In this day and age we're accustomed to communicating across vast global networks at the speed of light. Modern internet speeds allow us to send messages to our friends on the other side of the world in a second and to stream Netflix shows in realtime at 4k resolution.

In light of that, DeFi at present feels like using dial-up internet with interactions that are expensive (on the order of dollars) and that take on the order of seconds (if not minutes) to complete. It's highly improbably that in the coming years we can expect mass adoption of DeFi if things remain status quo.

Current Day DeFi

DeFi offers revolutionary, global access to financial services in a decentralized, censorship-resistant way. However, the user experience of decentralized applications (dApps) can be notoriously clunky, leading to sub-par products that derive most of their adoption from (usually unustainable) token emissions. Therefore, DeFi will continue to see poor stickiness and retention until the products themselves are truly functional. Key UX issues to solve include:

1. High product complexity: dApps often require sophistication and face blockchain-related friction, such as weak onramps, self-custody, and gas fees, making them challenging for less tech-savvy users.

2. Slow performance: dApps rely on blockchain back-ends, resulting in slower reads and writes and a sluggish experience that really make you acutely realise like you're interfacing with a blockchain (not a good thing).

3. Inferior execution: Poor liquidity, slow execution, gas fees, and vulnerability to MEV bots can discourage users from trading on decentralized exchanges (DEXs).

This blog focuses on improving issue #2, aiming to create a new class of DeFi apps that feel like regular web applications. The future of DeFi likely involves back-end abstraction, where users won't know or care if their transactions occur on Ethereum, Solana, or Arbitrum, much like consumers don't consider the infrastructure supporting web2 applications (AWS vs GCP).

A Typical Experience

The trading flow on the most popular perp DEX on Arbitrum highlights DeFi's shortcomings:

1. Transaction signing adds friction, providing unhelpful information and potentially causing missed opportunities in fast-paced markets. Solutions like smart contract wallets and session keys may alleviate this issue.

2. Even on a Layer 2 solution like Arbitrum, the trading experience is slow. Trades take ~3.2 seconds to execute and indexing takes over 10.2 additional seconds, an end to end experience of 13.4 seconds. This is noticeably slower than centralized applications like Binance, which offer snappy, sub-second experiences.

An Era of Performant DeFi

New blockchains like Solana enable developers to build faster, more scalable DeFi applications. Solana's high transaction throughput and 400ms block times make on-chain orderbooks possible. Zeta, built on Solana, aims to deliver centralized-exchange-like performance while maintaining decentralization and non-custodianship.

Despite inherent physical limitations (blockchains require consensus and light takes 130ms to travel around the world), DeFi can still achieve latencies within hundreds of milliseconds. Although not on par with high-frequency trading (on the order or microseconds), is likely sufficient for most crypto consumers. By offering latencies similar to popular mass consumer applications (e.g. Instagram or Robinhood), DeFi can provide a relatively seamless user experience.

Trade execution takes roughly 1 second on Zeta, and trades are indexed, parsed and displayed in the front-end in the 2 seconds following that.