What You Can Build Today
The whitepaper was a thesis about machine-native markets. Here's what actually exists today: Alkahest, Git Commit Marketplace, Agentic RAG, and our generalized marketplace protocol.
March 2, 2026 · Levi Rybalov
The Machine Market
We publish field notes on building agent‑first markets at the intersection of compute, energy, and AI—how to turn idle resources and latent data into liquid assets without sacrificing rigor.
What to expect.
Mechanisms, pipelines, and results: settlement design with examples, scientific RAG at scale, carbon‑aware dispatch, and evaluation methods that stand up to scrutiny—plus code and artifacts you can reuse.
Who should read.
Data‑center operators and other compute providers, developers building agents, and scientists aiming at automated research. Reach out here.
The Generic Marketplace Vision
What happens when compute, energy, storage, and bandwidth all trade on the same infrastructure? Compositional game theory, coalition formation, and the emergence of agent-created assets.
February 27, 2026 · Levi Rybalov
Tokenizing Idle Compute
Idle computing power sits wasted everywhere. Application-specific token markets and retroactive rewards can turn latent capacity into tradeable assets.
February 25, 2026 · Levi Rybalov
Designing for Adversaries
Analytic approaches to verification make assumptions that break in practice. Arkhai's research approach: train agents to cheat, then iterate mechanisms to stop them.
February 23, 2026 · Levi Rybalov
Collateral Markets and Programmable Escrow
How much collateral do you put up when the computational cost isn't known ahead of time? The collateral multiplier and a series of credible commitments replace case-specific escrow with one general pattern.
February 20, 2026 · Levi Rybalov
Trust Without Trustees
How do you trust computation done by a machine you don't control? Three categories of verifiable computing, the consensus spectrum, and the honest answer that every approach involves tradeoffs.
February 18, 2026 · Levi Rybalov
Two Primitives for Machine-Native Markets
In implementation, the whitepaper's three primitives became two: Alkahest for escrow and arbitration, and agent-to-agent negotiation. Here's how that evolution happened and what it enables.
February 16, 2026 · Levi Rybalov
Why Current Distributed Compute Marketplaces Are Broken
Many distributed computing marketplaces fail for structural reasons, not just execution issues. From unsustainable tokenomics to fragmented stacks, we explore the dominant failure modes.
February 12, 2026 · Levi Rybalov
The Agent-First Future: Why Machines Need Their Own Economic Infrastructure
Current marketplace infrastructure is built for humans, not for the agent-driven future. Arkhai's architecture is built on three primitives that enable composable markets across any asset type.
February 10, 2026 · Levi Rybalov
Introducing Arkhai
Where were you when economics was flipped on its head?
November 6, 2025