"Can you say anything about incentives in Algorand?"
That question was directed to Silvio Micali, an MIT professor who had just delivered a keynote on his theoretical proof-of-stake (PoS) system at the Financial Cryptography and Data Security conference in Malta, yesterday. And the Turing-award winner's answer set a few back on their heels.
"Incentives are the hardest thing to do," Micali said.
In 30 years as a cryptographer, he had spent the last 10 working on just that issue.
As he explained, when you put incentives out there, people learn how to use those incentives for making money in ways that are nearly impossible to predict. He pointed to bitcoin as a prime example, saying its creator probably never imagined bitcoin's incentive structure would lead to industrial-scale mining pools.
Micali also argued those using a system do not necessarily need to be rewarded for trivial computations. And that, while bitcoin miners are compensated for their work, validators, who in contrast do not have to invest in expensive equipment and electricity, are not rewarded.
He said:
Intended as a public blockchain, Algorand contains a novel version of Byzantine agreement with nine steps, where players are replaced in each round of communication. The protocol tolerates one-third bad actors, and Micali said he assumes the majority of the system's users are honest.
Against the grain
Yet, the idea of a consensus algorithm that offers no incentives runs counter to the thinking of many, including those working on the decentralized application network ethereum, a blockchain project working on a PoS system of its own called Casper.
"We basically explicitly put incentives front and center," said Vitalik Buterin, founder of ethereum. He described ethereum's approach as fundamentally about how much money stakeholders can lose, as opposed to the approach taken with Algorand.
"One thing I would be concerned about, is if you have no incentives at all, then that means you have no incentive not to just be lazy and go offline," Buterin told CoinDesk.
Ethereum developer Vlad Zamfir, who is heavily invested in building Casper, had stronger words. He stated he simply did not think such a system would work.
"My whole perspective on the space is, like, polar opposite. I don't believe the 'majority of people are honest' assumption," he said, adding:
Cornell associate professor Emin Gün Sirer, also questioned the idea Micali put forth.
Sirer pointed out that, while it's true bitcoin's participants are not always fully incentivized, a lot of people run full nodes altruistically.
"But to go from that and to say, since bitcoin works and it is not fully incentivized, ergo, any system will work – that is not incentivized. There is a gap there," he said.
Launch and learn?
But some had an altogether different take on the matter.
Charles Hoskinson, CEO of blockchain technology firm IOHK, pointed to BitTorrent as an example of a system that works just fine while not being incentivized at all. No token exists and you aren't paid anything to share files on the network.
Another example he points to is folding@home, where thousands of people freely donate idle processing power for disease research.
"It is an open question of whether you need incentives or not, and I don’t think it can be determined in an academic model. It is actually going to be determined by evidence. You launch something and you see what happens," he said.
The comments nonetheless suggest that Algorand's eventual launch could be one to watch, especially for those who question whether the platform's incentive plan will work or not.
Correction: This article has been revised to reflect Emin Gün Sirer's title.
Disclaimer: CoinDesk received a subsidy to attend the Financial Cryptography and Data Security conference from the event's organizers.
Image via Amy Castor for CoinDesk