TechRank as a Contribution Counter

Next: Keystones, silicon individualism, types of diversities (alpha, beta and gamma)

TL;DR Technology is composite in nature. PageRank algorithm can be modified to show how valuable different technology pieces are for society.

In the previous post series we presented decentralised organisation types and concluded that as crowdsourced models, they will put customer needs at the center. Things like planned obsolesce of today are on the way out.

And we claimed that the best option from the customer centered viewpoint is to release results for open use as open designs/open source software.

Open designs ultimately reduce the need for money as most everything will be available for free. Later we see this is not quite, but the general trend is clear.

If money loses much of its power to represent value, something is still needed to understand the impact different contributions make. But what?

Concept

When money no longer is a good way of measuring contributions to give fair rewards, some new method needs to come up. TechRank is one.

Things and services build upon subcomponents and technologies. Products are composites of other products and technologies.

Take for example the internal combustion engine. It is made from a wide variety of components such as cylinder heads, pistons, crankshaft, bearings, valves, engine control unit (ECU) controlling air-fuel mixture and ignition timing, bag of sensors and so on.

ECU itself consist of a microprocessor and flash or EPROM memory with the control program, all placed on a printed circuit board. Printed circuit boards in turn are made with a bunch of production technologies.  If you look inside the microprocessor, there are things like transistors, resistors, diodes that originally were separate components until we learned how to integrate all of them into a single silicon wafer. Silicon wafer manufacturing is own industry heavily relying on an industry producing pure silicon.

The combustion engine itself is a building block for new solutions such as powering vehicles, turning water pumps or generators for electricity. Generators are used in power plants that are part of the national grid.

Same is true of software products. A modern web application can use a bewildering number of different libraries – several databases tailored for slightly different data and usages, different microservices implemented possible with different languages, load balancers, API gateways, client libraries for user interface, continuous integration tools etc. All of these themselves are products made of other libraries and technologies. All of these run often on a cloud environment – itself complex combinations of technologies running finally on top of different hardware combinations such as racks containing power feed, cooling, servers, switches, routers etc.

Technologies create a complex web of interconnections that build on top of other complex webs of technologies.

These interconnections can be thought of as linking in the same fashion as web pages link to each other.

To understand the value of a single piece in this mix of interconnections, one can loan a concept from web search engines when they decide what web page is most relevant. A use of a technology inside a product is like a link from a web page to another.

By calculating all occurrences of a technology and summing their weights, one can calculate the ‘weight’ of a that technology. This is an iterative process as in the beginning all links have equal weight and then you calculate the weight of each node based on how much it is linked to (used) and this continues until the numbers stabilise.

This way one can see how important each technology is – in other words how fundamental it is. And this is a discrete number that can be used in a number of ways. It tells how valuable different open-source designs or software modules are and this can be used to allocate compensations, when a sensible mechanism to compensate is figured out. It also can be used to prioritise what technologies to focus on in hack in labs (= concept for continuous testing own vulnerability presented here) as these are the most widely used and hence it they are compromised, likely to have biggest impact.

Intuitively you can think this type of analysis calculating roughly how likely you are to use a certain technology either directly or as embedded – i.e., how much time or how many interactions you need to do with the world before you encounter it.

Weighted TechRank

How often something is used also has an impact on its value. Let’s say that any product design is available for self-manufacturing or use. These designs could be anything – medicine, smart contracts, poems, books, applications, machine learning algorithms for language understanding, cancer detection from medical images or for products such as home appliances, power stations, trains, furniture, clothes etc. Weighted TechRank is based on idea of counting every time a someone uses a design - every time it brings utility, convenience or joy. Then the weight for all technologies embedded by it is increased by 1.

This isn’t quite right as some products are used daily and hence often used. Some seldom used ones can be life critical, but not all.

The weighted system can also be gamed in a number of ways. Digital content can be artificially consumed. One could pump-up usage numbers for product designs by downloading same design multiple times (mitigation: measure from production, not download). Another way is to make a minor addition and claim that I have “contributed” to this design (mitigation: original designers accept change requests).

But even with these shortcomings, the value of technologies can be calculated by modelling their interconnections, apply the old Page Rank like algorithms used by web search companies and modify it to calculate technology value. You can use TechRank or its weighted cousin - TechRankExtended or just T-Rex between friends. The difference is that the weighted value is more volatile whereas non-weighted changes fairly slowly, particularly with physical products.

TechRank also puts all contributions to the same level. This has not been the case traditionally. Many new cool inventions have their foundation in university research that in some cases started decades earlier until they mature to a level where commercial products can be made. TechRank exposes those links and all contributions are counted equally. As a counter argument one might argue that university researcher do not have similar job security risks as startup employees and they gain grades and reputation in addition, so some balancing might be needed, but all this is highly subjective.

