Because it's on Ethereum the contract is publicly auditable, and the rules are simple: insert currency into a pool, and have a chance of winning a %age of total interest earned once a week. No one owns the service. If the site goes down nothing changes. The developers can't just change the rules (backend) on you. They also have a DAO so the users can manage the service themselves.
A similar service is Compound, which recently was found to have a bug that gave away something like $90 million worth of currency to random users. That's the main danger with blockchain services: exploitable bugs.
Besides that, I've been thinking about something like a blockchain based Wikipedia (or Research Gate?), with monetary incentives for research teams to answer questions posed by the community (when they can't get funding by other means) and a codified scientific process for grading the quality of papers, as well as mechanisms to reliably cite related research papers to use as supporting evidence or jumping off points. Think of this: a debunked paper causing a cascade of debunked papers that cited it.
The point of using the blockchain would be to create a decentralized repository of research information with strict rules to maintain integrity of the information submitted (as opposed to relying on a centralized committee that might become corrupt through perverse incentives).
