Vitalik Buterin, who helped start Ethereum, talks about his ideas for layer-3 protocols.

Although Ethereum-based layer-2 solutions have designed to increase the speed of scaling, Ethereum co-founder Vitalik Buterin believes that layer 3s serve a different purpose offering “customized functionality.”

Buterin expressed his thoughts on the topic in a Saturday blog post, providing three “visions” of what layer 3s might be utilized in the near future.

The Ethereum co-founders said a 3rd layer on the blockchain would make sense only if it serves distinct functions from layer 2s. Layer 2s are used mostly to improve scaling using zero-knowledge (zk) rollup technology.

“A three-layer scaling structure that is built by stacking the same scaling strategy on top of itself doesn’t work very well. Rollups on top of rolls in which the 2 layers utilize the same technology, definitely aren’t.”

But, “a three-layer architecture where the second layer and third layer have different purposes, however, can work,” Buterin said.

Related: Post-Merge Ethereum Hits 2-Month Low, Bitcoin Back Around $18K

One of layer 3’s usage instances is the type of application Buterin refers to as “customized functionality” — refers to privacy-related applications that use zk proofs to send transactions that are privacy-preserving to layer 2.

Another possible use-case is “customized scaling” for specialized applications that don’t wish to utilize Ethereum Virtual Machine (EVM) for computation. Ethereum Virtual Machine (EVM) to do computation.

Buterin added that layer 3 can be used to provide “weakly-trusted” scaling through Validiums ZK-proof technology. Buterin believes that this is useful to “enterprise blockchain” applications by making use of “a centralized server that runs a validium prover and regularly commits hashes to chain.”

But However, Buterin said that it’s not clear if layer-3 structures are superior to the existing layer-2 model in regards to developing custom applications on Ethereum.

“One possible argument for the three-layer model over the two-layer model is: a three-layer model allows an entire sub-ecosystem to exist within a single rollup, which allows cross-domain operations within that ecosystem to happen very cheaply, without needing to go through the expensive layer 1,” Buterin explained.

But Buterin stated that since cross-chain transactions can be easily and inexpensively executed in two layers who have joined an identical chain the addition of layers 3 may not enhance the effectiveness of the network.

Buterin’s remarks on possible use cases for layer 3 are in the context of StarkWare’s recently released validity proofs that recurse could have ended Ethereum’s scalability worries.

Declan Fox, the product manager of Ethereum software company ConsenSys, recently told Asian Trade that “with recursive rollups and proofs, we theoretically can infinitely scale.”

Related: This October, Uniglo (GLO), Ethereum (ETH), and Aave (AAVE) could hit ATH.

Recursive proofs like these have been thoroughly tested in real-world production and have been proven to work, with StarkWare co-founder Eli Ben Sasson saying to Asian Trade that the company’s proofs recursive have rolled up around 600,000 token mints that are non-fungible in a single transaction Immutable X and that 60 million transactions may soon be in the pipeline “with more engineering and tweaking.”