Mezcal, A New Vision For Celo

In its current state, yes, transactions aren’t a lot per block due to less chain activity.

But we will run into this issue over the long run as the ecosystem grows. Polygon’s transaction costs increased by a lot since launch. The Coinbase Earn campaign actually caused a huge spike in the transaction cost of the Celo blockchain briefly.

The idea is to experiment on a time horizon of a few years and see what works. Yes, I believe Celestia solves that problem, open to hearing other suggestions as well (I did give a few other examples). We are still too early in the blockchain space from a users-onboarded perspective. Imagine a billion users using the blockchain, will transactions stay cheap in the current design? Especially with lots of composable protocols like DeFi making multiple transaction calls and increasing the price.

A design that sustains long-term scalability and maintaining the cheap transaction costs is long term project, and it starts with experimenting rapidly on a testnet.

@techboiafrica - I’d encourage you to take a step back and evaluate whether feedback you’re getting is “mud-slinging” or a “smear campaign”, or if the feedback is coming because people are reacting to/surprised by how this initiative is being presented/championed. People are asking questions/voicing concerns from what they see/from their context.

Could @papa_raw have been more respectful in the “you’re not stengthening your position…” comment? Yes.

Is this a smear campaign orchestrated by people camp who are against Ocelot and its mission? 99.9% certain that’s not what’s happening here.

“Mezcal, A New Vision For Celo” and its contents were very surprising/disorienting for me, and probably for many others. Surprise/disorientation is going to bring a strong response. I encourage you consider meeting that response with “humans are humaning” over “these people are out to get us”.

7 Likes

@brian comments around conflict of interest were definitely flung at Yaz both here and on twitter, as well as calling us arrogant and a number of other things like calling us unethical and morally bankrupt.

Brandy just recently accused us of purposely trying to uproot celo connect and crashing parties??? This is an employee of the foundation lying about our intentions and smearing us and introducing false ideas to the community.

I think much of the community is as you described, i do not believe that that is the case with everyone here on this forum. Thank you for engaing and joining the conversation though, truly.

I also tried to clarify that we were humans and that we never expected this much attention for a blog post - so we had not yet put in all the effort needed to make all aspects of the design clear. I don’t think we can be faulted for not knowing a blog post would go viral. However, i’m glad it did, it shows. there’s real interest in this idea from a technical perspective.

Alright, I can get onboard with focusing on increasing the transaction per second rate. Celo’s TPS is determined mostly by single threaded execution performance.
There’s a separate proposal in the short term to increase Celo block gas limit to address TPS: Propose to increase Block Gas Limit to 35.000.000 gas - #2 by diwu1989

I would be very interested in running validator in your new testnet with the new concensus algorithm.

2 Likes

Wow, I definitely didn’t expect to see so much contention when I pulled this forum up…

When I saw the proposal I thought it was a good idea, especially as a testnet experiment - I can’t comment on much of the backlash as I don’t know the history behind the interpersonal aspects (and frankly it’s none of my business). But I’ll comment on the technical aspect itself:

To @papa_raw’s point about moving to Eth instead of Celestia, I think it’s fair but incomplete. By the time the experiment runs its course (I believe the proposal said ~2 years?) I expect the community will know for sure which direction things should go towards, whether it’s Celestia, Eth, or Polygon Avail. That being said, from what I’ve been following, I do expect that the data availability sampling on Celestia will likely be vastly more scalable than that of Eth (but we shall see!).

(Also I expect the testnet experiment would be quite valuable even if Ethereum ends up being the right decision)

Regardless, I support this proposal as I generally support alt-L1s moving from a monolothic to a modular architecture, and fully expect it to be the norm going forward. Current scalability is nowhere near usable for the world yet.

2 Likes

Thanks for sharing your experience - it’s very helpful.

