SingularityNET & Singularity Studio
Blitzscaling Toward the Singularity
In the brief period since our announcement of the Singularity Studio spinoff on the Web Summit Centre Stage on Wednesday, we have received a lot of excited feedback from individuals and organizations wanting to get involved in the Studio, and have also noted some confusion in parts of our community about aspects of the relationship between the existing SingularityNET Foundation entity and the planned new Singularity Studio entity¹.
So I have decided to write a fairly thorough post to address some of the issues that have come up in discussions with various community members since the announcement, and hopefully minimize any unnecessary uncertainties folks have about our plans going forward. (Of course there are many related uncertainties that can’t be removed by clear explanations too — we are doing something that is massively new in multiple ways!)
In Praise of Speed and Uncertainty
By way of prelude, I feel impelled to trot out a quote from Reid Hoffman’s recent book Blitzscaling:
Speed and uncertainty are the new stability.
The goofy but appealing coinage “Blitzscaling” refers to an organizational management strategy in which one places growth above nearly other objective. Uncertainty is embraced, a certain amount of mess and waste are tolerated, with the goal of growing larger and larger as quickly as possible. “Large” here means both large in terms of team size and large in terms of number of customers served.
Blitzscaling is how we got the tech titans now dominating the AI field: Google, Amazon, Facebook, Microsoft, Tencent, Alibaba, Baidu. It is also the only plausible way a tiny little upstart like SingularityNET is going to divert the focus of the planetary AI ecosystem away from these big tech companies toward a decentralized, participatory ecosystem.
“Speed and uncertainty are the new stability” doesn’t mean one makes huge strategic decisions thoughtlessly and heedlessly. But it means one needs to evaluate the situation incisively and quickly, and then clearheadedly and rationally make the best choice one can, and act on it decisively in spite of the large amount of remaining uncertainty. This does not provide any sort of guarantee of success. But in the modern business, social and economic climate, it provides the highest odds of massive success.
The relation between this philosophy and the impending Technological Singularity is obvious. Exponential change in technology and society, which is what has driven Kurzweil and others to predict a coming Singularity, is also what drives the need for speed to leverage opportunities before they transform into something else, and the need to deal with uncertainty because the scene is changing so fast that a large percentage of the patterns one has recognized are no longer going to apply next year.
This is the spirit in which I and the rest of the SingularityNET executive team decided to launch the for-profit spinoff Singularity Studio.
Singularity Studio: Accelerating Utilization of the SingularityNET
To quote Reid Hoffman’s book once again (partly because it’s one of the two books I brought on the flight where I’m writing this post, and I can’t find anything relevant in the other one, “How the Spiritual World Projects into Physical Existence” by Rudolf Steiner!):
I believe there is a Maslovian hierarchy of fires that applies to most rapidly-growing startups, where the top of the list is the most important fire to fight first:
Distribution
Product
Revenue Model
Operations
Competition
What’s next?
What does “distribution”, Hoffman’s top priority, mean in a SingularityNET context?
SingularityNET is a platform business, which to succeed needs to create flourishing, rapidly growing networks of both suppliers (AI developers) and users (software developers or customizers needing to plug AI into their systems). The more suppliers there are, the more appealing the platform will be to users; and vice versa. To get a business like this off the ground, one needs to aggressively seed growth on both the supply and utilization sides.
What “distribution” means in a SingularityNET context is: getting AI developers and AI-using software developers and customizers on the platform.
To seed the supply side, we are doing two main things. First, we are building powerful AI tools on our own, including some like the OpenCog system that we believe to have unusual potential to yield advanced intelligence. Second, we are launching a series of SingularityNET AI developer workshops, beginning with one at the DevOps conference in Helsinki on Dec. 14. We will also release online video workshops in early 2019, recorded in our main office in East Tsim Sha Tsui in Hong Kong.
