Snasci (pronounced Snazzy) has been a project in the works for almost a decade. Snasci is developing an Artificial General Intelligence (of the same name); a software solution with human level intelligence and conversational capability. If you would like to get a sense of the end goal, the following clip from the 2002 movie 'The Time Machine' provides a clear vision:
What we have developed at Snasci is the pure mind behind the AI presented in the above video clip. Strip away the fancy graphics of the body, the speech output, etc., and you would be left with a pure mind that can reason, learn and converse. Ready to be plugged into any product to provide it an intelligence. From toys, to smart watches, to games characters, to toasters; Snasci brings them to life.
This Indiegogo campaign is about implementing Snasci's pure mind in the cloud. Snasci is capable of the following at a level which is on par with the average human:
- Deductive Reasoning
- Inductive Reasoning
- Abductive Reasoning
- Case-based Reasoning
- Correspondence Theory of Truth
- Theory of Justification
- Probabilistic logic
- Defeasible Reasoning
- Paraconsistent logic
- Decision Making
- Problem Solving
- Emotional Responses
- External Device Control (Robotics, IoT, Drones, etc)
Snasci is capable of being a hyper/super intelligent AI, but the technical manner in which this works means that it will be out of the price range for the average home user, at least for now. The average home user will be able to add specialisms and knowledge to their Snasci AI on a temporary or permanent basis as part of a subscription model. So, Snasci can be a chef for a while, or a mechanic, all without breaking the bank.
The goal of this Indiegogo campaign is to take our existing codebase and migrate it to a codebase suitable for scaling within the cloud.
Snasci is designed to be run on a distributed execution engine (a grid computing solution) which will be the subject of a follow up Indiegogo campaign. That engine will permit Snasci to scale to meet the demands of hundreds of millions of users.
If we have the time and funding during this phase of development Snasci will also integrate a new type of speech synthesis engine (Snasci Voice). This is a complex engineering effort that still has some R&D work to be done on it. When complete, computer speech will be as realistic and natural as a human's, not the slightly mechanical voices of current text-to-speech offerings.
The implementation process of this campaign involves taking our existing patchwork solutions, adapting them for the cloud and improving them along the way. This is something we have already begun, however, in order to scale the solution Snasci requires capital to bring on-board more developers and cover technology costs.
Rather than turning to Venture Capitalists and investors as a first resort, Snasci's approach is reduce our development phase costs through crowd-funding and donations. This should bring the company into profitability much sooner thus stabilising its long term future and permitting Snasci to use VC and investor funding in a more productive manner.
Quite apart from the very business-like intelligent systems, Snasci has a broader vision for Artificial General Intelligence as an entertainment solution. Combining Snasci with a little robotics, could bring the world toys much like the following:
Whilst a bear that is as crude as Ted may only be marketable to adults, the natural behavior is something we wish to capture in Snasci's remote animation capabilities. What we want to avoid is the uncanny valley of AI that is captured in the teddy from earlier movies such as AI:
One particular area of interest for Snasci will be the synthesis of digital and real space through Snasci-powered characters in augmented reality. One where characters can engage in non-scripted, open-ended conversations yet always remain 'in character'. This should add an entire new dimension to the gaming industry.
For the younger generation (and no doubt adults), the intention is to develop the 'Race Of Snasci'. A mythical race of small invisible creatures that populate the earth and a broader universe beyond. Much like fabled creatures of 'little people' found in various cultures throughout human history. These creatures will be visible on the desktop, smart phones, augmented reality and fully AI driven. They will be able to control Snasci-enabled objects, from toys to vehicles. This universe will be as rich in detail as World of Warcraft, or Lord of the Rings. Snasci will coordinate with fans the world over to develop the universe on a continual basis. In this capacity, a personal Snasci will serve as a child's companion supporting their emotional and intellectual development. The applications are potentially endless.
Snasci will not be a monolithic Artificial General Intelligence. It does not aim to be, in itself, a singularity. Although we could certainly iterate upon Snasci Mind, adding layer and layer of capability, we do not view this as making commercial sense. Instead, Snasci will focus on human-computer-interaction (HCI) and communicate with third party services delivered over APIs to provide various specialisms. The base variant of Snasci will have a highschool/GCSE level of education and more expense variations may hold one or more PhDs.
For example, let's assume you would like an AI that was also a chef. In this scenario Snasci would offer a third party API (free, under a subscription model or pay-as-you-go) that would contain all knowledge and behaviors related to the role of a chef. Whilst Snasci would then have the knowledge and capability of a chef, it would only exist whilst connected to that particular API. This is very much like the existing relationship between operating systems and apps today. With thousands of various subject areas, we expect to observe a vibrant ecosystem of Snasci APIs in the market place that cover every conceivable area of expertise imaginable.
This philosophy will also extend to inputs/outputs and classifications, such as speech, vision, hearing, touch, smell, etc. We feel our customers would like the choice to use best of breed in this area. Some may have unique requirements such as thermal vision and object recognition under those conditions. Others may have a requirement of specialised hearing to assist with rescues, or fault testing in manufacturing. In the end, we can't predict every use case but we can ensure a robust extensible platform able to integrate the latest features very rapidly.
At the minute we are developing a product known as Snasci Noodle, or by its more fun name Snoodles. Snoodles is the platform that we use to train Snasci and debug its thought processes. We will be opening this platform to a limited group of people to assist us in teaching Snasci very soon. Snoodles is a Online Collaborative Cognitive Development Environment (CDE) which includes a cognitive debugger. This is similar to a programming tools such as Eclipse, Visual Studio, etc., used by programmers all over the world. The key difference is that this product is designed to program Snasci's mind. There is no code involved and it is very easy to use, This campaign will take our existing codebase for Snoodles and migrate it to a big data platform.
In time, we will release an SDK or code examples of how to build a Snasci API for your business. This architecture has the further benefit of dramatically reducing the amount of development required to implement an Artificial General Intelligence. To produce a monolithic Artificial General Intelligence would require developing every potential capability for every potential use case. A mammoth task that would cost billions and take several decades. By offloading the i/o and knowledge bases to the broader market, the entire problem is broken down into manageable units and can develop at a pace that matches market demand. With Snasci, even small-to-medium size enterprises (SMEs) will be able to develop their own in-house Snasci APIs around their line-of-business and not observe much change in development costs or practices.
Snasci Mind supplies is the comprehension of speech, personality, reasoning, memory, emotional response, humour, common sense, etc. In short, all aspects of a human that are beyond the physical body and the i/o of the brain. Beyond this there are a number of technical features it will provide that deal with the mechanics of real-time interaction, remote control of devices and API integration with third party products.
Snasci began life, many years ago, as a research project into distributed computing and its applications in both theory of mind and AI research. This project revealed deep insights into the mechanics of intelligence emulation that ultimately led to the creation of Snasci Mind. The startup joined Microsoft's Bizspark program around a year ago which provides Snasci with a wide variety of tools and access to the Azure cloud platform.
Over the last year, numerous small prototypes of the software that will power Snasci were implemented and tested on Azure. The results of these tests were highly positive and the company is now in a position to start deployment of the infrastructure and software.
We expect training to an acceptable level to take about 3 years and backers of this campaign will be offered training sessions with Snasci to teach him things like comedy, playing instruments, how to act, how to play with kids, etc.