The power of knowledge graphs

It’s Friday night so I’m diving into the rabbit hole of weird/cool technology. Started by exploring Tools for Thought, then realized that I’m behind on my Anki and note reviews, started diving into that and reviewed several economic ideas such as:

  • If you cannot count what is important, you make what you count important.
  • There’s a discrepancy between the set of all possible transactions and the transactions that are realized; this is defined by the infrastructure (physical, legal, etc…) that enables efficient markets.
  • If we do not measure something, it is undervalued.
  • The basic rule, out here in normal finance, is that if a future event (1) will predictably happen and (2) will predictably increase the price of an asset, then the price of the asset should have already increased, so that when the event actually happens the effect on the price will be unpredictable.
  • Money is at the core of every society. It is how we transact. It is how we measure the value of something. It is how we understand the power and potential of our resources. Simply, money is the most interesting topic in the world.

note: these are quotes from interesting papers/articles I read, not universal laws

This made me think of SourceCred, and particularly how it allows you to measure a graph of contributions in a network. The more connections and references, the higher the value of the data. This means that someone can create data, then receive ongoing rewards for that data over time as it’s used and/or connected to. This creates a “market” for data and contributions that otherwise might be too granular of nebulous to measure. SourceCred acts as a market maker because it prices every data point on the graph. The participants in that graph (people posting/creating/commenting on stuff) then affect that price by interacting with data, raising or lowing it’s “price” (value).

Anyways, the more I think about it the more novel, groundbreaking, and essential this technology seems. In an interconnected and increasingly complicated world, being able to measure and graph the connections between things, not just the things, is essential. It solve so many problems, not just in open source, but also in social apps and potentially finance too.

  • For social apps, it could help reward contributors and give them a say in governance vs just the platform owners.
  • For finance you could tell which financial applications were providing value for others downstream (like in DeFi on Ethereum where dApps are “money legos”), which could help investors more accurately price protocols and/or asses risk.

Is there already a thread started to brainstorm all the potential use cases for SourceCred?

brainstorming for fun, not to change the roadmap and distract from the deep>wide focus of the project on first trying to serve open source communities


I wonder if this may be a more natural way to create incentives. Measure the usage of the thing you want to incentivize better, so it has more “cred potential”?

For example, to encourage development of new plugins like Discord Plugin + Cred should the plugin authors be rewarded depending on how many graph nodes/edges are generated by people using the plugin?

And for code in general, perhaps we should reward people based on how often a certain line of code is executed to measure it’s value better?


I’ve been thinking of making a new category on the Discourse, “Feature Vision” (a riff on “Future Vision” from Steven Universe). The idea for Feature Vision is that every thread would basically be a brainstorm on how SourceCred could look in the (near or far) future, and what impacts it could have on society and the world.

For example, I’ve been thinking lately that a “cred-based” approach to environmentalism could be a lot more effective at protecting natural ecosystems than the tax/subsidy framing that we’re currently working with. Everyone can flow cred to their environment, it flows through the ecosystem to keystone species, people who help protect the keystone species get a lot of cred, etc. The basic idea being that plants or species don’t have bank accounts, so they don’t participate in capitalism except as a resource to be extracted, but they could flow cred, and be vital participants in “cred-ism”. Would elaborate on this in a thread in “feature vision”. What do you think?

I know you anticipated my response here with “these are quotes… not universal laws”. But I still want to say: money is at the core of our present society. Not all societies have money.


Re: SourceCred for the environment. +1 to all that. In addition, every company or organization that uses an animal for their products or logos (California state flag, Geico gecko, sports team mascots) would have to contribute to preserving the animals that they benefit from. Using SourceCred they could flow cred to the people/organizations that support and protect those animals. The animals themselves can’t upvote or engage with people/orgs that support them, but maybe communities that are passionate about those things could rank contributors or something.

Re: money. Regardless of deeper abstract beliefs as to “what is money,” overall I think our goals are the same: help people create, share, and capture value in a way that is fair and reasonable. SourceCred enables that because it’s transparent and the community that engages with the system can change the system. This directly empowers people, but indirectly it can also help with credit attribution, maintaining the commons, and correctly pricing/measuring things that are currently hard to price or are mispriced (externalities) :slight_smile:


That’s very very interesting. Normally, you have to first figure out how to measure a thing, then create an incentive model around that measurement. With SourceCred, the measurement actually creates the incentives because then cred can flow to that data point. That’s amazing!

1 Like

I really like this framing of SourceCred. Once I get a chance I want to write an essay about SourceCred as an “incentive compiler”, and I think this really gets at the heart of how things will work: if we have an adaptable, general-purpose system for measuring value, then its somewhat trivial to use it to build adaptable, general-purpose incentives.

1 Like

Having thought about this a bit more, I want to turn the framing on its head: cred is at the core of every society and social structure. Money is one particular (highly scalable and legible) implementation of cred.

Imagine any society or social structure:

  • an open-source project
  • a military
  • a corporate bureaucracy
  • a dance community
  • a relationship
  • an art collective
  • an economy

Not all of these have money. However, I think that cred is integral to all of them.

  • In the open-source project, you get cred for valuable contributions, maintainership, etc.
  • In the military, you get cred for winning battles.
  • In the corporate bureaucracy, you get cred for attaching your name to high profile projects
  • In the dance community, you get cred for having engaging and beautiful dances with other people. Also throwing dance parties.
  • In the art collective, you get cred for making interesting, provocative art
  • In the economy, you get cred for making a lot of money (a proxy for providing economic value)

I think society has been developing in very unbalanced ways in the past few hundred years because we have a system for scaling economic cred (i.e. money and global finance) but we don’t have systems for scaling other forms of cred. So the economic sphere is coming to predominate all our other social or communal spheres. A key vision for SourceCred is to enable other kinds of cred and other kinds of values and communities to scale as effectively as economic entities do.

1 Like

This. This is so incredibly valuable. Really really want to explore more in this direction :slight_smile:

When you say “incentive compiler” is this a reference to code compilers that take a bunch of stuff and optimize it so that it’s more compact and useful (taking lots of contributions, analyzing their connections, and distilling the value of contributions to numbers we can measure), or were you thinking that it would create an incentive for people to engage with a system and create value (thus bringing people to a system/topic/problem and “compiling” their knowledge/energy/thinking into that topic/thing to produce results), or was there something else you meant here?

I really like the way you frame this. Humans create and exchange value, and cred is a way to measure that value. Money is another tool we use to measure and exchange value. The measurement tool itself does have some value/utility, but what’s important s the underlying value that it represents and enables us to work with more effectively. Not all things need to be hammered and not all value measurements are best served with the money tool. “Money-ness” is a completely different topic… but the point is that the heart of the matter is the value itself, not the tool that measures it. Losing sight of the underlying value and just focusing on the tool is when things usually get weird.

The term “economy” is a broader and more dynamic than they’re being given credit for here, but overall it’s clear that most economic models ignore or poorly measure lots of positive and negative value creation. In addition to the tools we have currently, Cred is another tool that can help us measure and manage value. At the end of the day the tools need to support and empower the people who use them, not the other way around. SourceCred has the potential to fill in the gaps where our current systems are deeply lacking. This is amazing. It’s like going from desktop to mobile, but for human coordination and value creation; it will enhance prior applications, but also create new things that never existed before.