Programming languages, the infinite garden, and great storytelling. I am Gabriel, researcher and engineer.

Welcome to my page.

Purpose

Writing is a formalized form of thinking, which makes this page an exercise in sensemaking. Here, I write my thoughts — mostly about technology. This page is also an attempt to document and share ideas that might benefit the broader industry.

About me

I design adversary-resilient algorithms and implement them end-to-end, from formal threat models to production smart contracts and off-chain systems.

Two examples. I co-authored the Dave algorithm, a dispute-resolution primitive with a proven exponential cost asymmetry against Sybil attacks. Dave came after our earlier primitive proved too slow to settle under sustained attack. And I designed, built, and deployed Cartesi's app-specific sequencer: low-latency, high-throughput rollup infrastructure that preserves proof-system compatibility and L1 forced inclusion.

In both examples, the systems were raised rather than architected top-down. I learned and practiced this discipline during my Master's in programming languages under Professor Roberto Ierusalimschy, creator of Lua, a programming language raised the same way.

The full record (papers, talks, and the rest) is in my CV.

Selected writing

My research and code are public and open-source. A few pieces of writing I'd point to:

  • Fraud Proofs Are Broken, but we can fix them (2024) — Established the safety/promptness/decentralization trilemma for permissionless fraud proofs. Drew responses from Ed Felten and OP Labs.
  • Best of Both Worlds? (2025) — Formal analysis of non-interactive ZK fraud proofs under adversarial conditions; drew responses from RISC Zero/Boundless and L2BEAT.
  • Enforceable Human-Readable Transactions (2025) — Proposed cryptographically binding human-readable descriptions to transactions, countering front-end spoofing like the $1.5B Bybit hack. Engaged Micah Zoltu and the Ethereum security community.
  • Scaling Content (2022) — Argued for RISC-V execution environments to give Ethereum access to decades of existing software, anticipating the ecosystem's current renewed interest in RISC-V execution.

Links