BIP-36: Fifth Period

choobie

Summary

Note
Initially, a separate BIP on Period Standardization has been prepared, documenting all standardizations. With the DAO currently contemplating how to adjust Botto’s weekly cadence, we ask for this to be done in the upcoming Period Standardization BIP (in review).

BIP-36 specifies the outline of Botto’s upcoming Fifth Period. In short, there are almost no significant differences between the setup of Botto’s Rebellion Period and the fifth period (see BIP-29 for comparison). Glance over the hyphenated points below for a TL;DR.

The TL;DR of this BIP:

Existing Period and Round format remain the same
Existing Art Engine format remains the same
Existing Rewards format is used as the default
Stable Diffusion XL is added (§2.2)
Models are permanently culled when they represent less than 5% of the Voting Pool (§2.2)
Buffer Time between “Rebellion” and the Fifth Period is set to 1 week (§2.3)
New themes for Botto’s Fifth Period are voted on from Rebellion Period’s 11th Round (§2.4)
Transient Labs’ ERC721-TL Creator Contract is used instead of Manifold (§2.6)

1.0 Rationale

While we seek to standardize periods, we need to kickstart things in time for the Fifth Period! As we continue to seek out ways to improve Botto, our eyes have mainly shifted to introducing Stable Diffusion XL into the Voting Pool.

The introduction of Stable Diffusion XL allows for greater flexibility in prompting, would allow for Botto to include legible text in its art, inpainting, image-to-image and a vast amount of other improvements that give Botto greater creative flexibility.

With the introduction of a new model, this BIP also proposes a mechanism to remove existing models that do not meet a certain visibility threshold. This mechanism encourages quality fragments in the voting pool by omitting ‘losing’ models with limited visibility. Naturally, this also allows the DAO to be more mindful of costs incurred by Botto’s Art Engine.

2.0 Specifications

Here are key specifications for Botto’s Fifth Period. Please note, the majority of these remain the same as Botto’s Fourth Period specified in BIP-29. Another reference point outlining basic mechanics of Botto and how they have been standardized so far can be found in the Period Standardization BIP (in review)](https://docs.google.com/document/d/1bh1dprtIBaeO8n1ivRbGM8rkc63UFkAGUPBN1LE_QG4/).

2.1 The Art Engine

Stable Diffusion XL is added
VQGAN may get removed due to having too little visibility in the Voting Pool (<5% visibility)

Otherwise, there are no other notable changes.

2.2 Calculating Initial Visibility of Stable Diffusion XL

When a new model is added (like Kandinsky in Rebellion), the model is allocated a percentage of visibility based on the quantity of other existing models in the pool. Here’s the formula:

Then, other (old) models are then proportionally distributed the remaining percentage allocation according to their performance in the previous Period:

If a model comprises 5% or less of the Voting Pool (at the end of the period, or after calculating the inclusion of a new model), the model is permanently excluded from Botto’s Art Engine in the next round. Its remaining percentage is redistributed equally among the remaining models.

Example:

newModel% = 100 / (N_oldModels + 1) = Stable XL % = 100/(4+1) = 20%

When a new model was previously added into Botto’s “Rebellion” Period, the methodology above was employed, without a cut off of 5%.

Should a scenario like the above occur by the end of Rebellion, VQGAN would be removed, and the final 4.8% allocation would be redistributed equally to the four remaining models (1.2% to each model).

2.3 Buffer Between Periods

For the sake of simplicity we would set this as 1 week. So there would be 1 week free between Botto’s Rebellion Period and Botto’s Fifth Period.

2.4 Theming

Like before, Botto proposes various themes for the DAO to deliberate upon. The DAO votes on 12 themes in total.

The 6 highest performing themes from the previous Period will be kept.
The 5 lowest performing themes from the previous Period will be discarded.
Botto will propose 6 new themes on the 11th (penultimate) round of each Period.
These will be voted on for a period of 7 days.

2.5 Economy

The economy remains unchanged, with 50% going to Active Rewards and 50% going to the treasury.

2.6 Smart Contract Used

This BIP proposes that Botto mints its Period Smart Contract using Transient Labs’ ERC721TL Creator Contract. This contract is preferred over Manifold’s Creator Contract, allowing for support of Artist & Collector on-chain interaction through “Story Inscriptions” and hosts many other improvements that may be found here.