I’m just now getting actively involved again in Celo after a ~1.5 year hiatus, so my experience is probably different than folks more involved day to day. But this is what I experienced:

  1. Mezcal announcement happens the same time as Celo Connect.
  2. Contents of the announcement feel like a radical departure from what I understood Celo’s protocol/architecture to be. They also felt very detached from the momentum/messaging at Connect. Like I said, I was very surprised & disoriented.
  3. Researched what Ocelot is. Realized it’s independent of the foundation. Also realized it got 3M Celo from the community fund.
  4. Researched Ocelot members. Learned it’s a lot of former Celo folks, including Yaz.
  5. Questions naturally form in my brain. “Didn’t Yaz go to work for Celestia? Is that why Celestia’s in the proposal? Why is Ocelot proposing something so different than how the protocol works today? Why are they announcing it as a new vision for Celo? Is this one of those architectural/political splits that happens in blockchain projects? Why the timing with Celo Connect + the “let’s pile on to CeloOrg’s stuff on Twitter” social media strategy? Is this really what these community funds were intended to be used for? Is this team a group of former Celo employees who are disgruntled/unhappy in some way? Did they raise community funds to go attempt to create a parallel Celo?”
  6. People share feedback in the forum. Intermingled with feedback to people’s questions/feedback was an undertone of “us vs. the Foundation” that was surprising/unexpected.

I’m gonna guess others had a similar experience.

Being humans, people will fill in these gaps/questions with assumptions. Sometimes people will act on those assumptions with accusations/certainties, which is very difficult to resist, but I agree isn’t fair to you/Yaz/Ocelot.

I hope me sharing my experience shows you how folks could land where they landed. There were a few things that looked like an outline of a duck, so people started calling it a duck. Not saying that means it actually is a duck. But there were things there that led to natural questions.

As you alluded to, I also hope this experience is a growth opportunity for Team Ocelot, hopefully learning more about the great power and responsibility you have. The projects you work on and the announcements you make signal to the ecosystem & the broader web3 community what’s going on at Celo & if Celo is a good bet for their project/investments.

HTH.

9 Likes

Yes, I sold a lot of celo tokens by Ocelot’s “celo’s vision”.
Now I know celo is still L1, so I bought more

I was also surprised by the framing as many have commented above. It read like some sort of “official” announcement rather than experimentation by a third-party and I was really disoriented with the message and timing.

My reaction initially was quite visceral and that gave me pause to reflect on why. I’m afraid to say I completely missed the discussion and voting on the governance proposal to create the treasury that became Ocelot. So I went back and read in detail the CGP and the Mezcal proposal is not out of line with the stated goals of radical experimentation that was voted on and passed muster through Celo’s governance process.

Experimentation and diversity of thought are very important to me in this space. The entire ecosystem is a giant pilot at this point, over a decade in. Moving fast and breaking things (in a safe space) should be encouraged. Engineering results should speak for themselves and should rarely have emotional valence. Mezcal (and other Ocelot technologies) will either be useful or not.

The second topic I see in the discourse here is a philosophical discussion about what “official” / “third-party” / “community” / etc mean. Ocelot knows they’re ruffling feathers (I would guess), and my gut feeling is this is part of the point. Shake the tree and start conversations about the nature of decentralization, “sanctioned” projects, the mission of Celo in general, and “who is a blockchain?”

Long term this kind of thing will be good for the ecosystem if we can get over these communication hurdles. After reading through this post multiple times, I’m fairly sure everyone here wishes Celo as a technology, and a mission, to succeed.

Yes, the framing was eyebrow-raising at least and I’m pretty sure this conversation would be focused on the technical merits of a new canary network configuration right now rather than this meta-conversation if it was handled a little differently.

This would be have been perfect.

Steel-manning Ocelot here, I’m going to look past perhaps a poorly worded article and evaluate the tech as it comes (or doesn’t).

10 Likes

@YazKhouryTest thanks for the proposal. And thank you for answering some of my questions at the CC after party! I still have a few lingering ones:

(1) if we were to move to Celestia, would the Celo Reserve lose its ability to enable the Tobin Tax and collect a portion of block rewards? Both of these are emergency mechanisms to bolster the reserve if it were to approach undercollateralization and should ideally be preserved or replaced with other cash flow levers.

