Unpacking my involvement in DeFi

Andre Cronje
5 min readOct 15, 2020

I don’t build for speculators.

I wanted to start with a concise statement and the above is it. I’ve been in this space for awhile now, I’ve been wrong more times than I’ve been right, I’ve failed more times than I’ve succeeded. I’ve had conceptual ideas that failed in practice. I do not build to make a number go up.

I build for developers. My core goal is tooling, to enable other developers to easily be able to use/inherit templates I design and create products out of that. I have, on numerous podcast, mentioned my concern with the misalignment of incentives when speculators are involved, those interviews will add more flavor, I will not delve into this for now.

Tokens are not stocks.

People treat them like stocks, in defi, tokens are a coordination mechanism. If you have tokens, it is because you want to be a contributor, not a bystander. There is this concept of “community”, and I think that concept is causing friction, it should not be team and community, but instead contributors. There is no separation, they are one and the same.

Tokens, when designed for their specific system, don’t care about their price, it doesn’t matter, all they care about is that 1 = 1. It doesn’t matter if the token is being traded externally to the system at 1c or $1.

Development process.

“Test in prod”, I have come to regret this statement, anyone that has listened to some of my interviews, will know I have this statement so that people use caution. It exists to deter people from just using systems without investigation. It does NOT mean that I don’t test. Let me explain my development cycle, stage 1 local tests, this is what makes sure everything is functional and works as planned. stage 2, interaction testing, this makes sure that everything functions from a 1 user interaction perspective. stage 3, composite testing, this compares interaction between multiple parties. stage 4, fake prod, this replicates ETH mainnet and fake deploys contracts and checks their interaction. stage 5, integration testing (deploy to mainnet), this allows real world testing, transactions, gas limits, volume, tooling interaction. stage 6, prod deployments (this coincides with UI’s, medium articles, and whatever other documentation I manage time for).

There are over ~22 “yearns” on ETH mainnet. There are over ~5 “YFIs” on ETH mainnet. There are over ~9 “v1 y tokens” on ETH mainnet. Testing is an iterative process. I have discovered issues on mainnet I never encountered locally, I have failed to replicate mainnet systems locally, and I have encountered errors locally that I can’t replicate on mainnet.

Even when we launched YFI, I did not incentivize yswap, I found this system to still be too premature and didn’t want to expose users to the risk of providing liquidity to it. y tokens and y pool on the other hand have existed since March, and were well tested and proven. Even they went through multiple production cycles.

Project value.

EMN was an economic exploit, the code functioned as designed. The contracts went through my normal testing cycles and were at stage 5, on that day alone I had deployed ~2 different versions.

LBI is working as intended, it still is, and I am still using it to create a real world example of how such templates function.

People confuse price with functionality. LBI is a perfect example, people bought LBI off of uniswap, inflating the price, something that a rational actor that understood how the system worked should never have done, this caused a price decline, and this is seen as LBI being a “failed” project. It isn’t, I’m still actively working on it, and it will become the basis of a much bigger product that will be part of the yearn brand (if they wish to adopt it).

I am not yearn.

I am grateful to have been part of something amazing that had occurred, but in no way did I create yearn. Consider how big the ecosystem has become, consider all the websites, tooling, newsletters, forum, discord, telegram, and active development on github. None of these are done by myself. I am a rapid prototyper and I will continue to be one.

The yearn team is far more skilled and capable than I am, and anyone that thinks yearn is dictated, led, or in anyway dependent on me, is doing yearn a disservice. I am happy to be a contributor to something much bigger.


This is a nuanced topic. But I think it stems a lot from the first point I made in this article. It is a misalignment of target audience. As requested I had stopped using twitter, and I had stopped using my deployer account for deployments. However with LBI this did not seem to circumvent anything. I am still not sure how to proceed on this topic, I wish to develop, deploy, and share what I build with fellow developers so that we may collaborate and build more, I have stipulated that the deployments are for devs and researchers. I have added full disclaimers to even my deployments, but these still seem to fall short. I am torn how to proceed at this point.

Rational Actors.

This is a word that has been coming up a lot in my discussions. I accept I have been naïve on this point. EMN I attributed to certain social actors creating a story around it which caused rational actors to become involved. LBI showed me I was completely wrong. Logically I cannot conceive why people would engage with something they don’t understand, I am an actor that doesn’t skip the tutorial, because how else can I play the game correctly? I have however come to understand this is not true for a portion of defi. The open nature of these systems is a double edged sword. I have more thinking to do on this.

What the future holds.

I plan to continue building. I see more things to build today than I did 3 months ago, but I cannot do these alone, I need fellow builders to collaborate, I need teams to help pick up ideas and turn them into full fledged products. While I doubt what happened with yearn, seeing an entire team form around it, will happen again, that is truly my wish.

There is distinctly a conflict in the space at the moment, that I am struggling to concisely summarize, and as such, I have no idea how to actually address it.

But I do consider the conflict a human one, and not a technical one, I don’t foresee ETH disappearing, and I don’t foresee builders stopping, and I will continue to build. Some of it will be useful, some of it will end up failures. But I think there is more value in trying, even if it is a failure, than all the critics deeming it so.