# Light lit review on Hypercerts and Impact Evaluators

# Impact Evaluators

#### References

Note: I recommend watching the talk from ~ 10 - 15 min mark, and reading the widget example in the Official Protocol Labs Whitepaper.

- [Internal Ceramic Thinking](https://threebox.notion.site/Impact-Evaluators-Over-Ceramic-Data-5ac733f30a454356914b22778f512381)
- [Ceramic Blog](https://blog.ceramic.network/apply-for-the-impact-evaluators-grant-on-ceramic/) same as above, just published
- Note - gives a shoutout to Govrn
- [Official Protocol Labs Whitepaper](https://research.protocol.ai/publications/generalized-impact-evaluators/)
- [Intro to Impact Evaluators Talk](https://www.youtube.com/watch?v=TdDHWv00Z4E)

Info from Holke (runs hypercerts at Protocol Labs)
- [SFF Speculation Grant Program](https://survivalandflourishing.fund/speculation-grants.html)
- [S Process Funding](https://www.youtube.com/watch?v=jWivz6KidkI)
- [CO2 Storage](https://co2.storage)
- [PL Labs Hypercerts Docs](https://testnet.hypercerts.org/docs/)
- [PL Labs Hypercerts White Paper](https://testnet.hypercerts.org/docs/assets/files/hypercerts_whitepaper_v0-3e54f05fe1358373c4f32610dd4fb391.pdf)
- [Notes from Aarons Meeting with Holke](https://docs.google.com/document/d/1qm_h43InJZYiy7zKu8uBXef2YpolOdvuyAlRR8VzdOA/edit?usp=sharing)
- [Arcological](https://arcological.xyz)

#### Christine notes

- Govrn is very aligned with impact evaluators (IE). I would argue that Govrn allows each DAO to be its own IE.
- we both aim for weights to be assigned to contributions, and then map weights to reward.
- Two key differences:
- reliance on an objective function and KPI
- IEs judge a contribution based on how much it contributed to an objective goal, judged with some KPI. Bitcoin mining is an example of this from the talk, or number of widgets created from the white paper example.
- Govrn takes a more flexible approach which I think is more realistic for DAOs; it doesn't require a singular goal with a quantifiable KPI but rather allows members to subjectively value contributions based on whatever goals they have for the DAO. I'd argue that impact (measured with a KPI against an objective function) is useful in very clear cut cases, but impact is actually much more subjective in real life DAOs).
- impact in IEs == explicitly quantified via a KPI prior to reward mapping
- impact in Govrn == implicitly backcalculated from reward mapping
- reward function
- Agree with Nick's point that IEs don't talk much about the reward function. Govrn is actually all about allowing a community to define their reward function together + emergently, rather than relying on a panel of judges (one ex from the IE talk).
- What we can learn from the white paper
- Protocol labs paper solved for how to make a finite reward pool positive sum, which had been stumping me for a while! They suggest you can program the reward pool to expand if people cooperate (essentially changing the percent of the reward pool available)
- We can consider instances where KPIs would be helpful for helping to assign weights to contributions
- This wasn't completely answered by IEs, but they were starting to think about how to evaluate if the IE was effective. We should think about this too.

#### Nick Notes

Nick's Notes:
- misses what I think are key points re: social-ness / governance-ness of value.
- I think you map this to our existing value frameworks to the Impact Evaluator framework by introducing a mechanism for voting on r, the reward function
- Or, in the case of the example in which many users assign tags to many contributions records and vote on the tags finding the reward function r is a sort of learning-from-data process over the voting/preference/governance data
- We should 100% cite this (and explain this comparison, perhaps) in detail in writing going forward. Ideally in a fully complementary, boosting up, fashion

#### Misc Notes

What is the goal of an impact evaluator?
- to distribute reward to community contributors who add value to the DAO (Ceramic Blog)

- 
- Explained at 11:18 in the video
- choose the signals from the world that you care to reward, based on a measure of how well the objective kpi is being maximized
- measure the individual contributor against what the contributed to the objective kpi
- map this to the reward
- "measure function extracts info out of the world; evaluate function translates that information into impact against a kpi per contributor; and the reward function maps the impact against the kpi to a payout"

- 

- implementation:
- dao treasury sets aside a reward budget for dao contributors
- rewards are given continuously over rounds (what we've been calling epochs)
- rewards could be governance shares (think moloch dao)

- - objectives can include both actions (leading indicators) and outcomes (lagging indicators)
- this paper really reminds me of OBDs
- recommends having multivariate objectives to increase resilience against bad actors

- reward based on how well you are doing in comparison to the goal
- if the impact evaluator is doing a good job, value creation can feed back into the reward reserve
- contributors can coordinate only if they trust the impact evaluator is long term
- there is a kpi that measures what the objective is

Questions

- What is the trusted seed?
- How do you know an IE is succeeding?
- How do IEs combine?

# Hypercerts

#### References

[Hypercerts: on-chain primitives for impact markets with David Dalrymple](https://www.youtube.com/watch?app=desktop&v=2hOhOdCbBlU)
[HyperCerts for Regenerative Cryptoeconomics by Evan Miyazono | Devcon Bogotá](https://www.youtube.com/watch?app=desktop&v=YpIQoq_bfu8)

**Aarons Comments:**
Hypercerts are an interesting superset use case on impact certificates. Both aim to "make fungible" the output or impact of certain tasks, and aim to be an evaluation on the intended impact of certain actions that take place.

Both impact certifacates and hypercerts, I'd argue are a superset or abstracted idea on top of govrn's contribution graph.
- I'd even argue that govrn's contribution graph increases the value of hypercerts themselves, via contributions nft's ability to be packaged into multiple certs.

Something that is worth exploring, are impact certs and hyper certs the signfier of contributions being linked towards a common dao, goal, or reward relationship.

I'd be interested in exploring hypercert markets that could be built on top of govrn. Hypercerts themselves, are a graph overlay on top of the Govrn graph. What gets interesting (and creates a value add for govrn), is if we can find ways to natively allow graphs to overlay our contribution graph, allow the overlayed graph to have value flow to the original contribution nft graph node(for example), and have both the hypercert overlayer earn a fee (for being a value add) as well as govrn earn revenue.

#### Christine Notes

- hypercerts are trying to solve the problem of retroactively funding public goods
- similar to govrn in that hypercerts can be periodically mint contributions
- e.g. minting your work on creating IPFS; a merge/rollup of these hypercerts would == the creation of IPFS
- impact assessors evaluate the value of a hypercert
- different from govrn in that

what we can learn:
- can govrn minted contributions be hypercerts? can we make them meet the standard?
- when someone assigns value, should this person (funder) be connected to the minted contribution
- hypercert funding is an extreme case where (I think?) the funder actually receives the hypercert in exchange for funding

 from david

from evan

^ contribution chains / sourcred type effects