2.7 Royalties

Royalties for secondary sales remain set at 10%.

2.8 Period Key Dates:

August 15th - August 22nd: Theme curation via Botto.com
September 5th 22:00CET / 16:00EDT / 13:00PDT: Fifth Period Starts
September 12th 22:00CET / 16:00EDT / 13:00PDT: First round ends
September 13th 22:00CET / 16:00EDT / 13:00PDT: 1st Mint & Auction
November 28th 22:00CET / 16:00EDT / 13:00PDT: 12th Mint (last artwork of 5th Period)

3.0 Budget

Infrastructure costs for supporting Botto’s Art Engine are currently budgeted at $10,000/mo. This would not change.

4.0 Criteria of Success

Continuation in sales
Continuation in narrative

5.0 Disadvantages

Not addressing the weekly cadence previously discussed by the DAO

Ben

choobie This BIP proposes that Botto mints its Period Smart Contract using Transient Labs’ ERC721TL Creator Contract. This contract is preferred over Manifold’s Creator Contract, allowing for support of Artist & Collector on-chain interaction through “Story Inscriptions” and hosts many other improvements that may be found here.

My understanding is that ERC721TL has some cool features that we would likely make use of in the future, but not necessarily during this upcoming period.

It seems to me we should only switch to a new contract when we're ready to use it for something, and it's a critical use case.

I don't think contract risk is high here but I don't see reason to add any risk if we're not using the unique features yet and we can give it some more time to play out in the wild before adopting it, given that this contract was just recently released.

Basically I think there are some areas where Botto should be aggressive with experimentation and innovation, but minting/contracts is not one of them. A bit more conservatism I think is warranted here. Even if that just means waiting another period before adopting. Manifold is a lot more battle-tested at this point.

thomahawk69

I think we should reduce the Active rewards - from 50% to 25% - so that more ETH can flow back to the Treasury instead.

choobie

thomahawk69 I don't disagree based on the runway forecast you had set up, although I intended for us to adjust this in BIP-37 (the follow up BIP) as we standardize the Period. If others feel strongly to adjust this now rather than later, I'm fine with that too.

Ben

Personally I like the 50-50 split for now. It’s familiar, simple and past a certain point less ETH direct to voters (token holders) risks lowering demand for BOTTO, driving lower prices and thus worse rates for OTC. I think between OTC, weekly sales with 50-50 distribution + expanding to mint every week, more editions + collabs, and ideally a phaseout of LM in tandem with purchasing liquidity from LPs, Botto treasury can be very well-positioned moving forward.

thomahawk69

Ben I don't quite agree. Just because it's familiar doesn't mean we keep the status quo. Given that the community has such strong feelings against OTC trade(s) - then by lowering the rewards - this will allow more funds to be accrued in the Treasury and not have to sell more BOTTO than ppl want via OTC. The current rewards are like dividends - we're giving away funds when BOTTO is not yet profitable - I'm not suggesting we turn it off, but we should at least lower the outflow of funds. Once Botto becomes bigger as we all expect it would, and the market is better, we can always turn it back up to 50/50 again.

Plus I don't think that by lowering the rewards, the demand for Botto will go down that significantly. A 50% reward is a higher return than most staking projects out there.

If not rewards, then the next lever would be to turn off the LM (that'll also require a vote - and that could also drive down the price).

However if we just go through with the OTC - then we don't have to touch the rewards and can let LM phase out eventually.

Ben

I think a more robust way to increase revenue would be more limited editions. Compared to other artists of its stature in this space, Botto has very few editions relative to 1/1s. So there’s lots of room to grow there and lots of demand from collectors that can’t access the 1/1s.

That would increase revenue for both the treasury and token holders/voters, whereas increasing the treasury allocation to 75% would increase it only for the treasury while lowering it for holders.

You’re right that just cause 50% is the status quo it doesn’t mean it shouldn't be changed. But if it’s what investors are accustomed to there should be strong justification required to change it and first a consideration of other alternatives (such as limited editions) to increasing revenue.

