We've looked at the root meaning of the word “token”, deciding that it indicates tools which transmit factual information intended to teach. The word “incentive” has a similarly interesting etymology, coming from the Latin incantare, which means “to chant or charm”, and the Latin adjective incentivus, which means "setting the tune".
We've already referenced a poetic translation of Chapter Seventeen of the Tao te Ching:
When the work's done right,
with no fuss or boasting,
ordinary people say,
Oh, we did it.
Ursula le Guin explains her choice of words here as follows:
“Again, it's a matter of 'doing without doing': uncompetitive, unworried, trustful accomplishment, power that is not force. An example or analogy might be a very good teacher, or the truest voice in a group of singers.”
Here, it becomes clear what incentives really are. Incentives set our shared tunes, as if by charm or incantation. They are the key to which our voices gravitate, the unseen flows which shape our collective behaviour. If the key is a harmonious one, that's wonderful, but there are as many examples of perverse incentives in our history as there are of wholesome ones.
What follows is not intended as a definitive statement about which kinds of incentives fall into the above categories - that's not how we think at Kernel - but rather an exploration of how making public and verifiable the way incentives are structured in any opt-in legal code enables us to create more persuasively wholesome structures together which nevertheless arise from and therefore continue to incentivise diverse ways of being and creating.Persuasion¶
What constitutes persuasive language? Words which perform what they say. If an essay about humour is itself funny, then that essay is more likely to convince you about what humour is.
What happens when your code says, "Transfer 50 DAI to Alice" and executes the transfer of value simultaneously? Well, you can build really persuasive incentives for any kind of economic behaviour. How do we know what behaviour we should incentivize though?
Q: Code which executes economic action can be used to build what kind of incentives?
Following on from a link we found in Andy Matuschak and Michael Nielsen’s essay, we can consider a new kind of economic essay programmed by Peter Norvig. What makes it interesting is that with a few lines of Python, Norvig proves a surprising result about wealth inequality. The initial distribution of wealth in the economy doesn't affect the long-run distribution of wealth nearly as much as the nature of the transactions. His code also suggests that constraining agents to trade only with people geographically near them makes little difference to the final distribution of wealth.
Results like these challenge your intuition. But instead of those challenges being on the basis of abstract arguments, you can engage his model. If you don't agree that the initial distribution of wealth doesn't affect long-run wealth inequality, you can try find a counterexample: an initial distribution which does affect inequality. You can experiment easily, making modifications to a few lines of code, trying to find instances where the initial distribution matters. No matter whether you succeed or fail, you will build a better understanding of the problem.
Used well, Jupyter notebooks become environments for transformative thought. They're a new media form, with different possibilities from either essays or code. And that's just a notebook, let alone a smart contract. Such contracts deployed on a shared, global network not only demonstrate new economic insights and ideas, they implement them.Programming with purpose¶
You must think like an engineer about the economic games your contracts incentivize. Ironically, the best place to go looking for clear thinking and research in this domain is neither economics, nor psychology. It is computer game design.
The reason for this is that contractual economic code on a shared, public ledger programs crowd behaviour. Don't get too excited by this, for with great power comes great responsibility - which is why the thinking patterns we outlined at the start of these modules have fundamentally to do with humility. You cannot, as an individual, know what the second and third order effects of such code will be. But you can be deeply aware that such effects exist, and you can use your awareness to inform the intention with which you write and deploy contracts.
Such intention is deeply linked to your purpose: it is no coincidence that many educators use the term "purpose-drive tokens". The purpose we can imbue our mechanisms with include reaching consensus on a network's state, contributing to lists, contributing to social media in various ways, and making it more expensive to pollute.
This last point is most illustrative: simply buying carbon credits does not directly contribute to reducing emissions. However, it does drive the price of carbon credits up, which makes it more expensive for companies who pollute to offset such behaviour. Mechanisms are not only about whom we reward and how: we only favour that way of thinking due to the way incentives are programmed in Proof of Work blockchains, and because rewards are easier to market to the masses. When, instead, we realise that tokens are only scarce if we design them to be that way, then we can begin thinking more creatively about the sorts of cost models so critical to censorship resistance on credibly neutral networks and the kinds of balanced, collective incentive structures that can keep public goods both good and truly public.
This more balanced and integral approach leads us further into software as service. When we really serve others with purpose, the distinction between "them" and "me" begins to dissolve and we can begin to understand the inner meaning of phrases like, "We are all Satoshi".
Q: In order to implement economic games like responsible engineers, what internal state of mind must we cultivate?