# Time tokens to orchestrate "Universal Digital Infrastructure"

## slide 1: Time tokens

Hey everyone, thank you for staying here all the way till the end.

This is my first presentation after I took a sabbatical just shy over a year ago. I'm very honoured to introduce some of the work done since here, today in the context of the Universal Basic Income track.

Universal Basic Income is a challenging topic, but one worth taking serious. Global poverty is not just unacceptable. Eradicating global poverty is a requirement for transitioning our world into a sustainable future. We cannot expect people to pay a green premium for their energy needs, if they need to worry about going to bed hungry.

In this talk, I want to take a step back from the problem of "Universal Basic Income" and think with you through how value can be generated that could possibly fund a Universal Basic Income.

## slide 2: delayed, balanced reciprocity

To achieve this we will unpack and build on a concept that has been called "delayed, balanced reciprocity".

## slide 3: delayed, balanced reciprocity continued

More concretely we'll introduce "time tokens" and see how they can be used as a method of social coordination.

## slide 4: a simple definition

To get started though, a simple definition can help guide through this talk.
Time tokens are tokens that get allocated to you continuously, but they also disappear again.

So these tokens by themselves are not a store of value, however if you use them smartly you might be able to generate value collectively.

## slide 5: we invented inequality

To set the stage for this idea of "social coordination to help generate value" it is worthwhile to take a short look back in history.
We often think of money and markets as the coordination mechanism that helps us organise the needs of society at large scales - beyond the 'human-scale' of family or a village.

Afterall, Adam Smith in the "Wealth of Nations" explains that before the advent of money "there was only barter economy". We learn that "we needed division of labour and the creation of a ruling elite" to organise civilisation.
New archeological and antropological evidence shows us this is not true. We have under-estimated our ancestors in the deep-past.
We now know that for the first few thousands of years after the advent of agriculture the first cities formed and technology advanced. However, these cities did not have kings or ruling politicians.
In fact, we can travel even further into our collective past.

## slide 6: oldest large-scale social network

At least as early as 33.000 years ago our ancestors invented the social technology of "gift-giving".

They would spend considerable effort to make jewelery and would travel up to hundreds of kilometers to go gift them to tribes at various distances and living in different environmental conditions.

This tradition lives on to today.

These gifts were not exchanged for goods, but instead to build and maintain social relationships. If hardships would befall a tribe, they could disperse and find shelter with the various tribes they had established bonds with.

Antropologists call this "delayed, balanced reciprocity".

## slide 7: Agenda & Bio

Hey everyone, my name is Ben. To learn more about history, you should not listen to me, but I can recommend some books later. Rather than studying history, for the last eight years I've been working on consensus algorithms.

So we now know that we've invented social coordination technologies well before kingdoms and markets were invented.

Today's talk is about a proposal to apply "delayed, balanced reciprocity" to a technical context of peer-to-peer networking, and later decentralised computation.

First though, we return to the topic of Universal Basic Income.

## slide 8: Reduce UBI

And we'll instead reduce the scope of the question.

"Can a person earn from their data, if their presence has value in a data economy?"

## slide 9: markets are a price-finding mechanism

To unpack that question, we first observe that markets are a price-finding mechanism.

While everyone has a digital trace, the problem is that separately each ones data by itself has no shared value, and the cost of copying data is marginally zero;

therefore the price is close to zero and markets don't work to filter what's valuable.

But if we can combine and refine people's data, into collective, actionable information, that could have markets for a price.

## slide 10: higher-order relevance filtering

Now we stumble onto a second problem. How can we filter The combinatorial space of actionable information is huge - and that's an understatement.

Different subsets of data can be included in different computations, and ultimately relevance is subjective to the context of the audience.

Think for example that your location data might help traffic jam information when you're in a car, or instead tell us whether a restaurant is more busy than usual.

To address this we must ensure that nodes can choose which data and information they deem worthwhile to support - according to whatever policy they have.

This in turn means we have start from the lowest layers and strengthen our peer-to-peer network as nodes might talk about very different topics.

## slide 11: knowledge should be a public good

Third problem, if we're trying to build a better world. Do we not want information to be open and free?

Here we might have a first proposal. The idea of a Harberger tax on private property was introduced to balance the benefits of private property with the interests of the public over a good.

To maintain ownership (and as such exclusivity) the owner needs to pay a tax into a fund that gets allocated to the public.

We propose that we can have a market for exclusive first-access to information streams, but with a Harberger tax to be paid into the data collective that organises the data contributors and the algorithm designers over that data.

## slide 12: from data to information

For brevity, we just note that there is no need for a data collective to internally trivial. Algorithms can be composed hierarchically, and even compete with eachother inside. A data collective can also set rules for the contributors to be able to enter and for algorithm designers on what they can do with the data.

Perhaps a better mental model would be that of a bazaar, bottom-up organised by with clear understanding of social norms for all participants. A "knowledge bazaar".

## slide 13:

As we said, the space to explore for finding useful information and how to compute it is huge. So naturally no single group of people will have all the answers.