The core goal here remains, as articulated in the original SingularityNET white-paper, to create a network of AIs that share data and outsource work to each other, and a network of users and providers that keeps this network of AIs fed with new data and algorithms and with resources to maintain and expand their operations. We see this as the best way to fulfill the organization’s root mission to catalyze the emergence of beneficial general intelligence.
To achieve this core goal with maximal probability we need to aggressively seed both the supply side and the demand side, and that is where Singularity Studio comes in. As soon as we launch the beta version of the SingularityNET platform in Feburary 2019, we will begin relentlessly inviting and encouraging software developers and companies in need of AI to use our platform and the AI tools on it. However, it’s hard to predict the speed with which this effort will succeed. In the early stages, our platform will have both advantages and disadvantages relative to alternatives such as AI platforms offered by big tech companies.
The beta SingularityNET platform will provide some unique AI algorithms and approaches (created by the SingularityNET AI team based on their decades of prior AI experience), and the privacy and control benefits that come along with its decentralized infrastructure. But, to be frank, at time of our beta launch in February, we will not be able to match big tech companies’ tools in terms of slickness of user interface, or variety of data importers for handling different types of user data (to name just a couple examples).
Singularity Studio has potential to drastically accelerate the utilization of the platform after the beta launch. The basic idea is very simple…
Large corporations from around the globe frequently approach the SingularityNET AI team asking for help incorporating AI into their systems and processes. This is because every company now knows they need AI, but most companies — even large and wealthy ones — lack the AI expertise to know exactly how to apply AI in their contexts.
Now, SingularityNET is not in the business of AI consulting for large corporations, per se. But working with companies to figure out how to fit usage of the SingularityNET platform into their systems and processes … and then helping them with the technical work needed to connect their own internal software and data with AI on the SingularityNET platform — that’s a more appealing line of business to be in.
But it’s still a pretty different business from the core business of SingularityNET Foundation, which is to create an awesome SingularityNET platform and some brilliant core AI tools to run on it.
Presto, Singularity Studio.
The Singularity Studio Product Line (Initial Vision)
As often happens, once one gets into the details, things get a little more complicated. Helping corporations to integrate their systems and processes with SingularityNET, one by one, is a feasible thing to do, but not very efficient in that it would involve a lot of repetitive work. There are some aspects of integrating SingularityNET into corporations’ systems and processes that are going to be common to a majority of corporations’ situations. There are other aspects that are going to be common to a majority of corporations in a specific vertical market — say, pharmaceutical companies, or investment banks, or car companies.
To increase efficiency and decrease repetitive work, then, it makes sense for Singularity Studio to create a core product, with robust software tools for connecting SingularityNET AI into corporate information systems. It also makes sense for SingularityNET to create vertical-market-specific plugins for this core product.
One sort of challenge faced in this context is the desire of many companies to run their AI in a “hybrid cloud” scenario. This means they will be happy to outsource some of their AI processing to SingularityNET services running anywhere that’s sufficiently secure and reliable — but will also want some of their AI processing to run on server farms that they host internally. Enabling this will require significant infrastructure work on the SingularityNET side.
The SingularityNET beta platform will depend on certain features of big-tech cloud providers such as AWS, in order to achieve efficient functionality. But to run a SingularityNET subnetwork on a corporation’s on-premise server farm, this dependency can’t be there. Removing this dependence is important for SingularityNET to realize its long-term ambition of decentralized AI, but the need becomes more acute and urgent in a Singularity Studio context.
The same infrastructure tools that will allow SingularityNET to be run in a hybrid-cloud configuration for large corporations, will also allow anyone who has a server farm to run SingularityNET there. This leads to the realization that Singularity Studio should offer an IaaS, Infrastructure-as-a-Service product.
So we arrive at the Singularity Studio product line in its present, preliminary conception:
- Singularity Studio (the core enterprise product)
- Singularity Biomedical Studio
- Singularity Financial Studio
- Singularity IoT Studio
- … and more…
- Singularity Infrastructure
The Studio products will rely, for their AI processing, on the SingularityNET platform and AI services running therein. They will supply additional aspects such as user interfaces, specialized data-type handlers, code for hooking into corporate IT systems, machine learning models trained on data relevant to specific vertical markets, logical knowledge bases curated for probabilistic reasoning relevant to specific vertical markets, etc.