It’s likely that in future much of designs will be computer produced through generative design. In that case attribution could be done through a system like curation markets where the people finding usable things from a sea of generated options are rewarded rather than the generation itself. Or the contribution is split between the finder, machine learning model creator and the person who contributed their computing resources.

Generative design may also fundamentally change how products are designed as computers do not think in the same way as humans. They might redesign everything from first principles where people like to assemble and use already proven solutions/technologies.

Credits: Brian W. Arthur’s book: “The Nature of Technology: What it is and How it Evolves” explains the composite nature of technology. Applying the PageRank algorithm on top was a small step from the author.

ReviewRank

TechRank algorithm or more specifically PageRank derivates can be used for many other domains, not just looking at the nested and interlinked structure of technology.

In a past post we introduced the concept of IP-NFT and various ways it can be used to incentivize replicating existing research results or to generate negative information in general. IP-NFTs can power quality assurance in science and research. Refresher: in science this works so that a cut of initial funding is set as a bounty for proving that the results are not replicable, in industry this works so that someone makes a claim that the product (like medicine) is not as effective as the producer claims and if proven right collects rewards from a short position on the stock markets (short positions pay if the value goes down within its set period).

Blockchains record every step along the path. The chain records the initial research hypothesis, funding application and decision makers, verification parties, who did challenge it and how did they end up with their conclusions. Over a long period, this allows to see the quality of different reviewers, scientists and analysts.

ReviewRank shows domains where individuals are good at, how their results stand the test of time, who are the high quality reviewers and fault finders etc.

IP-NFTs are crowdsourced quality assurance for research and industrial production. ReviewRank is quality assurance for people launching and working with IP-NFTs.

ContribRank

We can expand the ReviewRank to include all meta-data around the process of creating new solutions.

Repositories track who has submitted or updated various items, when a release was made and who made it. Likewise, error reports, suggestions for improvement and discussions around them are public and usually done via repositories. Votes and funding granted to various new developments are also public.

But this is not enough. The world is clearly moving towards using digital collaboration tools, but this will take time.

In open-source design a sizeable portion is captured on the repository, but not all. There are a myriad of services on the Internet where other not contributions happen. Blogs, video sharing services, discussion forums, newsletters, independent spin-off projects that may get later get partially merged back. Not to forget in-person events, meetup groups etc. New people get their know-how physically meetings other people at all kinds of events, teams get formed, ideas are created and improved. Events are in constant flux. New ones raising and old ones eclipsing.

Or think of a movie. Teams producing films have their own tools for managing staff, not accessible to any automated tools. A lot of people participate into movie making: screen writer, director, producer, associate producer, casting director, camera crew, sound crew, costume team, catering, actors and actresses. They are today manually added as credits to the end of the film (+ shorter list as opening credits).

For a contribution rewarding system to work, data needs to be sometimes manually added directly or about sources where additional data is available and how.

Minimum is to know a contributors’ identities in different places where they are active and direct links to their content (think linking to their blog, video channel or meetup group). Some of this material can be automatically found out by scanning ‘the usual suspects’, but this type of scanning tends to favour the already big and dominant destinations. Most likely individuals need to maintain their own meta-data about personal holdouts.

The above concept is influenced somewhat by what hackmd is doing:

https://hackmd.io/MKskCuXbQT2t9s8lou_GKQ?view

TechRank as a Metric for Issuing Governance Tokens

The idea in many DAOs is to distribute the governance of future activities to the community. A meritocratic system grants more governance tokens to people who contribute more. They have a bigger stake in the project and more to lose should it falter.

The issuing can be done simply in relation to the revenue their assets or contribution brings. But for open design/source there is no or very little monetization. In these cases TechRank gives a method of quantifying the value of each contribution and could be used as an alternate metric. Not a perfect one as you’ve seen from above, but a starting point for the the gentle reader can work on in finding an improved system.

Product Meta-Data

As side note, TechRank and its siblings uncover meta-data about products that today is not readily available.

As a contrast if I buy a carton of milk, it will contain detailed description of substances it contains (protein, milk sugars, minerals). Same is true for all food products. They even contain information where the contents are from and where it was packaged. Not so for technical products.

Highly concentrated global production capability has made it possible to use access to components and technology a powerful tool for intimidation and pressure. Small and medium size nations need to be aware of the dangers and meta data where parts are manufactured is foundational data in understanding risks. TechRank discover this information as a byproduct.

Crossing the chasm with Schrödinger Fund

When there are large repositories of open-source designs and working reward models for creators, there will be plenty of people wanting to contribute.

Most inertia is in the beginning. Bootstrapping an open-source repository is a clear case of two-sided markets. No one wants to contribute if there are no users and no one visits unless there is plenty of content.

The Schrödinger Fund is a possible mechanism of starting the system and filling the repository with enough high-quality content before the systems ignites and starts running on its own.

—

In next post we look for use cases how Tech Rank can be used to drive national technology policy

Next: Keystones, silicon individualism, types of diversities (alpha, beta and gamma)