"Okay, ethereum's done, let's go back to ethereum classic."

The jokes were frequent during ethereum creator Vitalik Buterin's keynote at the project's Shanghai developer conference, Devcon2, on Monday. There, despite the complexity of the technical changes the project will face in the years ahead (and the controversies in recent past), Buterin conveyed a confidence and personality onstage that was unique among the day's presenters, even as he outlined some of the more nuanced proposals.

That scaling was the subject of his talk, entitled "The Mauve Revolution", is not a surprise given the increased relevancy of the issue across all blockchain networks. As more banks and enterprise firms seek to use blockchain systems, the scaling question calls to light the inefficiency of the nascent technology.

But the idea that ethereum may be performing below the expectations of this new audience was the frequent butt of Buterin's barbs, with slides featuring titles like "What sucks about ethereum?" and an analogy comparing the network to a "smartphone from 1999".

To begin, Buterin rifled through a number of issues he sees with the blockchain-based decentralized application platform today, including how few transactions it can process.

Still, he told the audience:

In the remainder of his presentation, Buterin discussed ways ethereum will be seeking to solve for scalability in the coming months and years. While he did not discuss any specification timeline or plan of execution, the talk felt charged with an overall sense of direction that seemed to resonate.

"[Buterin] has a very special ability in that he's able to come up with the theoretical solution to a situation that is very much pressing today. He can play devil's advocate to himself," Kesem Frank, COO of Toronto-based enterprise blockchain firm Nuco, told CoinDesk.

Lower on the agenda for the talk was discussion of ethereum's proposed effort to implement sharding, a concept that would find it splitting in such a way where multiple blockchains would, almost like miners, form their own consensus on a larger state.

Here, Buterin deferred attendees to the latest version of his "mauve paper" outlining his current thesis on the state of the network, the third edition of which was announced prior to the conference.

Key to ethereum’s vision for expanding its network of users is transitioning from the transaction validation algorithm popularized by bitcoin (proof-of-work) to an alternative (proof-of-stake) that does not require the purchasing of hardware.

Buterin explained the transition as one that will seek to replicate bitcoin's mining process virtually without "wasting electricity".

Essentially, Buterin sees his fix as one that will find consumers purchasing ethers (the unit of account on the protocol) in exchange for virtual miners, which would then be governed in such a way as to replicate a competitive verification process.

"Virtual miners are kept track of in the state of the protocol itself," Buterin explained.

However, Buterin's version of the idea offers a number of fixes to what he called the “supposed fundamental flaws” of this long-attempted validation mechanism.

First, he outlined that it is possible to make such a system more difficult to game if those who buy virtual miners have to wait to join the validation pool, thus gaining eligibility to the rewards produced by the protocol.

Buterin also envisions restrictions on both withdrawals and the transactions that these addresses can as execute as other ways to ensure that those who are validating are doing so in ways that will not be harmful to the computing network.

"If you're done mining, you can call this function called start withdrawal. Then after another, something like a few months, you can withdraw and take your ether out," he said.

Perhaps Buterin's most powerful critique, however, was to the "nothing at stake" problem whereby proof-of-stake algorithms have historically struggled to align virtual miners.

Key to solving this, he predicts, will be constructing proof-of-stake in a way that incentivizes participants to continue backing the “winning” version of the transaction history. A feature he proposed as a solution is the inclusion of so-called "dark uncles" or "dunkles" in the protocol.

"The fact that you made a block on another chain means you get penalized and it hurts you," he explained

A tongue-in-cheek play on the term "uncles" (which refers to blocks that are mined but not added to a winning blockchain), he foresees dunkles incurring steep penalties, to the point where losses would even be 1,000% larger than rewards.

Buterin sees the proof-of-stake protocol aiming to encourage the production of a network where the winning blockchain would be the one with the most "value at stake". In this way, he said, validators could bet on the blockchain they think will be the winner by continuing to back it with value.

"Bets start off being conservative, but then over time expand. As validators see that everyone is betting 10-1 on a particular block, 20-1, 40-1, the value lost on a particular block will expand exponentially," he said.

Despite the heavy emphasis on theory, however, Buterin was keen to state the goals of the effort in simplistic terms.

He concluded:

Image via Pete Rizzo for CoinDesk