So how can we coordinate compute across projects to avoid falling into the silo-ed mistakes of the past?

## slide 14: money creates boundaries

To call the devil by his name: money logically creates boundaries. By requiring any one particular money in one region, by construction the remainder is excluded.

This is not only true for nation states and their currencies. The same pattern has been rebuilt in the digital realm. Over an open, and money-free TCP/IP network we've built Bitcoin, Ethereum and many other networks that each require their pound of flesh.

This is not a moral statement, it is a logical one for each region to operate according to markets.

## slide 15: typing data and strict write-access

We can already overcome these boundaries fairly easy though:

by strict namespacing and strict typing of data, we can unambiguously read across boundaries. We do want to enforce strict write-access - for too many reasons to expand on here.

This gives a picture where different projects can disjointly co-exist, but their valuable output does not have to be contained to their private database.

We call this use of open-standards "web3" 😉

## slide 16: money creates boundaries - part 2

That said, we still need a measure of accounting resources that everybody has access to, and that is equal for everyone.

We do have one such quantity, it's called time.

## slide 17: time tokens to coordinate Operator-Consensus

So perhaps we can imagine a network where the currency is time, and we use it to coordinate compute algoritms over data.

Note that the old saying "time is money" is no longer holds, because these time tokens won't fit into any DEX.

So we still want the Harberger tax to be paid in the project tokens that built the data types, the algorithms, and interfaces for people to help them solve their problem.

## slide 18: finally, time tokens, or delayed, balanced reciprocity

Thanks for staying with me so long: finally what are these "time tokens"?

## slide 19: graph-fungibility

At the basis, and what's not captured in the name "time tokens" is that they are defined over a social graph. Here each edge is an attestation that you trust the other entity is real and has a single account, ie. it is not sybil.

Everyone who registers gets their own unique issuance of time tokens. At a constant rate. However, these tokens are not fungible. There is no liquidity for "random coin".

However, by asserting social relations you also declare that your tokens are of equal worth to their tokens. Taken together over a graph, we get a concept of graph-fungibility: transitively over a social path in the graph we can move tokens further away.

## slide 20: trust relations

I'll keep this brief. It turns out that social relations tend to form what are adorably called "small-world" graphs. They're neither fully random, nor boring and regular.

I'll just mention here that have a useful property. Namely, that the average path length scales logarithmically with the size of the network, all the while still clustering non-trivially.

## slide 21: temporally discounted tokens

Now for the final ingredient in our soup: these tokens don't stay in your wallet.

While you get time tokens continuously issued to you, they also fade out again.

So what happens to the social graph we talked about just earlier: think of these edges now as little springs. If you give someone time tokens, this moves the tokens around and builds up tension along the edges.

Importantly edges can even become depleted, and no more tokens can move through them.

However, by discounting the tokens over time; by reducing the value of older tokens, we fade out these tensions and the graph slowly discards older memories of unbalanced movements.

## slide 22: local sybil resistance

The last slide, but an important one. Remember how we spoke about the vastness of the informational landscape;
as a results we need different nodes to talk about different topics. If we want to enable that we need new ways to strengthen our peer-to-peer networks to shield them from abuse.

Let's think now about how this could work:

as a node I need to connect to other nodes on the internet - all I will node is their nodeId and IP. I would have no idea who this peer is.
On the picture think of the sending node as one of the big, blue circles on the right and they're trying to contact the big red circle on the left, the receiving node.

The blue nodes can send message requests at network speed to our red node, as is standard.
However with time tokens we can now require them to attach a payment-intent to each message. The red node can collect these intents for payment in time tokens, and forward them to a solver who can aggregate the intents and rebalance the time tokens over the paths along the graph.

Remember though that time tokens travel over the social connections, only and that connections can be drained - temporarilly.

What follows is that if there is an imbalance of resource usage across the social graph, any node will be locally shielded by its trusted connections and a domain wall will appear.
At every node in the graph, each node is incentivised to monitor its connections and their imbalanced use of resources, because they themselves will be paying the cost. As a result this domain wall will move towards highest resource usage and cut off endured, above-balance resource demand of other nodes in the network.

## slide 23: a moment to recap

ok. let's recap.

## slide 24: delayed, balanced reciprocity

We explained how one application of this social technology of "delayed, balanced reciprocity" can be imagined in a technical context, specifically of balancing resource usage in a peer-to-peer network.

We implemented this as "time tokens", tokens that don't hold value because they are issued to you, but will also decay away again.

Therefore, these time tokens have no value in any sense understood as money; however, because they are tied to your social trust connections, they allow you to coordinate over larger social distances to collaborate with scarce resources to produce something of value collectively.

Our proposal to generate such value is based on a Harberger tax for first-access to information streams. These taxes into public funds can be paid in project tokens that address specific problems to help people.

## slide 25: thanks to

lastly, I want to thank many people, but two friends in particular for their work on these topics.

I ve skipped over many technical topics, but reach out and we can discuss at length.