Crypto’s Potentially Fatal Interdependency Risk
Exploring Composability and ‘Exposure to the Weakest Link’
If you enjoy the article, please consider smashing that follow button on the right. For more, check out my Twitter, LinkedIn and Personal Website. Now, let’s get into it.
Composability — the idea that protocols can utilise and build on the products developed by other protocols — is a core feature of the cryptocurrency ecosystem
For those building new solutions, it allows software to ‘plug in’ to existing solutions.
This substantially speeds up product development and reduces wasted time spent on technological replication.
In contrast, traditional capitalism has incentivised monetisation of private information.
Companies are default-noncomposable.
Facebook leverages its own proprietary technology and data to generate revenue for the business.
It is a closed system, which contributes to its competitive advantage and business moat.
By generating larger data sets than Google, for example, Facebook has a greater value proposition to potential advertisers on the platform.
Facebook goes to these advertisers and says, ‘hey, we can give you the most targeted product and increase the conversion ratios on your ad-spend’.
Crypto, on the other hand, values openness.
Crypto protocols are (mostly) default-composable, building modular code for new protocols to plug and play.
Social Acceptance of Failure
This reflects a broader shift in capitalism towards transparency and openness.
Modern social media and telecommunication enables real-time distribution of content. It promotes sharing the good, bad, and the ugly while building a business.
The common saying is to ‘build in public’.
Management teams are producing regular social media content to connect with users and humanise executives.
This humanisation has demystified company-building.
When social media really started growing, it became valued as a platform to show your ‘best’ self in every post.
As time has gone on, we’ve acknowledged that nobody is perfect, and begun to attribute positive economic value to founders and teams that share successes and failures.
Companies have realised that consumer capital (sales, loyalty, engagement) increases when the business, and founding team, is humanised.
Sharing successes and failures enables other companies to learn, and avoid making the same mistakes.
One Step Further
But back to composability.
The crypto ecosystem has taken the transparency phenomenon one step further by promoting open-source development and composability.
This is predicated on a belief that consumer capital is maximised when it provides direct value to consumers and businesses that use the platform (refer to poorly drawn diagram below).
This general trend is extremely encouraging, but it also introduces new challenges, particularly in the realm of security.
In traditional technology firms, security is outward-facing.
Let me explain this using a medieval analogy.
Companies set up their IT infrastructure as a castle which repels and protects the platform from external attackers.
The castle’s management build walls and fortify the position.
Composable systems, on the other hand, have visiting hours and ‘visa-free’ migration.
Anybody can come into the castle and build new systems e.g. a bathroom which leverages the castle’s water system.
This new relationship creates interdependency which disincentivises malicious behaviour from either party because consumers require both systems simultaneously.
If the castle shuts off water, nobody can use bathroom. If the bathroom provider infects water supply with sewage, everyone in the castle suffers.
Classic game theory, but with a value-accretive outcome.
The Weakest Link
This system works well most of the time.
In this article, I’m hoping to highlight a fundamental issue which doesn’t receive enough attention: modular products are reliant on the security and trustworthiness of the weakest link.
The image above shows a typical product stack where each layer represents a different community or product.
Layer 3 relies on both Layer 2 and the Base layer. If either of these layers fails, every layer above also fails.
Let’s use a crypto example:
- An NFT collection (L3) wants to issue an ERC20 token using Uniswap (‘L2’) as a liquidity provider , which fundamentally settles on the Ethereum network (‘Base’).
- If Ethereum stops validating transactions, Uniswap can’t settle token swaps, and the NFT collection loses price visibility on their NFTs (which has knock on effects as holders begin to panic and sell).
- If Uniswap fails, the NFT collection loses price visibility on their token, but Ethereum continues to operate.
- In either scenario, the NFT collection is reliant upon layers beneath them in the chain.
The following image from Multicoin Capital research illustrates the complex interdependency of crypto systems well. This has nothing to do with the example above.
It simply shows that these systems are even more complicated in practice.
Any failure of an individual dependency creates ripple effects which impacts the stability of all others.
Centralisation Risk
Centralisation risk is the biggest challenge to default-composable systems.
As you move further up the stack, the higher the likelihood of centralisation in some form.
New protocols lack the scale to effectively decentralise operations, so require a management team to bootstrap development and growth.
The further you go up the stack, the greater the aggregate centralisation risk (i.e. the sum of all ‘centralisation risk’ in layers beneath).
The counterargument is simply that decentralisation of control will fix the problem long-term.
While true, it is difficult to attain true decentralisation the higher up the stack you go (i.e. layer 4, 5, 6).
This is because your total addressable user base decreases by an order of magnitude at each higher layer.
Higher layer solutions have a more niche and specialised use case. The challenge is inescapable.
Modularity and composability is not just found in the tech stack, but also in the broader ecosystem.
A second order effect of the FTX fallout is that several organisations which received grants are now forced to shut down.
This is a form of ‘funding composability’ where layers of the stack often fund higher layers to expand the reach of their product.
From the example above, it’s like Uniswap creating a fund to assist protocols building on Uniswap. Even if Uniswap’s systems are running perfectly fine, mismanagement of the Uniswap DAO or fraud may lead to loss of those funds. This then leaves the L3 without funding.
Smart Contract Risk
Assuming fully decentralised systems at each layer, smart contract risk is also an issue.
Each layer requires reliance on smart contracts and code developed by the layers beneath it.
Any bugs, hacks, or breaches threaten the stability of higher layers.
Intra-Industry Funding
Above I mentioned that funding can evaporate when organisations are no longer able to give open-ended grants.
The situation is even more complex when it comes to lending — a contractual relationship where repayment of funds is expected.
Lending within the ecosystem opens a whole new can of worms.
It amplifies the already-present challenges of composability. Most recently, we’ve seen the contagion effects of FTX’s collapse.
FTX was the recipient of loans from several platforms, including BlockFi, which recently filed for bankruptcy.
Allegedly, Alameda (FTX’s trading arm) defaulted on a $680m loan to BlockFi, causing them to go under.
Intra-industry lending amplifies interdependence.
In some sense, intra-industry funding is incentivised with default-composable protocols.
Lower-layer protocols will fund higher level platforms to incentivise development.
This is why several layer 1 protocols have foundations (e.g. Ethereum Foundation, Solana Foundation, Cardano C Fund).
Funding for growth is a central component of TradFi, and Crypto isn’t changing that.
Intra-industry lending is a challenge which can’t fully be solved by decentralisation, but rather prudent balance sheet and risk management.
Wrap up
Composability is a feature which will supercharge crypto development and unlock immense value to consumers and developers in the ecosystem.
Unfortunately, it also introduces challenges which re-emphasise the importance of decentralisation and thorough code auditing in protocols operating at L1 and L2.
Modular systems are only as secure as the most vulnerable component.
This will pose more of a challenge as we begin to build L3, 4, and 5.
A building is only as strong as its foundation. But for the first time, a fire on Level 3 can bring the whole building down.
