The Prophecy of Layer 1 Chains
Ethereum L1 vs Alt-L1s (and what about side chains and ZK rollups)
Ethereum L1 vs Alt-L1s vs Side Chains vs ZK Rollups
Over the weekend we had a bit of a splash when Zhu Su tweeted the below:
Of course, this created a ton of controversy given the dig on the Ethereum community and the discounting of a lot of the hard work to scale, led by numerous teams in the ecosystem. However, regardless of the wording, there is an element of truth where Etheruem is pricing out a lot of users at scale. This tweet from Sisyphus was some anecdotal evidence as well about rising gas costs:
In case you haven't done the counting that's north of $100m+. I wrote an article almost a year ago when yield farming came out and it's worth reading to get a perspective on the state of the world at the time:
Back then the situation was the following:
A quick breakdown of the cost to even get started mining:
Depositing $USDC = $2-3
Opening leveraged position = $8-$10
Collecting harvest (claiming $COMP) = $5-10 per transaction
Selling $COMP for $USDC = $10-$15 in transaction fees
Unwinding leveraged stable coins position = $50
Summing up the numbers, it’s very reasonable to think I spent over $100 in transaction fees over the course of 2-3 days. Some of these transactions were definitely more expensive using InstaDapp since you have to deposit everything to a smart contract wallet and then pay for extra added overhead. If you do this without a contract wallet you can probably reduce a lot of these costs but gas prices are still going to average 30gwei - 50gwei.
Back then it seemed like it was a major problem but the easy solution given out was "don't be poor" aka have more than $10k which was doable given the insane money being made during DeFi Summer. However, when the price of being poor is less than $100m then I think that excuse doesn't scale.
So fast forward to a year and a half later where we see major attacks on the Ethereum community and chain, the alternative layer 1s are being shilled relentlessly as a solution. On face value, yes they are. But don't be fooled, there's a lot of bandaids being applied in this thinking.
But newer chain = newer tech = faster & cheaper?
No, not really. In one of the most legendary posts written on this topic is Vitalik's one called "The Scalability Trilemma": https://vitalik.ca/general/2021/04/07/sharding.html
The crux of the article comes down to the following graphic:
You can achieve two, but typically never three. So what a lot of chains have done is basically sacrifice Decentralisaton for scalability and security. Now on the surface, you may be thinking, who cares about decentralisation? I'm not doing anything illegal, I just want cheap and fast transactions. Of course, however I think it's important to note what decentralisation gives you:
The first two are actually what people don't realise they're trading-off when they also sacrifice decentralisation. And when things go south (which they always do), you end up with situations where the blockchain itself goes down or the chain becomes out of state with different parties. So now you can't access your money or your balance is reported differently on different sites. Not ideal at all.
So okay, how exactly does decentralisation get compromised? Well typically you have two vectors:
Increase the requirements to run a node so that you have more high powered machines
Use fewer nodes to achieve consensus faster and quicker
To use an analogy, it's like trying to reach a decision for a country by reducing the number of people who are allowed to vote and then making sure they have to be rich as well. You'll have less contention to reach your final decision and to communicate changes.
The future is nearer than we think
At the start of this article I talked about the state of Ethereum a year ago and how gas fees were kind of high but "just don't be poor" was the excuse or "wait for gas to come down" would be fine. It's interesting to see that as of today Avalanche is going through a similar phase.
I did some basic digging into SnowTrace (Etherscan but for Avalanche) and this Uniswap-esque trade already costs $15 (for a $40 swap): https://snowtrace.io/tx/0x2511aa05b1aeb7f51c53e0a31742a1f064c117eb063ca25daa4ea9077b85f9db
This isn't really a dig on Avalanche but rather what the future of most chains holds for them since the scalability trilemma hasn't been solved at a base layer. Some chains make more extreme trade-offs to the point where this point will take longer to reach, but make no mistake — as demand increases the same issues will arise.
So what's the answer?
My best guess for how the future scales out is through the three following vectors:
Computation is done through zero-knowledge roll-ups
The final hash is saved on decentralised layer 1s
Apps that grow to become large enough will create their own specialised chains (Axie Infinity style)
Of course each of the above statements are chunky enough to get their own article so I'll be writing about it in more detail. However I hope for now that this piece serves as a reference for how layer 1s are a game where the end is predictable unless something significantly better is done.