(2) given that CELO is used to secure and govern the network (in addition to the celo reserve), I would imagine this proposal would significantly reduce the inherent value of celo (for the foreseeable future) if it were to pass. Consequently, this would be bad for the Reserve, given that CELO makes up ~50% of its holdings. Do you agree? If not, why?

(3) I agree with others that this proposal is a pretty significant change in direction for Celo—which is okay :slight_smile: but I ultimately see this as a decision between the following: do we believe that Celo stablecoins will scale faster as a layer 2 on Celestia or with Mysten Labs’ upcoming upgrade (eta end of year). Do you think this is an accurate distillation of the core decision we must make?

Would love to hear your thoughts on these questions. And perhaps you could get some of the other multisig signers to share their thoughts as well—I know there are many :slight_smile:

5 Likes

Greetings folks. I’m DC, member of the Celo community since the incentivized testnet. I cut my teeth in web3 as COO of the Electric Coin Company (creators of Zcash), and am active in a few (primarily privacy and security focused) Cosmos ecosystem projects. Prior to joining ECC I was CSO at SendGrid, and prior to that was founding CEO at JumpCloud. While I’m excited about Celestia coming online, I was surprised to see this proposal suggest that we move Celo (which has been battle tested for some time now) to be a L2 on a yet-to-be-launched project.

Having been involved in Celo from the early days, I believe that the project has a ton of potential, and that much of that potential is not yet realized. I believe that in 2022, consensus is becoming commoditized. Celo consensus uses an innovative Proof of Stake consensus protocol that offers low block times, one block finality, high throughput, and compact SNARK friendly signature schemes (through BLS aggregation) that make light clients feasible. Despite all the novel research and engineering that went into building this protocol, Celo’s mainnet has never gone down since its launch close to 2 years ago – no small feat especially when compared to other proof of stake networks that have launched in similar time frames. The recent halting of the Baklava testnet was concerning and frustrating, but is by no means a reason to throw in the towel on Celo consensus overall.

The proposal focuses primarily on Plumo and Phone Verification as reasons to move towards Celestia. At first glance, this seems strange as Celestia is a protocol focused on scalability and data availability. Furthermore, the proposal calls for a test network to be set up before Celestia is live which I don’t fully understand. Without Celestia, Mezcal will have to do its own transaction ordering and data availability in which case it would look just like the Baklava testnet.

I participated in both phases of the Plumo trusted setup and hold the cryptographers that created the protocol in high regard. I agree that it’s frustrating that it’s taken so long to get Plumo from paper to protocol to production, but this is frequently the case with novel cryptography. Again, the slow progress with Plumo is no reason to abandon the Celo L1, particularly since Plumo has finally shipped.

My understanding of Celestia is that it is focusing on increasing the data availability guarantees to light clients running on L2 nodes and not targeting mobile phones. Since Tendermint (Celestia’s consensus protocol) does not use signature aggregation or SNARK friendly curves, its light client is significantly heavier than Plumo and may not be directly useful to dapps or wallets running on resource constrained devices such as mobile phones. Note that this is not a criticism of Tendermint or Celestia. I’m just pointing out that Plumo solves a different problem than what Celestia and Tendermint are focused on.

Another distinguishing characteristic of Celo that got me so excited from day 0 is the decentralized phone verification protocol. As we strive for better UX to drive mass adoption, binding wallet addresses to phone numbers is critically important. As an operator of Celo attestation service nodes, I recognize that there is room for improvement in this area, but I don’t see how moving Celo onto Celestia would help at all.

Finally, I’ll go ahead and point out the elephant in the room here. The Celo community funded Ocelot (née the Celo Ecosystem Treasury) with the stated objective of “aggressively positioning Celo at the forefront of the crypto space. To this end the Celo Ecosystem Treasury will energize activity around research and development in open-source, as well as through local communities and special projects utilizing this technology.”

The community voted to entrust Ocelot with 3M CELO for this objective. The areas of focus stated in the proposal are spot on (identity research, privacy, improve EVM, support Geth, etc.) However, one of the original proponents of the Treasury Proposal (Yaz) then left his devrel/community role at the Celo Foundation, took a similar role at Celestia, and floats a proposal to use Celo Ecosystem Treasury funds to move Celo to be an L2 on a not-yet-live L1?