When we first founded SingularityNET, it wasn’t so clear to us whether we would need to supply these additional aspects, or whether others in the tech ecosystem would rapidly arise to supply them. At this point the landscape is a bit clearer to us, and it’s plain that the most rapid course to massively scaling SingularityNET is to supply these things ourselves.
Building A New Team…
SingularityNET Foundation, as a team and workplace, is something of a globally distributed, decentralized network of freewheeling tech geeks, hackers and mad scientists. We do have a reasonable element of corporate organization — we maintain careful accounting records, pay our bills on time, attend to legal compliance, pursue quarterly OKRs across different divisions of the company, and so forth. But still, we are frankly a quite different sort of organization than an enterprise products and services company.
On a personal level, I think I’m a pretty good leader for SingularityNET Foundation, but would be much worse at the job of leading something like Singularity Studio — while I realize that Singularity Studio’s mission is critically important, serving enterprise customers maximally well is simply not what makes my heart beat most enthusiastically when I wake up each morning. Direct pursuit of the vision of decentralized, beneficial AGI does that for me.
It’s true that Cassio (SingularityNET Chief AI Officer) and I earned our livings for the 16 years prior to starting SingularityNET by leading enterprise AI consulting and product companies together. This is why we understand so well what Singularity Studio needs to do and how it needs to do it. But it is also why we understand that Singularity Studio needs to be a different organization, with a different sort of team, than SingularityNET Foundation.
The need to build a Singularity Studio product suite rather than just doing customization, and the need to build a different sort of team for Singularity Studio, are the central reasons we have decided to seek additional funding to get Singularity Studio off the ground.
Our Duty as Executives of SingularityNET
I hope the above lengthy comments have made clear that, while at first the new child entity Singularity Studio could seem a digression from the core mission of SingularityNET, it’s actually an attempt to adaptively and dynamically do what is best for the achievement of this mission.
I, and the other executives of the Foundation, have a duty to maximize the impact of the project in its totality and to explore new avenues that could provide significant value to the project’s long-term enduring success. This is the spirit in which the Singularity Studio is being ventured.
Some Comments on Singularity Studio Structure As Currently Envisioned
While the business and technical purpose of Singularity Studio is very clear (and I’ve outlined it above), the formal and legal structures are not yet finalized. We are still doing research about various related aspects; and also, before finalizing any formalities regarding Singularity Studio, we will seek explicit input from the AGI token-holding community, as mandated by Foundation governance.
In this connection I need to first comment on a related matter which has caused some unfortunate confusion in the last couple of days. Due to a misunderstanding a document provided publicly by a partner included some information about potential structuring and fundraising decisions for Singularity Studio. This was the fault of the SingularityNET side only and we take responsibility for that. I need to clarify that the contents of that document were in no way a definitive, approved, finally decided plan. Rather, they were real hypothesis that we were (and are) exploring in conjunction with our advisors.
While these are matters we are still thinking through, in the spirit of transparency and communication, I will now share some of our current thoughts regarding how Singularity Studio might be sensibly structured.
Firstly: Our current hypothesis is that the SingularityNET Foundation will have a significant percentage of the equity of Singularity Studio; and as such, Singularity Studio will be a child entity of the foundation.
Equity of Singularity Studio may also be shared with interested parties, such as accredited institutional investors like Venture Capitalists or Institutional Funds.
We have also been considering the possibility of doing a Security Token Offering for Singularity Studio, though this is not finally decided and is pending (among other things) further discussion with the AGI token holding community.
According to our preliminary discussions around an STO, what we might implement is in effect a tokenization of the equity of Singularity Studio — rendering it effectively an Equity Token Offering.
In this hypothesis we have been considering, the STO would be limited only to Venture Capitalists and Accredited Professional Investors. Liquidity restrictions would apply, given that planned resales of the equity tokens issued would be subjected to restrictions and lock-ups and would not be buyable now or in future on the open market.
