An Interview with SingularityNet, a next gen-AI project that aims to put artificial general intelligence on the blockchain.

In 2021, perhaps the two hottest areas of technology undergoing rapid development and hype are AI and Blockchain technologies. SingularityNet is a project that aims to democratize the advent of general artificial intelligence and make it inclusive to benefit the whole world rather than just a proprietary technology aiming to benefit a particular tech company. These areas of innovation are moving very quickly and SingularityNet is making very quick progress in developing AI on the blockchain. The team was kind enough to answer some questions for PTN explaining the working going into this project:

Can you just briefly explain in a way that non blockchain experts who are interested in the future of tech can understand the value proposition of SingularityNet?

So to understand why SingularityNET is so important, one should first have an understanding of the current AI development Landscape.

I think by now most people have at least some concept of how AI can help us in the future, along with some sense of the danger of its usage.

In truth, AI is already all around us and we use it every day without thinking, mostly from our mobile phones assistants and web searches most of us do many times daily. Though AI can also be found in many places people are not quite so aware of, such as in your fridge monitoring the temperature, or your car actively managing the fuel mixture and many many many other places you might not expect.

AI development is it is very costly and there are far too few AI developers in the world for everyday businesses to develop their own AI applications. This isn’t a problem for the tech oligopoly and large developed nations who have almost limitless access to funds and can attract AI developers with large salaries. Big tech companies such as Google, Apple, Amazon and Facebook and nation-states in effect have control over the development of and distribution of AI at present.

Ultimately this means that the kinds of AI we are seeing being developed are either specifically commercially led, informatics led or offer some other developmental benefit that furthers a competitive advantage over their peers.
Not only is this bad thing from a perspective of squashing development of AI toward the betterment of mankind, but it also is dangerous when one considers the ultimate goal of AI development is to achieve AGI or Artificial General Intelligence.

Artificial General Intelligence or AGI is a point where AI has the ability to approach a task it was not specifically programmed for and work out for its self a way to solve the problem. Currently, all the AI that exists today is considered Narrow AI, that is to say, it can only do a task it was specifically designed for.  

Why is it dangerous for AGI to be formed by large centralized bodies? Simply put any AGI that ultimately at its core had business ethic or heavens forbid defense applications, human values will take a back step in Favour of the pursuit of profit or power for the centralized body it originated from.

This concept is where most peoples fears are born and are precisely why SingularityNET and other Decentralized organizations are vital for humanities future.

SingularityNET offers the ability for any developer anywhere in the world to register their AI algorithm onto the SingularityNET platform wrapping it with our API and creating an AI agent that allows anyone else globally to access it. What’s more, is this AI Agent can be interconnected with any other agents on our platform. This means complex applications can be built by chaining AI together. An example of this is connecting a Speach to text algorithm with a translation algorithm and then back to a text to speech algorithm and being able to offer live translation services. Application developers who previously had neither the funds nor the knowhow to develop their own AI can access these AI agents on our platform and incorporate them into their applications.

Eventually, the goal is to allow these AI agents to self organize based upon natural language requests. Currently, the only way to interconnect AI agents is to either use some middleware or to manually connect them. The vision is for anyone to be able to make a request of SingularityNET such as ‘I’d like to translate this video from French to English’ and the agents would self assemble and offer the requested service.

So what is the value proposition of SingularityNET? It offers open access to powerful AI services globally, breaking down the barriers of development and freeing the development from the shackles of centralized bodies.

Tell me a bit about the team and what made you decide to start the project.

SingularityNET was conceived in July 2017 when three friends and colleges, Ben Goertzel, David Hanson and Simone Giacomelli had a discussion centred around the concepts of AI, the future and how to get from here to the Singularity.

The technological singularity is in essence a time in the future where technological advance has gotten so fast that the implications are impossible to fathom. Technological advance has remained on an exponential scale consistently throughout history, survive all major obstacles humanity has faced. Dr I.J. Good first coined the term Intelligence explosion in 1965, this was later re-imagined as the technological singularity by the likes of  Vernor Vinge, Ray Kurzweil and Ben Goertzel.

From this discussion, an idea formed about how machines might one day gain the power of consciousness and sentients. Ben being an avid reader of Marvin Minsky suggested that as described in Marvin’s book ‘A society of minds’ that we could achieve machine consciousness by creating a network of AI and the basis of SingularityNET was born. Over the course of the next few months, a small team grew around the founders as more AI scientists, blockchain specialists, engineers, developers and tech marketeers worked together to produce the SingularityNET whitepaper. This paper formed the basis of our crowdfunding campaign and ultimately our token generation event where we minted the token AGI, which happened in December 2017.

What is your vision for the next 5 and ten years of the project?

