Catalyst isn't Broken, it is Unsolved

A few days ago, the Catalyst Fund 10 results were published, and this has caused quite a stir within the Cardano community.

At the time, I was on a small personal vacation, attending the fantastic StrangeLoop conference. I didn’t have time to organize my thoughts on the topic, and so I tweeted out that a thread would be forthcoming. As I went to do so, it became very long, and it seemed like it would be better suited as a blog.

Before we dive into it, let’s get a few housekeeping items out of the way:

• First, congratulations to every team that received funding and plans to push the ecosystem further; We’re right there in the trenches digging with you, and we know first hand how hard it is.
• Second, deep condolences to anyone who worked really hard on their Catalyst proposals and ultimately didn’t receive funding. Fundraising in public is an incredibly vulnerable thing to do, and I really commend those who keep building; some of the best pieces of technology in this space are from projects that didn’t receive Catalyst funding, but went on to build anyway.
• Third, I want to acknowledge the deeply positive and difficult work that the Catalyst team, and Daniel Ribar in particular, do. You consistently turn up with positivity in an extremely negative environment, and I do believe that you are good actors in the space.
• Finally, thank you to everyone who voted for the Sundae Labs proposals; We did not expect to get all 7 funded, but we’re really excited to deliver substantial value to the broader Cardano community across all 7 proposals.

With all that being said, it’s clear to me that Catalyst is currently deeply flawed. While it accomplishes some of its objective (as many very useful projects have been funded and delivered value to the ecosystem), there are clearly ways it falls significantly short. I’ll share my thoughts on what these areas are, but the central thesis I want to convince you of first is that Catalyst isn’t broken, just unsolved.

I know that this blog post will be tainted for some by the fact that Sundae received funding for its proposals. Perhaps that overlays a bias on me, but I don’t think so: I’ve been consistently optimistic in the space, and that’s not limited to this issue alone. Additionally, a significant portion of this article is devoted to criticisms of the Catalyst process. The point I’m trying to get across is that I believe it is far more productive to recognize those flaws in a healthy and constructive way, engaging in good faith to try to solve them, compared to some of the defeatism I’ve seen.

In particular, what Catalyst is trying to accomplish is extremely hard; In fact, given the historical track record of societal attempts to disrupt power, it may be one of the hardest things that people have tried to accomplish. It has always been the case that wealth is an accumulator, and there has never been a wildly successful system that is truly “anti-whale”.

This is why we say that we are attempting to “push power to the edges”; because that implies that it is a very active and difficult process. Power is already centrally concentrated, and needs a very broad and collaborative effort of people to disrupt.

Catalyst is one attempt to do so, and will be significantly flawed as we figure out how to solve this problem. But, much as you wouldn’t describe a partially assembled engine, or an early prototype “broken”, it is unfair to apply this label to Catalyst. If we throw up our hands on the first (or 10th) attempt that doesn’t work towards the goal, then we have allowed that coagulation of power to succeed.

And in fact, despite its flaws (which I will talk about soon), I believe Catalyst is overall, fairly succesful. Contrary to a narrative that I’ve seen on Twitter, as I skim through the results, I see a healthy mixture of established companies with a track record of delivery, along with new, exciting, and innovative projects. Just because the project you were most hoping for wasn’t funded, or just because the funding is a bit top heavy in some areas (I didn’t not expect MLabs to succeed on 14 proposals, or Sundae Labs to succeed on all 7, for example), doesn’t mean that it is a complete failure, or completely fallen to corruption.

I’m not sure how someone can say, for example, that Catalyst is the most centralized funding platform that exists, when close to 100% of all investments that have ever been made have each been the decision of a small number of people, be it a private investor, a hedge fund, a VC firm, etc.

Catalyst certainly seems subject to outsized influence from large ADA holders, but it still invites participation from an extremely large body of voters. As with all revolutions, the power held collectively is far greater than the power that is in control; it is only through inaction and apathy that that greater power remains unleveraged.

For example, of the roughly 1.3 million staked wallets, representing 22.79 billion total ADA, only 60,367 of those wallets with 4.4 billion ADA actually registered to vote. That’s not considering whether some subset of those participants registered, but didn’t cast their vote. And this brings me to my first substantive criticism of Catalyst: I do not think it does enough to enable and encourage participation, and in fact seems to actively discourage it in many ways.

This isn’t to say that this is intentional. To extend the popular maxim: “Never attribute to malice that which can instead be attributed to incompetence; and never attribute to incompetence that which instead can be attributed to systemic forces”. Those running Catalyst are taking actions to try to improve this, from better and more engaged marketing, to redesigning the voting flow, etc.

However, they are also making misteps that work against those efforts. Waiting a full year between Catalyst rounds means that those who participated in the past become disengaged and disenfranchised, and need to be recaptured. We’re 3 years into Catalyst, with 10 Catalyst rounds under our belt and we’re still using Ideascale; even though Ideascale was identified as a significant problem in very early Catalyst rounds. We’ve changed the key format for the on-chain signatures yet again, forcing people to re-register with confusing rules about whether or not they’ll be included in the snapshot or receive rewards. We force voting to happen through a proprietary mobile app, which locks huge swaths of the world out of the ecosystem. This proprietary app has significant and very easily solvable UX and stability issues. Particularly egregiously, I opened the app several days after casting my vote to find that, despite displaying “confirmed” previously, half of my votes were never finalized and I needed to resubmit them. I wonder how many people never bothered to open the app again, thinking their vote was cast.

