This episode reveals why a former Firedancer core engineer believes Solana isn't performant enough, introducing a new leaderless blockchain architecture designed to solve the fundamental economic flaws of proof-of-stake.
The State of Firedancer and Solana Client Diversity
- The Challenge of a Moving Target: Building a new client requires constantly monitoring and adapting to changes in the primary codebase (Agave), making it difficult to get ahead without a complete specification. Liam notes, "it's really hard to build against someone else's specification. It's really hard."
- The Two Sides of Client Diversity: Liam presents a nuanced view on client diversity.
- The Social Benefit: The competition between Firedancer and Anza (Solana's core developers) has pushed both teams to improve, leading to significantly better testing and overall network quality.
- The Technical Risk: Unlike Ethereum, Solana's consensus algorithm can halt if two clients with significant stake disagree on the state of the chain due to a bug. This makes client diversity on Solana a technically riskier proposition compared to Ethereum, which has built-in mechanisms to continue operating in such scenarios.
Why Solana Isn't Performant Enough
- Liam frames the need for a new, more performant L1 not as a critique of Solana's success but as a response to the ever-increasing demand for on-chain products. He argues that just as the internet evolved beyond simple text messaging, blockchain infrastructure must scale to meet future demand. When the demand for blockspace outstrips the supply, it creates a serious problem that his new project, Through, aims to solve.
- Strategic Implication: Liam's perspective suggests that the current performance ceilings of even the fastest L1s like Solana will become a bottleneck. Investors should look for architectures designed not just for incremental improvements but for fundamentally different scaling models that can accommodate future, unpredicted demand.
Introducing the Through VM: A Focus on Developer Adoption
- The core of the new chain is the Through VM, which is built on RISC-V, an open-source instruction set architecture (a blueprint for how software communicates with hardware). Liam explains this choice was less about raw performance and more about expanding the developer pool.
- Broad Language Support: By using RISC-V, Through can support a wide array of existing programming languages like C, C++, Rust, and even newer ones like Zig and Nim. This strategy aims to onboard developers by meeting them where they are, rather than forcing them to learn a new, niche language.
- A Different Approach than Ethereum: Liam contrasts this with Vitalik Buterin's proposal to bring RISC-V to Ethereum. He argues that layering it on top of the existing EVM would not improve performance or developer experience due to the constraints of backward compatibility.
A Leaderless Consensus Model
- Through introduces a radical departure from the consensus mechanisms of Solana and Ethereum by being completely leaderless. Liam critiques single-leader models (like Solana's) and even multiple concurrent leader (MCL) models for creating monopolies or oligopolies over blockspace, where a small number of entities control supply.
- No Leaders, Open Block Production: In Through's model, anyone with a small amount of tokens can become a block producer. This separates the role of block production from validation.
- Aligning Profit with Performance: The core problem with traditional Proof-of-Stake, according to Liam, is that it "incentivizes people to accumulate capital" rather than focusing on network performance. Through's model realigns these incentives.
- Validators vote on blocks proposed by producers.
- Producers include a price (a bid) in their block to pay validators.
- Validators earn fees from these blocks and can convert those earnings directly into more "validator power" (voting power) on the network.
- Actionable Insight: This creates a competitive dynamic where validators who are most effective at quickly and efficiently processing valuable blocks earn more, allowing them to increase their influence. This direct link between performance and power is a key architectural difference for researchers to analyze, as it could create a more meritocratic and performant validator set.
Decentralized Block Production and Zero-Fee Transactions
- The separation of block producers and validators creates a new dynamic for the network. Block producers can range from sophisticated professional firms (like the block builders in Ethereum's ecosystem) to the applications themselves.
- Application-Level Sequencing: An application, like a DEX, can act as its own block producer. This allows it to capture its own order flow, eliminate MEV like sandwich attacks, and offer users extremely low or even zero-fee transactions by fronting the cost to the validator network.
- Solving the Spam Problem: Liam dismisses the idea that fees are the primary defense against spam. He argues that spam is a solved problem from a technical perspective.
- Validators and network infrastructure can efficiently filter out low-value or invalid transactions before they consume significant resources.
- He attributes much of the "spam" on Solana to bots trying to overcome the high "jitter" (unpredictability) of transaction landing spots. By creating a more predictable system, the incentive to spam is reduced.
Market-Driven Block Ordering and Variable Block Times
- Through determines block order via a market mechanism, moving away from the fixed block times seen on chains like Solana (400ms).
- Price-Based Ordering: Validators choose which blocks to include based on a combination of the price offered by the producer and the computational resources required. Producers can also set deadlines, creating a push-and-pull dynamic.
- Variable Block Times: This model is inspired by traditional financial markets, which operate in bursts of activity rather than at a fixed cadence. Liam states, "we transact...it's in bursts." Through can process many blocks during high-activity periods and remain quiet during lulls, better matching the supply of blockspace with real-world demand.
- Centralization Concerns: When asked if this would lead to centralization around a few sophisticated block producers, Liam points to traditional markets, where large players coexist with smaller, specialized firms that identify and exploit niche opportunities. He believes a similar power-law distribution will emerge, with applications themselves also acting as a decentralizing force in block production.
The Through Token and Value Accrual
- While details are still early, Liam confirms Through will have a native token. He addresses the "L2 problem" where applications might use their own token for fees, preventing value from accruing to the L1.
- Value Accrual Mechanism: Even if an application like "Through Swap" collects fees in its own token, it must still compete to get its blocks included on the L1. To do this, it will need to acquire Through's native token to pay the block producers' bid to the validators, ensuring value flows back to the core network's security and execution layer.
The Path to a 10x Performance Improvement
- Liam argues that a true 10x improvement over Solana won't come from incremental tweaks but from fundamental architectural shifts. The key lies in aligning incentives and removing artificial bottlenecks.
- The Most Overlooked Bottleneck: The Read Layer: Liam identifies the most "slept on" area for blockchain improvement as the read layer, or observability. He argues that without robust tools to measure and understand what is happening on-chain, it's impossible to optimize effectively.
- Strategic Insight: "How do you optimize without observability?" This highlights a critical, non-obvious bottleneck. For researchers and investors, this suggests that projects investing heavily in high-performance data infrastructure and analytics (the "read layer") may unlock significant performance gains that others miss.
Launching an L1 in 2025: Strategy and Timeline
- Liam acknowledges the difficulty of launching a new L1 in a competitive market. Unto's strategy focuses on deep collaboration with early ecosystem partners and a continuous, iterative launch process.
- Ecosystem Strategy: The team is bringing in EIRs (Entrepreneurs in Residence) with deep product experience and working closely in-person with the first few core projects to ensure the protocol meets their needs.
- Launch Philosophy: Instead of a "big bang" launch, the plan is to ship continuously. They have already built "Breakthrough," a functional L2 sequencer network, and are targeting a mainnet launch by July 2025. The process will be transparent, with continuous updates on SDKs, wallets, and partner integrations.
Conclusion
This episode details a new L1 architecture challenging leader-based consensus. Investors and researchers should monitor Through's market-driven block production and its potential to redefine network incentives and MEV capture. This model offers a new paradigm for achieving scalable performance and application-level sovereignty in blockchain design.
