Exclusive Coinvision Interview with SingularityNET’s CEO Ben Goertzel
Coinvision sat down with Ben Goertzel, CEO and Chief Scientist of SingularityNET, a project leading the way in integrating the blockchain and artificial intelligence (AI) by creating a decentralized marketplace for knowledge sharing in AI research. After a USD$36 million initial coin offering that sold out in just over a minute, a lot of expectation surrounds one of the most forward looking projects in the blockchain sphere today.
Yesterday was a big day for SingularityNET as the team is in Toronto to announce the launch a new platform named DAIA, an Industry Alliance Leveraging Blockchain to Democratize AI. The project comes from a partnership between SingularityNET and AIDecentralized and will be working with over 100 projects in AI development.
This is what Ben had to stay about the state of affairs of SingularityNET right now.
Coinvision: What do you regard as the main transforming aspects the SingularityNET platform will achieve?
Ben: There are two main aspects to the platform. The first is the platform itself, which has the potential to open exchanges within the AI research community. The other main shift is the set of AI technologies that our team has been working on and that we will be deploying on the platform, in open source, for everyone to experiment with and use.
In a way, just like the Android system allows developers to build applications on top of it but at the same time rolls out its own applications like Google Maps and Gmail, we will aim to do the same within AI, which will also give extra motivation for people to start playing with the platform to begin with.
”In the medium term our goal is to offer something that is better than the cloud-based platforms that Amazon, Google and Microsoft are offering but that isn’t fundamentally owned by anyone”
Not just decentralized in terms of its physical infrastructure but decentralized in governance.
Once we manage to deploy the platform, the resources needed for AI research will dramatically reduce while the scope of research can be significantly increased compared to the situation now.
”Today, because AI research is centralized in a hand full of massive corporations and their shareholders, sectors of AI that can be very interesting for the wider audience remain underdeveloped as they are perceived to be less lucrative”
In this sense, there were two goals that were at the centre of this project. One is the potential for the acceleration of AI research, creating smarter and smarter machines moving towards general intelligence, and the other is to democratize AI, allowing for anyone with knowledge of AI technology to contribute for the development of these projects.
Coinvision: What sort of sectors are you referring to?
Ben: As an example, in 2014 I co-founded Icog Labs, which was the first AI research centre in Ethiopia. There are so many brilliant students coming out of universities with computer science degrees and engineering degrees that are not able to find jobs locally and depend on finding scholarships to go abroad. There is no Google there to hire these people.
So, we started deploying some of these technologies there. Using AI to analyse crop diseases with data gathered from local farms for instance, or helping to adapt software to some of the local languages that haven’t been addressed by mainstream companies. These issues are not the focus of major tech firms and therefore these avenues never get explored.
”So much money is spent on face recognition software for example, which has its main applications in government control and advertising, while so little goes to analyse plant diseases that could contribute to help communities be more sustainable”
Our ultimate goal with SingulartyNET is to create a marketplace where the scope of research can be widened to reach those avenues that the centralized companies are not interested in covering.
Coinvision: What is the main challenge you are facing in the platform’s development right now?
Ben: Today, one of the biggest challenges that we have is just how complex the usage of blockchain-based applications can be. We see that even people with AI backgrounds that love the concept of SingularityNET and are supporters of the project, are using Google’s Tensorflow to work rather than Singularity’s Alpha, because it is much slicker, faster and efficient at this point. Much like most people use Windows or OSX rather than Linux. While Linux allows greater functionality and even has less bugs than the other operating systems, not to mention that it is not owned by anyone, it is also much more complicated to use, and so, most people will prefer the easier options despite all their shortfalls.
So, it is a central point for us to be able to provide a platform that works well, flows, doesn’t crash and ultimately operates better than the centralized solutions. We are working on building a platform that is smarter and more capable than what already exists out there but that is always easy to use to be able to attract a wide community of users.
Coinvision: When can we see this platform deployed?
Ben: We have just released a second iteration of the Alpha version of the platform on Github, which is much more friendly for people to build upon. The Alpha version we released last year, while it worked and expressed clearly the concept of SingularityNET, had various limitations that we have addressed to an extent now in the new version, which I would call Alpha++, rather than a Beta version.
At this point, it doesn’t add a lot of new functionality, but it has a different architecture that reflects our careful thinking about how to make a system that can be gradually growable into the full-scale platform that we need. Early-adopter AI developers would be well-served by crafting their AI code to make it work with our new Alpha code; with some technical caveats, this would result in their code also working with our scalable platform when it’s ready.
Later in the year, while we don’t have a set date yet, we will release a scalable version of the platform that is directed both at the end users and developers and that we can classify as Beta.
BOX
Yesterday, SingularityNET launched DAIA, an alliance of companies and foundations and labs working at the intersection of AI and blockchain. It has been created with the aim of fostering cooperation on projects, standards, events and IP and media and other matters of common interest.
Ben had this to say about this brand new endeavour:
“To me, DAIA is more than just another industry organization. It’s a movement — a movement by a network of adventurous technologists and entrepreneurs around the globe, aimed at wresting control of the world’s AI and the data and compute power that feeds it from the handful of big tech companies and big governments that are currently threatening dominance.
DAIA is founded on the premise that making AI more decentralized, democratic and participatory is important for combating current issues such as wealth inequality and data privacy, and also for accelerating the development of AI technologies toward a broad scope of applications, not only those matching the business models of big tech and big government.”
Coinvision: What sort of SingularityNET-developed applications are you planning to deploy on your platform?
Ben: I have been working for a long time on an AI architecture called OpenCog, directed at achieving general intelligence. It is a probabilistic-based AI system, which is not the most popular field of research right now but has enormous potential applications.
”We are planning to use SingularityNET as a venue for making OpenCog into a scalable and distributed system which will help us make it smarter”
I believe this could become a really defining step for SingularityNET because it is based on a very different AI architecture than any of the projects the major corporations like google and Microsoft are working on. Most big companies are now focused on deep learning software using neural nets, which is very good for processing audio, video and images. We will also have them on Singularity as Sophia uses deep neural nets for vision processing and for head movement control. However, what we are working with right now is an AI architecture that is much more powerful in terms of abstract reasoning and learning and that has many more applications.
Now, all that we put on SingularityNET will be open source. So nothing will stop anyone or any company from just copying the code and throw money into it to create a really slick interface and market it, but somehow I think that these major corporations lack the flexibility and agility to really adjust gears and shift directions. After all, Microsoft could have copied Linux’s code decades ago and only now do you see some Microsoft servers running Linux. There is always a time lapse because their structures are too rigid.
Coinvision: One of the most interesting announcements of the last few months is that you will be deploying Sophia’s (the Robot) mind on the network. What exactly does that mean?
Ben: We will effectively upload Sophia’s mind into OpenCog on the SingularityNET platform. At this point, its system is a combination of deep neural nets and a sort of logic-based decision-making protocol. When we link her to the network, she will be able to connect to a number of other deep neural nets. In this way, Opencog will be used to control her dialogue, while using the deep neural nets to recognize emotions in someone’s face, for instance. It will allow for the coupling of additional, more complex systems to her AI.
As an example of continuous improvements we are making to the system, we have used motion captures to capture an actress’s neck and head movements, which I “deep neural net modelled” so Sophia can generate head movements from that data, which gives her a much more organic movements.
At the end, the system works through the combination of a set of deep neural nets that record emotions, movement, images etc, and that are linked to OpenCog, which works as a logic system that chooses what to say and how to react and move. Using SingularityNET to interact with these systems will be much more flexible than the technology that is available in the market today.
The Robot Operating System that we are using now has a centralized node, which works well if you try to create one Sophia, but is totally insufficient if you want to create 1000 Sophias.
”If we envision a future when anyone can have a Sophia on their phone or in any operating system, then we need a decentralized infrastructure to be able to make the system run”
Coinvision: You have mentioned before that the Ethereum’s scalability limitations were insufficient for SingularityNET’s needs in the long-term and that you were working on developing your own blockchain. What is the state of that development?
Ben: There isn’t a single blockchain out there right now that will be sufficient to serve SingularityNET’s needs in the future. There is a lot of hype around projects that claim great scalability capabilities, but the truth is that when we try those networks, none is capable of giving the structure that we need. So we have two choices: We either build a better blockchain or help someone else build a better blockchain, or we choose to design around it while waiting for the next blockchain to come. At this moment, all our blockchain development is on Ethereum. It is a robust system and is well understood. So, for now it works, even if we are doing a lot of things off-chain rather than deploying everything on the blockchain. As the technology develops we hope we will see a system emerge that is capable to really scale to the level we need.
Coinvision: Where do you see SingularityNET in the long-term?
Ben: When the first personal computers started to be made available and we started to be able to just type commands and get the machine to operate, there was no way we could see just how impacting this technology would become today. The same for when electricity first emerged. So, I expect the same thing will happen with AI. Right now we are looking at robotics, we are looking at biomedical AI or even at social media analytics. We are also looking specifically at agricultural issues. Being able to take a photo of a leaf and instantly know what kind of disease the plant suffers from or what treatment does it need. But that is just the start. The potential applications are endless and impossible to predict at this point.
And that’s it from our conversation with Ben Goertzel, CEO and Chief Scientist of SingularityNET. Stay tuned for our updates and join our Coinvision Premium community to be the first to get alerts about high ROI opportunities.
Disclaimer
We are stating our vision and opinions, we are not giving you financial advice and if you invest in this project it is your full responsibility. You are investing at your own risk. Always invest only what you can afford to lose and try to diversify your investments. Finally, do your own homework and learn about the project use case, roadmap and team.
