Proposal: Add support for Golem Network Token (GLM) (updated version)

[neptune]

@Mat7ias

The protocol and application of Golem are mostly off-chain where GLM is used for payments on-chain on the Golem Network (there are payment drivers for Ethereum, ZkSync and Polygon). Golem is very different from Aave, for example, where on-chain governance is much more beneficial (since many of the components are on-chain). Golem is a bit more comparable to geth and nodes running the geth implementation. Naturally in that comparison, users similarly have the opportunity to decide not to run an update / upgrade. Geth development itself doesn’t necessarily need governance when anyone can fork or decide not to run an upgrade. The Golem Network can be governed by social consensus comparable to Ethereum and Geth.

The main on-chain governance potential for GLM would be in the token contract. Golem Factory GmbH does not have any control over the token contract. The community does in a sense since they can migrate their GNT → GLM by using the migratory agent which has no ownership (i.e. decentralized). Ownership was renounced prior to the migration start announcement and migration is purely opt-in, based on the consensus of the community (here are the stats for migration).

It’s also important to keep in mind that any holder of GNT has the potential to migrate that GNT to GLM at any given point in time (migration is permissionless and censorship-resistant, based on smart contracts). So let’s say a DAO was created in the future, any GNT holder would have the potential to instantly migrate (within the blocktime of Ethereum) and vote on said DAO, therefore the GNT distribution can’t be ignored in a potential DAO or in a scenario that governance was to be created.

I’m sorry, I really don’t want to come off as antagonistic but isn’t then the only use case for GLM to pay for/be rewarded for processing power? There is no DAO, there is no on-chain governance? There is basically no reason for someone to hold this token, unless they’re planning to purchase processing power? So the only borrowing use case would be if I want to purchase processing power right now and speculate that the price of GLM will go down in the future, so that I can rebuy and repay at a more favourable $ value?

The token contract being non custodial is good. But I don’t quite get the point of voting by migrating? Wouldn’t it then be a bit concerning that within a full year, only 55% have migrated and the rest have seemingly forgotten about their holdings? And how would you integrate that into a risk analysis? A risk analysis is in ANY case just a snapshot of the current risk. It can change in the future, yes. But that isn’t what’s being evaluated. Right now GLM has 14k holders. GNT holders are irrelevant. Perceived lesser risk in the future is irrelevant.

@Mat7ias

Golem Factory has control over the GitHub repository in a comparable way to the Ethereum Foundation has control over Geth, which is open to be forked or releases can not be upgraded to if the community deems that the prefer an upgrade not to be representative of the community. In the Golem Stats there are users who decide to upgrade or not to upgrade from previous versions. Social consensus and the ability to fork at any time is as an important factor in Golem as it is in Ethereum. The Golem Network has 14% (426) of the number of nodes as Ethereum (3003 ) so social consensus should be considered very significant.

The social consensus of the grid computing platform is not at issue. It doesn’t really have anything to do with the token. Counterparty risk assesses if some entity or governance attack would be able to mess with the token contract. An example would be infinite minting of GLM. If Golem Factory GmbH has no control over the token contract - who does?

[Mat7ias]

@neptune

after 5 years are still in beta?

Just going back quickly to the point about 5 years are still in Beta since I missed responding, this is also similar to how Bitcoin is still in beta, it just means that development is ongoing (it will be in beta until there are no developers with features to add). In this sense, being in beta is a good thing.

Golem is up to its 2nd implementation (Yagna) which itself has had 4 major releases and a number of minor releases between each.

The first implementation of Golem, Brass (later code-named Clay), includes more than 20 major and minor releases (not including Alpha releases, link to the Brass and Clay communications archive).

@neptune

but isn’t then the only use case for GLM to pay for/be rewarded for processing power?

In a way, although Golem’s potential goes much beyond a singular application. In Golem there’s not a single use-case - it’s a self-sustainable ecosystem. There’s already a myriad of different types of applications on top of the platform, built by the Golem community. For example, password recovery, gaming backends (e.g. Chess), data analysis, simulation and optimization, tooling, etc. Linking Awesome Golem for the entire list of community created tools and applications.

@neptune

There is no DAO, there is no on-chain governance?

Exactly, the consensus is derived from the nodes of the network and developers (comparable to Bitcoin or Ethereum).

@neptune

Counterparty risk assesses if some entity or governance attack would be able to mess with the token contract. An example would be infinite minting of GLM. If Golem Factory GmbH has no control over the token contract - who does?

No individual has control, the GNTMigrationAgent contract has control over minting GLM (in exchange for GNT burned 1:1). The original post includes audits, both procedural and smart contract focused.

• audits both procedural and smart contract focused https://blog.golemproject.net/gnt-migration-audit-report/ https://github.com/golemfactory/gnt2/blob/add-audits/CertiK - Golem Audit Report 1.1.0.pdf https://github.com/golemfactory/gnt2/blob/add-audits/Golem - New Token and Migration App audit report.pdf https://github.com/golemfactory/gnt2/blob/add-audits/TrailOfBits - Golem Executive Summary.pdf

@neptune

borrowing use case would be if I want to purchase processing power right now and speculate that the price of GLM will go down in the future, so that I can rebuy and repay at a more favourable $ value?

Being able to speculate that way on the cost of computational resources would be quite beneficial and something that doesn’t exist currently. I would consider this very significant in our world where computational resources are being bought and sold more every day!

@neptune

the point of voting by migrating?

It can be viewed as making the entire migration opt-in and not centrally forced. It’s not necessarily a vote. The old implementation (Brass / Clay) is permissionless and censorship-resistant, technically speaking it’s still possible to use.

The main friction stated by the community around the migration is fees on Ethereum, so it’s slow because a portion of the community prefer to wait for lower fees. There’s not been anyone who has stated that they still want to use the old Golem implementation over the new one.

@neptune

And how would you integrate that into a risk analysis? A risk analysis is in ANY case just a snapshot of the current risk. It can change in the future, yes. But that isn’t what’s being evaluated. Right now GLM has 14k holders. GNT holders are irrelevant. Perceived lesser risk in the future is irrelevant.

Understood! So let’s agree there and evaluate it as a snapshot of GLM :)
I’ll create it and tag you for review if this is okay with you?

By the way, do you happen to know if any other risk analyses have included holders and transactions on Layer 2 (e.g. ZkSync or Polygon)?