Neural Governance: Misprint Correction

It turns out the Census Committee of Kekistan has been infiltrated by blind and color blind individuals. As a result, a number of NPs have been misprinted.

We believe now is the right moment to address the “misprint” issue and decide the future of NP metadata and “trait discovery program”.

Census Committee of Kekistan

Misprint Correction

Currently, the most common and obvious misprint is background color. Several dozens of NPs which have non-white background have it marked as “White”. Some NP holders also consider certain “Special” traits are misclassified as well. However, if background color can be reclassified more or less subjectively, special traits are absolutely subjective and considering the nature of AI-generated imagery abstract elements can be interpreted in many different way. What looks like “hair” for one person can be “fin” or “hat” for another.

Therefore, we’re convinced that “background color” is the attribute we should target with our misprint correction initiative. NP backgrounds have different background hints a sometimes it’s hard to choose between “Gray” and “Grayish” or “Purple” and “Violet”. So, in order not to change the rarity distribution too dramatically, we believe rare background colors such as Greenish, Violet, Orange, Red, Brownish, Yellow and Silver must be kept intact. In other verge cases ( e.g. Ultramarine vs Blue) the choice should be made in favor of more common color. Verge cases of White vs Other should be resolved in favor of White.

All 4 NPs have very different color hints, but all of them are quite correctly called “Gray”

In order to make classification unbiased, we’re going to use Effect Force. It’s a platform which lets us create a special classification task along with classification criteria and instructions and hire thousands of random real people to classify NPs. For instance, we can require processing by 5 random people. If the majority of 5 selects the same color, we implement the change. If opinions diverge, we can expand the selection.

Metadata

At the moment, NP metadata (except images) is being stored in a centralized way which has both advantages and disadvantages. On one hand, it lets us easily correct misprints and makes possible future trait discovery (if we decide to pursue this path). On the other hand, it’s rather bad for NFT longevity as metadata is bound to disappear if our server goes offline and some NFT marketplaces such as PancakeSwap NFT require storing metadata on IPFS (immutable and decentralized).

Governance Scenarios

We see 4 possible scenarios of how to tackle misprint and metadata dilemma. Each solution has its own pros and cons. We decided to break them down for each option.

A) Keep traits as they are. Lock in metadata on IPFS

Pros:

  • Almost no stress for NP holders and community as rarity ranking stays unchanged
  • Locked metadata lets us get listed on PancakeSwap NFT marketplace

Cons:

  • Even most obvious misprints will stay which may be found misleading and unfair by the future NP holders
  • Locking metadata will make future trait discovery impossible
  • Some marketplaces won’t be able to display NP names as metadata will be locked and displaying names would require a custom solution

B) Fix background color misprints only. Then lock in metadata on IPFS

Pros:

  • Most obvious misprints will be fixed
  • Locked metadata lets us get listed on PancakeSwap NFT marketplace

Cons:

  • Moderate stress for NP holders and community as rarity ranking will shift (especially, for Green and Grayish)
  • Locking metadata will make future trait discovery impossible
  • Some marketplaces won’t be able to display NP names as metadata will be locked and displaying names would require a custom solution

C) Reclassify all traits. Then lock in metadata on IPFS

Pros:

  • Full misprint correction (subject to community accepts the Effect Force classification result)
  • Locked metadata lets us get listed on PancakeSwap NFT marketplace

Cons:

  • High stress for NP holders and community as rarity ranking will shift dramatically and classification criteria and outcome may cause arguements
  • Locking metadata will make future trait discovery impossible
  • Some marketplaces won’t be able to display NP names as metadata will be locked and displaying names would require a custom solution

D) Keep NP all traits changeable through Trait Discovery Program. Don’t lock metadata.

Pros:

  • Full misprint correction (subject to community accepts the Effect Force classification result)
  • We can run Trait Discovery program and come up with new traits to improve NP classification
  • We’ll be able to keep name updates dynamic on almost every marketplace

Cons:

  • Constant rarity distribution shifting which may deter potential investors which rely on rarity score in their decision-making
  • Unlocked metadata makes PCS listing impossible (although we can enable finalizing metadata on IPFS in the future)

Voting

We’ve created a governance proposal which will be running for 5 days until 20:00 UTC, October, 24th.

The voting is weighted which means it’s possible to distribute voting power across several options based on your preference.

You can cast your vote here: https://snapshot.org/#/neuralmeme.eth/proposal/QmY1P23oHSqhXAAq5bPbpe1J4R97ZootWRCjd6oYmEijjt

--

--

--

The first AI-(de)generated art project on Binance Smart Chain

Love podcasts or audiobooks? Learn on the go with our new app.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Neural Frens

Neural Frens

The first AI-(de)generated art project on Binance Smart Chain

More from Medium

Coinlegs Black Fish Algorithm

Hydro Development Updates 24 February 2022

Understanding Diagon’s Reward Model

2021, a review of the MIM Swarm