I know that each one of these likely has a reasonable sounding explanation for why we’re in this state, or how it will get better in the future. I know all too well how it can feel to have someone armchair CEO from the outside. Such critics have very little perspective for exactly what contributed to where we are today. If someone from the Catalyst team is reading this, I really want to emphasize that I can have empathy, while at the same time having criticisms that I want to hold you accountable to.

If you are a member of the community who is also angered by these things, what I’d like to emphasize to you is that you can be angry about and demand movement on these, while also having empathy and compassion for those in the trenches trying to make it happen. And if possible, instead of shovelling shit into those trenches, ask for a shovel and get in the trenches with them.

My second substantive criticism of Catalyst is that we are not experimenting rapidly enough. The only way to solve incredibly hard problems is to find out and eliminate options that don’t work. This is called “failing fast”, and Catalyst has not set itself up to fail fast.

In particular, there was over a year between Fund 9 and Fund 10, and the only substantial meaningful change that I was able to find was changing the sort order of the proposals from community review stars to random. I reached out to Daniel via twitter DMs for other examples of meaningful changes made based on community feedback. As I haven’t received a response in the 4 days it took me to write this post, so there may be some I missed. But by and large, Fund 10 was operated very similarly to Fund 9.

This means that not only did we lose the opportunity to learn a new set of things that don’t work, we have repeated all of the same mistakes from Fund 9, and further disenfranchised members of the community. And, in some cases, with the added time and lower participation, these problems were significantly worse.

One incredibly difficult challenge the Catalyst team is facing is that so many people have ideas for how to change it, making it nearly impossible to sort through the noise and figure out which of these largely incompatible ideas is “the right one” to adopt. It could be that the Catalyst team is left with decision paralysis, with no way to cut through the noise, and worried that picking something unilaterally may anger more people than it pacifies.

My suggestion to the Catalyst team would be to find more ways to experiment: run funding rounds more frequently; run smaller intermediate rounds; run them with different rules. And then open the data in those rounds and be reflective on the feedback. I think this kind of experimentation would both allow us to iterate a lot better, and demonstrate your willingness to tackle this problem in good faith.

To the broader Cardano Community, however, this only works if the Catalyst team has some grace to experiment with different approaches, where criticisms and suggestions are delivered from one collaborator to another, not angry tirades and bitter catechisms.

To that end, one of the promising aspects of Voltaire is that it opens up a broader channel for this experimentation as well. Having a single Catalyst team organizing funding within Cardano means that it is a lot more cohesive and in theory the learning from each round can be better integrated in future rounds; This comes at the cost of putting an incredible amount of strain on a single team. A broader, more diverse set of funding channels that each have their own goals and philosophies about how to approach the problem will be noisier, but also allow us to explore the problem space a lot faster. And because of that increased noise, we as a community need to become a lot more sophisticated in how we digest and deliver feedback from these kinds of experiments, and avoid just instantly labeling everything that didn’t have the outcome we want as “broken” or “corrupt”.

Many of you may know that I have argued strongly against any kind of quadratic voting or “one person one vote” systems in Voltaire. My main reason for being against these is that nobody has yet shown me a concrete proposal and analysis for how they can be secured, when the integrity of our entire protocol and treasury are at risk. But, the dynamics for a single funding round through Catalyst are much different, in my mind, where the upper bound on potential damage is well known. So, contrary to the position I took when discussing CIP-1694, I very strongly believe that Catalyst is the perfect avenue to explore alternative voting schemes. Lets try quadratic voting, or time-weighted voting, potentially with smaller and more regular funding rounds, to get real world measurements of how aggressively these systems are exploited compared to the size of the fund.

The third substantive criticism I want to raise is related to the ability to downvote. I believe the intention behind the feature is good: provide the community a mechanism to signal and filter out active scams. However, it is clear that this mechanism is not working as intended. I have worked very hard, as have many others, to foster a “rising-tide, all-boats” culture of competition within the Cardano ecosystem. I strongly believe that our corner of the internet and the goals we’re trying to accomplish are better helped in the long term by a strong sense of collaboration and support between projects that would classically be competitors. And the downvote feature, when usable by projects themselves, works directly counter to this, and amplifies, rather than smooths, the influence of whale wallets.

The scope of this blog post is probably too small to contain a full analysis of this; Instead, I’ll defer to the other excellent slueths (such as Xerberus ❤️) to enumerate this case. Nor do I have a concrete solution, beyond a number of half-baked ideas to explore. I would love to see a deeper analysis of a number of approaches, and meaningful commitment from the Catalyst team to follow the data where it leads. I also understand that some of these proposals may not be technically feasible, depending on the zero knowledge structure of how votes are cast.

For example:

• Downvotes are removed all together, and scams are instead identified during community review.
• Downvotes dilute across proposals, while upvotes do not.
• Downvotes proportionately lower all upvotes from a voter.
• Upvotes are ADA weighted, but downvotes are delegation time weighted.
• Whether a proposal is funded is decided entirely by upvotes; but downvotes can scale down the reward amount along a log curve.
• etc.

I’ll close out this article with a reflection on a mission statement for myself I arrived at a while back:

I am building in crypto to contribute to a “nuclear disarmament treaty” with existing power structures, outsourcing what we can to technology for transparency, robustness, and efficiency.

I am building on Cardano specifically because it proports to have a mission of self-sovereignty and sustainability that I find lacking in other chains.

I am building DeFi protocols specifically to build financial tools and passive incomes for those who usually lack them, as a means to help distribute power to the edges.

While I believe Catalyst is attempting and succeeding in many ways that are aligned with my goals, that isn’t mutually exclusive with having substantive and meaningful criticisms and calling for the Catalyst team to do better. If we abandon nuance in favor of outrage or supplication, we will never accomplish great things.