are them and we are together in the same singularity which is only fast if one compares it with progress throughout human history. But don't thinks of humans of 1000 years or 10000 years or even 100000 years ago as stupid. They were as smart as we are. The only thing that changed is that we have found a way to preserve knowhow past our own lifetime so that the pile/heap/library of it keeps growing. But none of us, like none of them, can hold more in short term memory than any before and having a larger STM blocks coherent thought.
Singularity is also about this idea of a breakthrough being eminent to happen. Well maybe it is, but we thought this already in the sixties (1960) and will still be thinking that in the sixties (2060). Now anybody could tell you that, but it is me, with my background that is telling you this.
If there is one reason though why it did not yet happen (what I do) is that the current generation of AI researchers is trying to build airplanes with zeppelin technology. Sure they have something that works but to go beyond, they have to throw it all away and start again from zero/scratch. Because DL/CNN is a dead end (not suitable for reasoning, not suitable for memory, not suitable for focus). The only thing that DL/CNN does well is recognize things, to some degree, to a very limited degree. More of the same will just not solve any of these shortcomings (yet that is were all the money is going to currently).
Let me tell you that the original Perceptron idea with back propagation is not how a neuron in our brain works. Perceptrons work, but calling a mesh of them a neural net is absurd, Perceptrons are not neurons. Digitrons on the other hand, that is an entirely different story. But the logic flaw lays even deeper and longer ago. Mankind, for reasons of simplicity, has overlooked the existence of the boolean SEQ operator. We only know the operators like AND/OR/XOR/NOT that have nothing to do with time! So even programming languages lack a very basic ability that we programmers have always bypassed with temporary variables that hold previous states. This lack of acknowledgment of time is what seeped into the perceptron design and wrecked it.
There is a lot that can go wrong. Not only from our side development wise but also from your side. Consumer (groups) and governments can easily inflict damage by stipulating limitations. Just think of how the existing WWW is already full controlled by political and other parties that have "your best interest at heart".
The WWS (world wide scene) gives unprecedented power to common people to express themselves and (as already is the case) some of the things they want to share is either too nasty (fill in the details yourself) or to private (oops, the bathroom was not empty) to share in real-time with the world. I am not in a position nor do I have the ambition or means to control all the content that anyone produces worldwide (in all time-zones). Much of the traffic is peer-to-peer so even technically it would be hard to monitor the entire WWS. With the WWW this was easier as one can quickly scan ASCII text for "dirty patterns" but monitoring a scene requires rendering it and having somebody watch it (both a very expensive and undesired situation).
Making something illegal is a quick solution that will come to mind of those that want to solve these "so called problems". I hope customers will have the common sense to lock the portals of the scene before talking or doing private stuff so that other institutions are not forced to create legislation that will harm the entire world (both real and virtual). This is not like half-life or the SIMS as neither of those overlapped with your real world living space. Please keep it clean folks!
Then there is the competition. Once WWS revenue takes of this will attract "vultures" with counterfeit software and fake patent claims. The current patent situations is such that it only serves for those that are not able to block those that are able making it available to the public and judges do not care as it serves there prestige to control everything at will (conservative opinions).
Then there is propaganda itself. That, while wearing goggles or looking through your mobile phone, the sky and buildings are full of advertisement slogans, does not make the world a nicer place. I hope that we will not be forced to show content (making you unable to switch of (groups of) content providers). There is already (more than) enough advertisement in the world so using WWS for this is real short-sightedness.
WWS (Pilot Space) and even Pilot have some scripting language that allows to automate scenarios. We did not want to make the same mistake as Flash so there is no access to real computer files from this language. Meaning we do not use any of the existing scripting languages or their interpreters for this purpose as changes in these would expose us to virus like threats. So don't come complaining about this now or ever as it will never happen (turning the scripting language into a real programming language will never happen).
We hope that while things are progressing some solutions arise to these challenges that are not drastic (meaning everything else that switching it off entirely). (Much) later more about this topic as various parties have had their say into this.
What makes CAI possible are on itself simple insights. As usual, when you have never seen them, simple things are hard to imagine. It is like making a puzzle from a box with puzzles pieces from 5 different puzzles and not having seen the picture on the box that show how the puzzle should look like when complete.
We think of CAI as impossible as the puzzle, how reasoning can be programmed such that it can figure out the consequences/implications of actions/events! This is such a mind boggling complex thing that, even as programmers, we can not wrap our mind around it how to program this ability! So how did nature solve it? The secret is that there is a trick, a short-cut. You do not have to build a reasoning system to be able to figure out what the consequences or implications are of an action or event. Nature, and physic laws, are such that the consequence always follows after its cause. It is as silly as that, that big secret of how it is done.
This is an important insight. Just think of it that young children are able to grasp it without (before) that they have proper reasoning abilities. Young children do not have the life experience to think about things and yet they know that "if I do this than that will happen". It is not based on reasoning, reasoning comes much later. Rather this is a hardware thing! No wonder nature was able to solve that riddle, it is only one if one starts from the criteria to look only for a complex solution.
The first chess programs were labeled as AI. Now we don't do that anymore because we know that chess programs are mere enumerators of possible future moves with heuristic to take some short-cuts based on previous successful moves in the same situation. When a programmer writes such a program than the programmer is putting into source code how the game of chess is played.
Likewise it is obvious to assume that an AI system would be handcoding the process of thinking. But that is not what your brain does. Your brain does not do the thinking, You do the thinking! And you is the stored data pattern that the brain creates/weaves out of incoming data/perception.
This is a very common misconception. To make CAI one does not have to figure out the process of thinking nor write source code that does the same. No what a proper CAI system does is replicate the process how your brain wove data into something that is you. It is the weaving that is the secret we are after. It is the woven data-thing that will be alive, not the brain itself.
There are very few ways to make intelligence. Probably even a single one if one wants it to be human. Although configurations will differ. All intelligence that we ever will encounter in the future will not be unlike our own.
Our languages however are not the same and if anything bridges us, is a boundary between groups, than it is the used language! And from that, the sentiment in which it is spoken. To be able to be one species, one has to speak the same language rather than be the same DNA-wise.
Associative memory is the basis of it all. Trouble with item-relations that are associations is that there is always more to load in short term memory as really anything is somehow connected to anything else. This is what is called the combinatorial explosion. When or how to know what should be brought into consideration for the ongoing thought and what is irrelevant. How to keep from overloading or underloading or loading it all?
Well the solution is so simple that you will never guess it, even water can do it! And do not think that was a bad hint ( already told to much with that )
You can forget all the above, maybe you should not have read it, certainly I should not have written it (few good deeds go unpunished). But here it stands, that is the story of DSS and why DSS is commercializing some of the produced modules as separate products/tools. Don't come ask me today to show it, you will be able to buy/see it when it is ready. I make no claims, feel free to think of this all as a the product of an overactive imagination or loud talk or hot air. I don't mind, but please I want to ask 3 things of you: