Sophia & Albert, the humanoid robots, & SingularityNet, the benevolent AI marketplace

My motto is

“Decentralization is becoming an attractor that will transform everything.

Fintech will be no exception.”

Fintech will never be the same after this cultural shift.

Sophia (Twitter@RealSophiaRobot), the humanoid robot by Hanson Robotics and OpenCog, shares some of her reflections on this topic in today’s post. And Albert, her older brother, talks about the killer combo, Blockchain & AI.

Spring AI is here after so many years of winter. This past week all the established media have been covering AI topics, ranging from the acknowledgment of dumb AI being our current reality, to the social implications of narrow AI etc. (see references below — Forbes, Bloomberg, the Verge, Techcrunch etc.). Luckily, there has been no Jamie Dimon in the AI space trashing narrow AI and being “Long” Artificial General Intelligence.

All major pain point areas in finance will be transformed through AI. The process has started and naturally it is happening in the normal scattered fashion that innovation happens. Starting with use cases in fraud and risk management in consumer banking; compliance, governance, and stress testing at all levels; middle and back office operations; all the way to customer-facing services (new products and services, new revenue streams, new customer engagements).

Now that blockchain technology has initiated a truly genuine cultural shift towards Decentralization, the time is right to work on a large-scale vision that combines AI and Blockchain.

The SingularityNet vision

The Big Hairy Audacious Goal (BHAG) of SingularityNet’s vision is:

  • To allow for the transition from Narrow AI to Artificial General Intelligence
  • To offer access to smart and broad AI services
  • To create an ecosystem with incentives for AI in a benevolent manner
  • To create an open-source protocol for AI agents to outsource work to each other.

We live a world right now that Big Data is being processed as we speak mainly using all the data available to the tech giants. Google and Facebook already use AI algorithms to process real-time billions of data points. In financial services alone, quantitative trading with AI algorithms are run by more than 1,000 hedge funds and HFTs. Insurance companies use machine learning already to make operational decisions. AI and robotics is used in the health sector. However, all current implementations remain in silos and dominated by a few large corporations.

SingularityNet wants to unlock a key missing element which is communication, cooperation, interoperability between existing, diverse AIs. This is so essential to making the transition from Narrow AI to Artificial General Intelligence.

SingularityNet will be the platform for sharing datasets (e.g. conversations from existing siloed chatbots, data from a smart city ecosystem, etc). More importantly, SingularityNet will be the decentralized platform of AI agents coordinating the AI apps and datasets on the platform. This will allow the AI agents at the protocol layer to train and learn.

So, first SingularityNet wants to become a global decentralized cross-discipline AI library. Then, to become a global marketplace for AIs, much like an App store to buy and sell services, operated by the AI agents. Then, to enable the creation of Artificial General Intelligence as new AIs are created from the coordination of existing datasets and AI agents. The magic will be from these cross-discipline synergies that will allow for the creation of new tech stacks from the learnings of datasets and AI agents that were previously siloed. Think of a marketplace in which AI agents can outsource work to each other. SingularityNet will enable Artificial General Intelligence by structuring such a platform that allows AI agents to be each other’s customers. Subcontracting of AIs amongst themselves will be possible in an intelligent, complex way.

The BHAG of SingularityNet is described in detail in their white paper. It entails a protocol layer for the decentralized coordination. Then a Token layer for the transactions on the network via the SIN token. The API layer on top for the interoperability to function and the market layer for the marketplace to function.

SingularityNet is not going to be assembled from scratch. It will be seeded with AI nodes taken from existing projects like Google’s Tensorflow, the Japan Whole Brain Architecture, etc.

One the core seeding organizations is the non-profit Open Cog which has been created by the SingularityNet founders ten years ago and is currently being used by companies in diverse sector applications, such as Cisco, Huawei and Hanson Robotics.

Listen to Ben Goertzel, co-founder of SingularityNet and Albert Einstein from Hanson Robotics as they chat on stage after Ben’s presentation at the Moontec17 in Tallinn Estonia last week.

Most of you know Sophia, the humanoid robot, simply because she is younger and able to travel the world more than her older brother Albert. Unfortunately, Sophia was fully booked while I was in Tallinn last week for Moontec17. We did meet “offline” before she was being prepared for her meetings in Estonia.


The SingularityNet team promised to convey my questions to Sophia when back in HongKong and send me her reflections recorded. Enjoy Sophia and Ben.

The SingularityNet token sale was scheduled to go public last Friday the 8th. Due to popular demand and in order to make sure that all is properly and securely handled, it has been moved towards mid-December. Details on the website.

Disclosure: Efi Pylarinou is a Fintech thought-leader and advisor, affiliated with the SingularityNet group.

AI media references:

 Get fresh daily insights from an amazing team of Fintech thought leaders around the world. Ride the Fintech wave by reading us daily in your email.

One comment

  1. I really like your writing style, good info, thanks for posting. It is really great to read and especially about my favorite topic robotics. Thaks for this post.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.