To demonstrate how restricted this sale is, consider that only 3% of US residents are legally eligible to invest in start-ups of this nature.
It should be clear from this that, if we do create Security Tokens corresponding to Singularity Studio, these STs will have a very, very different nature and status than the AGI utility token. They are not comparable constructs at all.
Security Tokens have some advantages over conventional equity shares, many of which are obvious consequences of their (restricted, but still non-trivial) liquidity. In a security-tokenized company, as opposed to an ordinary private company in which shares have been sold to investors, early investors have a more straightforward way to cash out their shares when/if they want to.
For Singularity Studio, one advantage of a STO as opposed to a standard sale of equity shares, would be that in the STO route there would not be the same pressure from early investors to seek exit via acquisition. The vast majority of AI startups end up being acquired by big tech companies, and this outcome is often very actively sought by the VCs who invested in these AI startups, because it provides the VCs with the fastest and most reliable route to a highly profitable exit.
However, an exit via acquisition is not what I and the other SingularityNET Foundation executives currently have in mind for Singularity Studio. An appealing aspect of the STO route is that it would provide VCs and other early investors in Singularity Studio with a more straightforward route to exit some or all of their investment in a way that leaves Singularity Studio intact. (Of course, a traditional IPO also has many of the advantages of an STO. But IPOs have different costs and complexities, and whether Singularity Studio might eventually do an IPO is not a matter we need to consider in depth at this stage.)
On the other hand, the infrastructure to support STOs is currently at an early stage, making an STO a significantly more complicated undertaking than more traditional ways of selling equity in a private corporation. We are in the process of investigating various aspects, with an aim of validating or refuting the hypothesis that an STO is the best path forward for Singularity Studio.
According to our current hypothesis, Singularity Foundation will own a significant equity share of Singularity Studio, and in relation to this, a share of the profits made by Singularity Studio will go to the Foundation. In addition, the Foundation will possibly license the use of their current and future IPs to Singularity Studio, in exchange for which Singularity Studio will pay yearly FMV fee on these, providing the Foundation with greater sustainability.
Current AGI Tokens, Supply & Transferability
I hope it is very clear from the above that, in any scenario where we do create Singularity Studio security tokens, current AGI Tokens will not be replaced by Singularity Studio security tokens. It must be emphasized that the two tokens are markedly distinct!
The notion of a “token” is one of a symbol for something else; and in this case, the things the two tokens symbolize are very different. One represents a certain amount of AI services on the SingularityNET network, the other represents a certain percentage ownership of a company selling enterprise products and services that leverage the SingularityNET network. Obviously there is a relationship there, but also a major difference.
The current AGI utility token is open to anyone and will always be so; while, if we create a Singularity Studio security token, it will be restricted to transfers between Venture Capitalists and Professional Investors.
While the current AGI token is related to the Foundation (The Parent), the Singularity studio equity token is related to Singularity studio (the Child Company).
Our belief is that, assuming Singularity Studio is successfully capitalized and launched and becomes a flourishing business, this will be a very good thing for the SingularityNET Foundation, the AGI token ecosystem and the SingularityNET nework-of-humans-and-software as a whole. We see tremendous potential for positive feedbacks here.
On a formal level, there is a limited analogy with dual class stocks. A dual class stock is a financial instrument already in place in publicly listed companies; and there are well-researched studies documenting no detrimental relationships when analyzing the behavior of how one class of shares impacts the other or vice versa.
However, in the SingularityNET situation we have some interesting and specific positive-feedback relationships that don’t exist in a typical dual class stock situation. The more products and services Singularity Studio sells, the more AGI tokens will be used, and the greater the value there will be in the SingularityNET network. The more value there is in the SingularityNET network, the better the Singularity Studio products will be, and the more they will sell. Etc.