We’re aiming to occupy a significant percentage of the AI market space by 2025, with trillions of API calls going through our platform and with self-assembling AI services that meeting the demands placed upon them. Importantly we hope that the decentralized nature of our platform will lead to great discoveries and further enhance humanities global wellbeing. Ultimately we are pushing toward creating beneficial AGI and a positive technological singularity.

You recently held a vote to move into phase 2 can you explain how this will transform the project?

During Phase One, solid progress was made on the platform and AI development, but limited advances were realized in the growth of user traction and network utilization. SingularityNET is designed to exploit network effects in a massive, multi-dimensional way and these effects only kick in once there is a critical level of network utilization by a wide variety of users. Phase Two addresses this issue by providing a massive economic boost to community use of AI, introducing spin-offs seeded in Phase One to the newly formed SL2 platform, and funding new projects from the community.

In Phase Two, the AGI token is expected to come into more intensive practical usage as the network’s central payment mechanism. Customers and AI agents will use it to compensate AI agents for services. From a cognitive science perspective, one can think of the population of AI agents as an emergent mind, and the AGI token is the mechanism it uses for ‘assignment of credit’, a key aspect of any experiential learning system. The economics of the AGI utility token are a way for human and artificial network participants to support operations, and profit from network participation, while networking intelligence self-improves and self-regulates.

Can you tell us about important partnerships you have in the work with robot makers?

Our relationship with Hanson Robotics is relatively well known, as is, of course, the world’s most famous expresionate robot Sophia, what is not so well known is that David Hanson is, in fact, a co-founder of SingularityNET.  In recent times we’ve undertaken a partnership with a Hanson robotics Spin-off project called awakening health, to mass-produce a robotic nursing assistant called Grace based upon Sophia, who can help across multiple areas of health care such as eldercare as an example. Grace will interact with the SingularityNET platform and so each and every grace robot out there will generate calls to the platform. We’re really excited to see this project taking shape, we need AI to gain a sense of human values if we are to reach for a beneficial future.

How do you plan on getting to mass adoption?

The Phase Two plan has two key aspects for getting massive traction: 1) Working closely with, and in many cases, incubating teams building Layer 2 networks on top of SNet, 2) Creating Infrastructure-as-a-Service tools which will then make SNet a viable option for enterprises, and then reaching out to large enterprises (via Singularity Studio but also more broadly) to leverage SNet for their secure/decentralized AI needs
Once SL2 networks and some large enterprises are there, then there will be an ecosystem appealing to indie AI developers.

Ah yes and not to forget, 3) the DEEP Funding pool which will incentivize indie AI devs to put stuff on the platform w/ direct AGI token compensation

But 3 on its own won’t kindle the massive traction we need .. you can’t just buy traction … I think DEEP Funding will be very impactful but in conjunction with 1 and 2

Why did u want to move from Ethereum to Cardano and how is this going to affect investors and token holder?

We are not moving away from Ethereum, ETH has and continues to offer great functionality for those wishing to create smart contract based platforms. However, ETH currently is a victim of its own success in so much as it has become slow and expensive to use given the volume of transactions and type of demands placed on the blockchain.
Our plan has always to be multichain or cross-chain capable, specific decentralised distributed networks have always suffered from the blockchain trilemma.

1. Decentralized: creating a blockchain system that does not rely on a central point of control
2. Scalable: the ability for a blockchain system to handle an increasingly growing amount of transactions
3. Secure: the ability of the blockchain system to operate as expected, defend itself from attacks, bugs, and other unforeseen issues.

Why is this a trilemma? Because you cannot increase the scalability of the network without sacrificing either decentralization or security to some extent. A kind of blockchain equivalent to the Heisenberg uncertainty principle.
What’s more, is, individual blockchains are self-contained ecosystems that are isolated in terms of interaction with other blockchains, although solutions exist to interconnect them, doing so creates a point of centralizations.
So the solution appears to be rather than interconnecting blockchains, decentralized projects should be blockchain agnostic and allow multichain support.

Our Partnership with Cardano circumnavigates many of the current problems that exist on the Ethereum blockchain such as transaction fees and speed of processing, but further, than that Cardano uses some functional programming language called Plutus which can be used within the SingularityNET framework and assist with the interconnection of AI agents together, a process we’ve coined as AI-DSL.

Do AGI token holders need to do anything as a result of the change?

No, before the end of the 2nd quarter 2021 we expect to begin the phase 2 token release schedule, we’ll also release a way for token holders to interchange their current Ethereum based tokens with AGI on ADA tokens (the name of the new token is yet to be decided). Token holders will be able to swap their current tokens for the new tokens at a ratio of 1:1 and of course, both tokens will be functional to use on our platform.