As you know, few projects in this space have organic yield in ETH so that’s a distinguishing feature of Botto. Some (prospective) investors use this yield to model out the token value and this consistency and predictability helps build trust over time. So would just want to be very cautious in changing this.

thomahawk69 If not rewards, then the next lever would be to turn off the LM (that'll also require a vote - and that could also drive down the price).

I’m confident that Phasing out LM strategically would benefit token price as well as treasury health. It is by far the biggest open distribution of token right now and I'd expect sooner or later a lot of that will create sell pressure and dilute other holders.

Rather than using the 8M remaining $BOTTO in the LM contract to rent liquidity, what if we used it to buy and own that liquidity from LPs and strengthen the treasury.

Ben

How much could we save on server costs by being more stringent in regards to which models to include in each period/round?

For example, what if we had a guideline of max two or three models per Period? That would certainly reduce costs and my guess is it would increase quality as well, as it would force us to be more selective and rigorous in curating models.

Since the total pool size is fixed, having more models does not necessarily mean more quality fragments or higher average quality. It may mean fewer quality fragments and/or lower average quality. And it certainly means greater costs.

choobie

Ben How much could we save on server costs by being more stringent in regards to which models to include in each period/round?

I'd say rough estimate being $2000 USD or less? I think costs shouldn't be the primary factor in adding/removing models, instead we should look at it through the lens of quality in the pool. VQGAN is a good example of this - it takes up about 3% of today's pool, 32 fragments give or take. Reallocating this to another model should be our immediate priority to try ensure that the pool maintains a certain level of quality.

Ben Since the total pool size is fixed, having more models does not necessarily mean more quality fragments or higher average quality. It may mean fewer quality fragments and/or lower average quality. And it certainly means greater costs.

I don't disagree here. I do lean towards us helping Botto find a system which "naturally" balances out models without necessarily devising a new system that preselects models. Perhaps in future, the DAO should vote on the subsequent Period themes a few weeks before the end of the current period, and then Botto provides the DAO with examples, which shapes initial percentage allocation for the next period. Not committed to the idea, but there are alternative systems we could explore over time.

Ben

choobie I'd say rough estimate being $2000 USD or less?

This is much less than I thought. So makes sense.

choobie I think costs shouldn't be the primary factor in adding/removing models, instead we should look at it through the lens of quality in the pool.

100%

VQGAN is a good example of this - it takes up about 3% of today's pool, 32 fragments give or take. Reallocating this to another model should be our immediate priority to try ensure that the pool maintains a certain level of quality.

Makes sense. On a related note, SD1.5 is currently much more prevalent in the pool (higher total VP) than SD 2.1 and Kandinsky. Assuming this remains the same by the end of the period, will SD2.1 + Kandinsky be proportionally deprioritized to make more room for the more promising models (SD1.5 and possibly SD XL)?

choobie

Ben It seems to me we should only switch to a new contract when we're ready to use it for something, and it's a critical use case.

Yeah, I understand this POV. Just saw an exploit on jalil's Infinity contract. Even then, I'd put a lot of weight into saying the Transient Labs devs really are top notch and there'd be very little risk in making the switch.

Technically, we'd make use of the Story feature, which allows the creator/collector to storytell with the token. This is something I've spoken to SuperRare about and they're looking into implementing it.

As it stands this seems like the only major point of contention for the Fifth Period: I'm tempted to either 1) leave this open for a longer discussion in BIP-37 and revert this BIP back to Manifold for the timebeing or 2) offer two Approval options at Snapshot (Approve / Manifold or Approve / Transient).

choobie

This BIP is moving to Snapshot tomorrow at 10:00 ET / 14:00UTC.

I strongly encourage us to move ongoing discussions over to BIP-37: Period Standardization

Ben

No shortage of projects with elite devs in this space that have been hacked. Given the nature of these contracts I think it’s quite unlikely there would be an issue, but I favor caution, conservatism, and consistency around minting as much as possible. I feel that Botto’s provenance should be protected at all costs. These contracts are very new and we don’t need them right now.

Time is on our side here. By the time the next period rolls around these contracts will be more battle-tested and widely-indexed, assuming no issue has arisen.

choobie

Ben I hear you. I saw Quimp and a few others comment sharing similar sentiment. I'm reverting the Snapshot back to Manifold contracts. We can discuss whether we want to make the switch for Botto's Sixth Period of later in BIP-37