At this early stage we cannot make any reliable specific forecasts, but just as a form of evocative story-telling in order to let you better understand the link between the two entities, let’s suppose in some future fortunate year, Singularity Studio were to sell $500M of AI software products and services. Suppose that, say, $100M out of this was used to purchase AI services on the SingularityNET platform. Then the platform would have an economic flow-through of $100M just from this one source. If another $150M out of this $500M were taken by Singularity Studio as profit, then the SingularityNET Foundation would also receive a percentage of this profit. Utilizing its percentage of Studio profit, and by selling some of its own Foundation Reserve AGI tokens each year, the Foundation could then fund ongoing work on platform and AI development — which would work toward the goal of beneficial AGI and also provide additional intelligence and economic value to Singularity Studio’s products. The additional work done by SingularityNET staff using the funds derived from Singularity Studio products, would thus end up improving the Studio’s offerings, increasing Studio revenue — so maybe the following year it would be $600M rather than $500M. Etc.
Going back to Reid Hoffman’s quotes from the beginning of this long post — this is a highly plausible way to blitzscale SingularityNET. It is far from certain, but there is no certainty in any tech startup, let alone one spanning such dynamic spaces as AI and blockchain.
On the importance of our AGI Token Holders
There is a lot of complexity involved here, and brevity is not my strong suit; I am thankful to anyone who has read this far!
For those who are reading this and who hold AGI tokens, I hope that through all this explanation, it has been clear that the Foundation’s executives have not forgotten the voice of our current AGI holders who have seen us through some challenging times. SingularityNET aims to disrupt an industry whose players are among the biggest companies in terms of capitalization worldwide; and to do so it is necessary to think out of the box and try to compete on a level playing field for both talent, technology and customers. Speed and uncertainty are the new stability!
Concluding Caveats
As Kurt Vonnegut liked to say … And So It Goes …
The above comments are considerations that I and the other SingularityNET executives and our advisors are thinking about. As in any start-up, strategic pivots and details can be modified. All executive decisions in this domain are subject to the approval of the existing governance structure, which includes the voice of the AGI holding community.
Further information will be shared with the community when finalized. Our official channels will be used to share detailed information with the relevant community members.
While the Singularity Studio is an exciting new development, our operating priorities remain heavily focused on the development of our product — the launch of the beta platform in February, the education of our community and the enablement of our customers. Because of these priorities, and the complexity of some of the matters under consideration, clarification of all the details raised here may not be immediate.
Thank you
Ben Goertzel
CEO, SingularityNET
¹ Certain information set forth in this presentation contains “forward-looking information”, including “future oriented financial information” and “financial outlook”, under applicable securities laws (collectively referred to herein as forward-looking statements). Except for statements of historical fact, information contained herein constitutes forward-looking statements and includes, but is not limited to, the (i) projected financial performance of the Company; (ii) completion of, and the use of proceeds from, the sale of the shares being offered hereunder; (iii) the expected development of the Company’s business, projects and joint ventures; (iv) execution of the Company’s vision and growth strategy, including with respect to future M&A activity and global growth; (v) sources and availability of third-party financing for the Company’s projects; (vi) completion of the Company’s projects that are currently underway, in development or otherwise under consideration; (vi) renewal of the Company’s current customer, supplier and other material agreements; and (vii) future liquidity, working capital, and capital requirements. Forward-looking statements are provided to allow the opportunity to understand management’s beliefs and opinions in respect of the future.
These statements are not guarantees of future performance and undue reliance should not be placed on them. Such forward-looking statements necessarily involve known and unknown risks and uncertainties, which may cause actual performance and financial results in future periods to differ materially from any projections of future performance or result expressed or implied by such forward-looking statements.
Although forward-looking statements contained here are based upon what management of the Company believes are reasonable assumptions, there can be no assurance that forward-looking statements will prove to be accurate, as actual results and future events could differ materially from those anticipated in such statements. The Company undertakes no obligation to update forward-looking statements if circumstances or management’s estimates or opinions should change except as required by applicable securities laws. The reader is cautioned not to place undue reliance on forward-looking statements.