Marginfi co-founder Edgar Pavlovsky departed suddenly from the Solana DeFi project in April.
The split wasn’t exactly amicable. “I don’t agree with the way things have been done internally or externally,” Pavlovsky wrote on X at the time.
Now, eight months later, Pavlovsky is a core contributor to Paladin, a Solana client meant to protect validators from so-called sandwich attacks while helping them “earn more in block rewards.”
Around the same time, Temporal, a crypto research shop with multiple marginfi-linked members, began advertising “Nozomi,” a proprietary quic client partially meant to prevent sandwich attacks.
Temporal is legally distinct from mrgn and marginfi. Temporal said most of its staff never worked at marginfi.
A Temporal researcher I spoke to was critical of the project being worked on by his former colleague. The two products aren’t the same, by both parties’ accounts, but they are both focused on the same niche. There’s one notable difference though — Pavlovsky’s Paladin has a token, while marginfi, controversially, has yet to debut a native coin.
Temporal’s Nozomi infrastructure is built to land transactions as fast as possible. It also prevents sandwich attacks by only sending transactions to “whitelisted” Solana validators who are known not to attack users. The project pitches this as a way to democratize trading.
“We have a ton of experience from our trading arm getting the best execution for what we want to do,” Jakob Povsic, a developer at Temporal, said. Temporal is “getting this level of sophistication for basically anyone.”
Paladin, on the other hand, is a fork of the Jito-Solana client that identifies and drops sandwich attacks from transaction bundles and helps Solana validators prioritize transactions with high priority fees attached, Pavlovsky told me. Access to Paladin validators will eventually be gated with the yet-to-launch PAL token.
I asked Pavlovsky if building a product with similar anti-sandwiching aims to his ex-colleagues at Temporal makes him want to compete more aggressively.
“No, man, they’re like my children,” Pavlovsky said with a chuckle. “I’m stoked for them to win, stoked for them to build.” He said blockchain infrastructure creators tend to have too much of a winner-take-all mentality when, in fact, multiple businesses can succeed at the same time.
Povsic showed less bonhomie. “I applaud them for going after sandwiching since [it’s] a big problem, but their approach is naive and very [self-serving] from what we are seeing,” he said of Paladin. Povsic added Paladin “arbitrarily” blocks some transactions based on a wide net of criteria, and sophisticated players will still be able to pull off sandwiches. And “forcing” its own token into the equation made matters worse for Povsic.
Pavlovsky responded to the criticism by saying that its anti-sandwich whitelist is “obviously systematic, not arbitrary.”
“[P]aladin is designed with sophisticated players in mind (and being tested comprehensively against sandwichers on the whole network). [W]e’re always happy to discuss tradeoffs,” Pavlovsky said in a text.
“[We] think [Paladin] is extractive and would rather focus on open-source initiatives rather than sending a % of priority fees to [Bloxroute], who own 20% of the token supply. Nozomi is quite different in that it’s simply a product to help land transactions fast,” Temporal partner and mrgn research co-founder Ben Coverston said.
The PAL token is an interesting part of the equation. Marginfi frequently draws criticism for its so-called points program that tracks user contribution to the platform. Points are frequently used to allocate token airdrops, but marginfi is one of the last major Solana DeFi protocols not to release a token. Pavlovsky never mentioned the lack of a marginfi token in his departure post in April 2024.
I asked him if frustration at the platform not launching a token contributed to his leaving. “[I] am pro token and think the right time to ship one was [Q4] 2023,” Pavlovsky said in response.