As a longstanding member of the Celo community, this feels like a hijacking. I’m all for radical experimentation, but this proposal suggests that we build atop unproven tech to solve problems that aren’t on the critical path (yet), relying on relationships with curious optics.

We should be leveraging the Celo Ecosystem Treasury to make incremental progress on the things that we know need improvement. We should not use Celo Ecosystem Treasury funds to fix things that are not broken with Celo on top of a pre-launch, untested L1.

I’m excited about Celo, and I’m excited about Celestia, but I do not support this proposed use of Celo Ecosystem Treasury funds.

19 Likes

I don’t think anyone can predict how these experiments will pan out. If they want to set up the celestial based incentivized testnet, then let that experiment proceed and at the end of the year, we’ll see how things shake out.

At the end of the day, it’s all up to where users take their apps / blockchain usage / economic activity that dictate which direction is the right choice.

Personally, I don’t ever use the current Celo testnets because:

  1. there’s no real economic activity going on there
  2. users aren’t really on the testnet
  3. mainnet is effectively my testnet because why not?

If this new Celestial testnet will have some actual incentives for participants (using that 3M celo grant), then it does give me real motivation to learn and try it out, and if the Celo foundation grant was already voted and handed out, they are free to use that as incentive for the new testnet right?

Or put it another way, if their new testnet has no incentives at all, I doubt many people will put up new infra and tooling just to play with an un-incentivized testnet that doesn’t have any real economic activities at all. It’d just be a dead testnet dead-on-arrival, so it make sense for them to put up some rewards to attract early adopters.

Edit: for transparency, I run liquidations, limit orders, and balance out all of the DEXs on Celo, so having actual incentivized testnet playground is beneficial in my perspective.

Hey All - just want to step in to reflect on how we have 2 different conversations happening in this thread:

  1. debate around how something should be communicated to the community - including language, tone, by whom, as well as what “official” means in a web3 world
  2. the actual proposal at hand - creating an incentivized testnet for radical experimentation.

I will be focusing here on #1.

I’m honestly not surprised to see #1 dominating this thread, since radical experimentation requires going outside of comfort zones and pushing edges, which can be scary and threatening. My experience is that the Celo community is generally conflict-adverse, more often than not suffocating diversity of thought and perspectives. This is to be expected in a space where we are are at the very forefront of technology, making up the rules, and still rather nascent as a community. For better or worse, we create our culture (and identity) every day through our contributions. It’s quite the responsibility.

And since I also contribute to that culture, I’d like to invite in reflection around our principles and values. I personally would love to see this community grow together to:

  1. be more tolerant and open to new ideas (and as said earlier by @Thylacine encouraging moving fast and breaking things)
  2. make it a habit to first reflect on emotional triggers, assumptions, and projections before engaging in conversation – especially in the imperfect online world we find ourselves in
  3. adopting the principles of Non-Violent Communication, which is about sharing observations instead of critiquing individuals
  4. assume positive intent - I think it’s safe for us to do that here in a community where our mission is global prosperity… :slight_smile:
  5. actively dismantle our own internal conditioning that enables a dominating power within a hierarchical system at the expense of the majority (aka, a system of oppression).
  • Somewhat relatedly, what is “official” communication in a decentralized Web3 community? Consensus? Does it even exist? Given this word came up several times as a point of confusion, it feels important to discuss.

My own opinions on the matter -

  • If you initially felt a strong reaction to this proposal, I invite you to reflect on why. What feels threatened?
  • Ocelot was created for the purpose of radical experimentation, which - if we are to succeed in the Web3 space - we need to embrace and create ample space for. This proposal is doing just that. What’s the problem?
  • I’m unsure why so much heat - as well as accusations of “self-dealing” - has been projected towards @YazKhouryTest who worked tirelessly as the #2 hire on Celo Foundation devrel for well over a year supporting community calls, governance, and protocol-level engineers. It’s because of him that a “core community” team was formed at the Foundation. I would have expected more respect towards him from people who know him here on this thread. Again, not sure what feels threatened? For the record @bcamacho, Yaz attended the Celo Connect afterparty at my invitation, as he is still a part of our community despite choosing employment outside the Foundation.

If others have suggestions as to how to address the issues underlying #1, I’d love to hear it. Perhaps this warrants its own community call.

6 Likes

@Adriana, thank you for the insight. TBH, I was confused by the proposal and timing of it. I was sensing division versus unity. Appreciate your work to keep us on track for meaningful conversation.

I was sharing my opinion on how the announcement was viewed from my perspective. It was confusing to me. When I reviewed this discussion, others were confused as well.

@techboiafrica, I value trust. My message was not a smear or a lie. I’m human with an opinion, just like the rest of us. Let’s respect one another and understand that there are times when we all make mistakes. I’m learning from this experience. Again, I’m just a little confused with all of this.

1 Like

I would like to re-iterate a question that has yet to be answered here:

How much engineering talent and how much of the grant fund is being put towards this single initiative?

For folks who see this as a possible instance of self-dealing these facts would help establish a baseline truth. From my perspective, if the amount allocated to this initiative – specifically in relation to Celestia-related work, as that’s the primary “self-dealing” concern of most folks – is significantly higher than the other grants issued by Ocelot, there is a compelling argument for self-dealing. That being said, if the amount is roughly proportional to other grants issued, this kind of argument has much less of a basis to stand on (I don’t personally see it as any great issue to allocate a similar amount to explore a potential technology, even if a multi-sig member is working for that project).

@Yaz If you could provide these numbers, thank you!

1 Like

Hi @alchemydc,

Thank you for the engaged and reasoned dissenting feedback here. Personally, I can’t tell you what a relieving breath of fresh air it is to see, as well as the sentiments expressed prior starting to shift here.

I hope you’ve been well since we got to interact with the Plumo setup. I, too, have deepest respect for the brilliant team of cryptographers behind Plumo’s design, and am so excited for them to enjoy the fruits of such hard work. I am certainly rooting for Plumo, and the question myself and many others are asking is why hasn’t it been demoed on Valora yet?

“The proposal focuses primarily on Plumo and Phone Verification as reasons to move towards Celestia.”

The post focuses on these points as they’re reasons to reconsider monolithic L1 architectural decisions, not as reasons to move like you’ve described. It is a subtle but key distinction.

“We should not use Celo Ecosystem Treasury funds to [xyz]”

With utmost respect, this and related points are antithetical to the whole point of why the treasury exists — in simplest terms, the decisions of what treasury funds should be used for are the purview of the multisig (and ours counts 11 brilliant people). That’s the whole point, because it’s the treasury, and not the community fund.

We created the treasury with the consented premise that the multisig has agency. It’s the same agency that was afforded to the Celo Community Fund 1 project, the same agency that the Celo foundation or Flori Labs utilizes when they opt to be held accountable to the community — all of whom made or make funding decisions a priori with no such analysis nor scrutiny.

For those who’ve expressed curiosity, the budget for Mezcal is not finalized and we first need to do further scoping. When that’s done then we would share the precise costs with the community, as we’ve always been happy to do. Right now, the working estimates for the testnet amount to ~5% of the treasury, which is about the same ratio as the ~5.5% total for grants Ocelot has committed so far. Again this is still preliminary and in an ongoing scoping process — and thanks to our collective relationships we’ve been thrilled to have a priority engagement for getting the testnet built.

“We should be leveraging the Celo Ecosystem Treasury to make incremental progress on the things that we know need improvement.”

My understanding of this sentiment is that your position affirms a more conservative approach to distribute funding. There is nothing inherently wrong with that, and please correct me if that read of mine is wrong. But this is not why Ocelot was created. While I respect your stance, it is not the approach we adopt, nor is it the kind of approach our commitment by requesting the treasury consists of — plainly, it is not our mandate. And the agency we should be afforded means that we should be able to adopt a more moderate or even radical approach to value generation.

This is a very important point that I want to reiterate here for emphasis. Whether the Mezcal testnet is funded by the Ocelot treasury or not, this approach is part-and-parcel of the mandate, spirit, and goals for Ocelot in general.

I, too, have been part of the Celo community for around 3 years now. And many people who I care about and respect deeply have been around here for even longer than me. Anyone who knows me, knows I’ve been sounding this drum since the day I joined Celo. The decentralization drum. The Celo protocol was created — and it’s also the story we have been continuously told — to not be run by just its same original “powers” nor dependent on the original team. And in that same vein, the Ocelot ethos is anchored around these same values. The charter of Ocelot is open source, decentralized community, emerging regions, and radical experimentation.

And an experiment, like a testnet, like this one, falls under the scope of radical experimentation (as some others have likewise keenly noted). We believe this to be a worthwhile experiment. Especially with a currently monolithic L1, radical experimentation is needed to scale the mission — because we need innovative projects that raise the bar and push the space forward.

This is the zeitgeist of our industry.

Because lastly, this also allows us to be thinking about something that we all haven’t necessarily touched on here yet, which is opportunity cost — if something is exciting to the wider web3 community as a whole, something that went organically viral, then that is a good vector for something that would be appealing to them to build on Celo. So here we have a prime opportunity for including and appealing to people who would’ve or would be part of the Celo community from wider web3, but right now, are not.

2 Likes

After attending the Ocelot community call I am now sure of the good intend of this proposal, which I was previously not. I don’t agree with the framing and communication of this proposal, but with the underlying goal of keeping gas-costs low long-term.

I don’t see any negative implications from Ocelot spinning up a testnet. I believe in radical experimentation as a way towards innovation and progress, and who knows what comes out of this. Without further commitments from the Celo community this seems likely to be valuable for the attention from the wider web3 ecosystem, especially developers, alone.

But I do wonder about the robustness and risk implications if Celo would become a Layer2. While I am in favour of this experimentation, I believe that centralising a lot of DeFi on only a few Layer1s is a fragile long-term approach. Zooming out a bit to take a look at the DeFi system in general, I believe having the redundancy of multiple independent Layer1s would a key improvement compared to the traditional financial system. In engineering and complex systems, redundancy of critical components provides a great degree of robustness by reducing single points of failures, which previously human engineered systems have often neglected in favour of short-term cost and profit optimization. Some good examples are probably the global financial crisis 2008, where ‘too big to fail’ banks made the crisis and cost to taxpayers significantly worse. Or the global supply chain crisis emerging with Covid-19. Or the 2021 Suez canal blockage.

From that perspective, I see validators, data availability and consensus not as burden to worry about but as a luxury to maintain.

5 Likes

Just wanted to quickly chime in here as there were a few questions asked that I know the answer to. This proposal pegs itself as a radical experiment that will help Celo with its mission. While I’m excited about experimentation and bringing new people in the the ecosystem, I’m struggling with two areas as they have been pitched:

Firstly, the repeated mention of Plumo and the Celestia light client makes it look like the proposal is proposing a solution that will work better for mobile phones than Celo in its current form. This is puzzling because as @alchemydc mentioned, the Celestia light client is not very light at all, certainly not light enough to use on mobile phones or web apps. So far, nobody has responded to this point other than @gabrielllemic who asked why Plumo hasn’t been demoed in Valora yet. The answer is simple: Plumo was only launched two weeks ago and I thought that demoing it on a web application (Celo Wallet) was much more powerful since mobile web apps are even more resource constrained than mobile apps. In addition to the more data heavy light client, by eliminating consensus on state roots, it appears that Celestia actually makes it harder to build decentralized mobile applications that can fetch and trustlessly verify chain state with merkle proofs. Now I’m not criticizing Celestia, I think it’s very interesting but it does look like its design choices are not as focused on mobile devices as the Celo ecosystem is. Given this, it does not feel like an obvious choice without more analysis.

Secondly, the proposal is proposing the following sequence of events:

Phase 1: Instantiate the Mezcal network with validators and a snapshot taken from Celo to help onboard the ecosystem and dapp developers to utilize the incentivized testnet.

Phase 2: Grow the ecosystem with incentives for developers to build on the Mezcal testnet first in order to test out new features. The Celo Reserve can also test the new tokens being deployed for diverse collateral.

Phase 3: The Merge, moving Mezcal Incentivized Testnet from an L1 to an L2 Rollup on Celestia. Phase 3 will take some time as it requires waiting on Celestia Mainnet to go live.

The first two phases appear to have nothing to do with Celestia and therefore may not add that much value to the Celo ecosystem and instead may even split attention away from the existing Celo networks. Usually, when you want to launch a new network with new features, you develop that new network before launching it, since it is much much easier to do so before it’s running. In this case, the development only happens in Phase 3 after Mezcal is live, requiring a difficult Merge operation that feels needlessly difficult to take on for a network that doesn’t exist yet. If having an incentivized testnet in the Celo ecosystem is the goal, then we already have two testnets to pick from that we could convert to take on an incentivized form.

Given these two points, it’s hard for me to understand the goals behind this proposal. If the goal is to radically experiment with new scalability solutions, I would be much more in favor of focusing on development and benchmarking before launching something. It would be great to use these benchmarks to compare the proposed approach with the current Celo roadmap that uses Narwhal for ordering transactions in a scalable manner.

10 Likes

Catching up on the thread here a bit late. I haven’t dug too deep into everything here but overall I share a lot of sentiments that @alchemydc mentioned. Choosing new completely unproven technology as a new experimentation ground and making a huge technological shift doesn’t really have much justification.

Also, I am not sure I agree with the overall premise that Celo doesn’t need to be separate L1 network. Both economically (for cXXX stable coins) and also overall technology wise, I think it is pretty important for Celo to continue to innovate in the core blockchain part.

Even as a high level idea, I am not really sure it is a great idea to separate “consensus” and “ordering” from “execution”. Successful L1 that can support real DeFI activity will require quite a tight integration across all of its stack to properly support needed throughput, QoS, availability and correctness.

Just to give an example, TX/s that a network can support isn’t hugely relevant, because no matter what potential TX/s throughput you have, it can all be filled up by bots if you don’t have other mechanisms in place to manage TX traffic. (As other L1-s are finding out with their fees increasing very quickly as soon as there is DeFI activity on the network).

4 Likes

The vast majority of the TX on Celo’s network are already bots, not humans. The ratio is like 20:1 or higher already.

There are 2 types of bot activity on Celo today:

  1. top of block priority gas bidding, there’s 3 or 4 players bidding on these, their bidding gasPrice is capped by the available arbitrage profit left over from the previous block
  2. backrunning bots listening to mempool and tailing human trades that cause unbalanced pairs

Over the last 6 months, there’s been a shift towards more of type2 bots vs type1 bots. They issue way more TX into the network than type1, but cumulative gas fee is actually drastically less because type2 backrunners do not bid up the gas price to insane levels. They are always submitting TX at the exact same gas price as the human trade.

Digging a bit deeper into what’s happening between competitive type2 bots, it’s converging to a winner take most dynamic where there’s one set of bots that have the fastest reaction time and accuracy, and is able to extract the bulk of the profits with just a few TX.

Other slower backrunners are not capturing profit and are just wasting gas.
Example block from yesterday https://explorer.celo.org/block/12569894/transactions has a $5k trade triggered 14 backrunner TX. Block was stuffed 2/3 full.

Swappa bots SwapExactInputForOutputWithPrecheck and 0xb24e6ae6 captured all of the profit, resulting in no profit for type1 priority gas auction bots in the next block.

Pro this means gas bidding on Celo will be kept to a minimal because the amount of profit to be extracted at top of block is minimal
Con you have at least a 10x+ ratio between human TX and backrunning TX, this is what’s eating most of the block space

My guess at how the block congestion will play out is:

  1. The dominant backrunning type2 bots will keep innovating on efficiency and drive down the number of TXs necessary to extract the profits (extra TX still cost the bot a little bit of gas)
  2. This is a low latency competition at the milliseconds level, only the fastest one wins and the rest are just spamming the network for no gains, hopefully the losers realize this and leave
3 Likes