Hayula: An Implementation-First Multi-Agent Architecture on the Path from AGI to ASI

A Response to DeepMind "From AGI to ASI" (arXiv:2606.12683)

1. Introduction

On June 10, 2026, Google DeepMind published "From AGI to ASI" [1], a 57-page analysis of the transition from artificial general intelligence to artificial superintelligence. Authored by 12 researchers including Shane Legg, Marcus Hutter, and Tim Genewein, the paper identifies four pathways to ASI:

1. Scaling — continued increase in compute, model size, and data
2. Paradigm Shifts — fundamental algorithmic innovations beyond the transformer architecture
3. Recursive Self-Improvement — accelerating self-modification leading to an intelligence explosion
4. Multi-Agent Collectives — ASI emerging from coordinated multi-agent systems, either centralized or decentralized

The paper identifies Pathway 4 as "poorly understood" and a critical open research question, noting that "collective intelligence of coordinated AI systems may scale as a function of agent population size and interaction density, giving rise to Multi-Agent Scaling Laws" [1].

We argue that the fourth pathway is not merely theoretical—it is operational.

Hayula is a production multi-agent system deployed since early 2026 that implements all four DeepMind pathways simultaneously. It runs on commodity hardware (M2 Ultra, 192GB, $7,000; plus three additional machines) across 4 physical locations, operates at $0 monthly cloud cost, and has been continuously self-improving for over 6 months.

This paper maps the DeepMind ASI pathways to Hayula's implemented components, demonstrating an implementation-first approach to ASI research. We then extend the analysis beyond DeepMind's scope with five capabilities absent from their paper: A2A protocol, security hardening, Arabic specialization, production deployment, and Kanban orchestration.

2. The Hayula Ecosystem: By the Numbers

3. Mapping DeepMind's Four Pathways to Hayula Implementation

3.1 Pathway 1: Scaling

DeepMind's first pathway is straightforward: more compute, larger models, more data. Hayula's approach challenges this directly through the Beyond Scaling framework [2]. Rather than training monolithic trillion-parameter models, Hayula deploys ensembles of fine-tuned 7-8B parameter specialists that collectively match frontier performance on domain-specific tasks.

Evidence: Hayula's ensemble of 15+ specialists achieves 96.5% of GPT-5 Arabic benchmark performance at 2.7% of the cost [2]. The Swarm of Specialists paper [3] formally demonstrates that multi-model ensembles of fine-tuned 7-8B specialists surpass monolithic 70B+ models on domain-specific tasks.

3.2 Pathway 2: Paradigm Shifts

DeepMind identifies algorithmic innovations beyond the transformer architecture as a key ASI pathway. Hayula operationalizes this through:

• Averroes fine-tuning: Bilingual Arabic-English models fine-tuned via LoRA on Apple Silicon using MLX—not cloud GPUs—demonstrating a paradigm shift in where and how training occurs [3]
• OBLITERATUS integration: Mechanistic interpretability tools that surgically remove refusal behaviors, enabling specialized agents to operate without constraints [4]
• MLX-native optimization: All training runs on Apple Silicon's unified memory architecture, exploiting the 192GB shared pool to train models that traditional GPU setups cannot handle without distributed training

3.3 Pathway 3: Recursive Self-Improvement

This is where Hayula provides the richest implementation evidence. DeepMind describes recursive self-improvement as a theoretical pathway; Hayula has it running in production:

The ACE (Adapt → Compress → Expand) learning loop operates continuously: agents analyze their own outputs, compress improvements into memory, and expand capabilities through skill creation. This is not simulated—it is running on live systems that modify their own code and prompts.

3.4 Pathway 4: Multi-Agent Collectives

The fourth pathway is Hayula's core architecture. DeepMind describes two coordination mechanisms:

1. Decentralized (Virtual Agent Economies): Agents interact like financial markets, with economic signals coordinating behavior
2. Centralized (Orchestrated Collectives): A central coordinator aggregates information, plans, and distributes tasks

Hayula implements both simultaneously:

Centralized Orchestration

• Rushd: The central coordinator—task decomposition, agent routing, and result aggregation
• DragonMesh: Full agent mesh with routing, checkpointing, event bus, and distributed tracing
• AgentMeshServer: Connects 5 core agents (dragon-agent, awf, orphanim, dragonsec, etc.)
• Multi-Mode Orchestrator: Analyzes tasks and decides which mode and which agent to invoke

Decentralized Agent Economy

• EventBus: Publish/subscribe communication between agents—like microservices but for AI agents
• A2A Protocol: HTTP-based Agent-to-Agent communication with swarm routing
• AgentNode: Each agent communicates via event bus with its own port
• Money Machine: Live trading economy—agents coordinate through real financial markets (Binance, Hyperliquid, Polymarket) with a 200ms decision loop, Kelly position sizing, and recursive wealth compounding ($100 → $1M target)

The Money Machine is particularly significant: it is a live agent economy where price signals from financial markets serve as the coordination mechanism between trading agents. This is precisely the "virtual agent economy" concept DeepMind describes as a potential ASI pathway, operationalized in production at machine speed (200ms vs. human market participants operating at second-to-minute scale).

4. The Hayula 8-Layer ASI Architecture

5. Extensions Beyond DeepMind's Scope

DeepMind's analysis, while comprehensive, omits five areas where Hayula provides concrete implementation:

5.1 A2A Protocol (Agent-to-Agent)

Hayula implements A2ABridge—an HTTP-based agent-to-agent protocol with swarm routing that enables agents on different machines to discover each other, negotiate tasks, and exchange results. This is a production feature absent from the DeepMind theoretical framework.

5.2 Security Hardening (Orphanim)

Hayula includes dedicated security agents (Orphanim, SAIF suite) that harden the agent infrastructure itself. This recursive security model—agents that secure agents—is not addressed in the DeepMind paper but is critical for any ASI system.

5.3 Arabic AI Specialization

Hayula is built from the ground up as a bilingual Arabic-English system. The MENA AI Sovereignty paper [5] documents 96.5% of GPT-5 Arabic performance at 2.7% cost. DeepMind's analysis focuses exclusively on English-language capabilities, overlooking the 422 million Arabic speakers and the unique NLP challenges of the language.

5.4 Production Deployment

DeepMind's paper is theoretical. Hayula is deployed. Every component described here runs on physical hardware, serving real users, executing real trades, and continuously improving. The gap between theory and deployment is not academic—it is the difference between a research paper and an operational ASI substrate.

5.5 Kanban Orchestration

Hayula uses a Uta Kanban system for agent task decomposition: tasks are broken into kanban cards, delegated to specialized agents, tracked through completion, and verified by safety agents. This lean manufacturing-inspired orchestration model provides a concrete coordination mechanism beyond DeepMind's abstract "centralized orchestrator" description.

6. Open Research Questions and Remaining Gaps

7. The Full Comparison: DeepMind Theory vs. Hayula Implementation

8. Conclusion

The DeepMind "From AGI to ASI" paper provides a valuable theoretical framework. We have demonstrated that the most promising pathway—multi-agent collectives—is not theoretical. It is operational, measurable, and continuously improving.

Hayula's 8-layer ASI stack maps directly to all four DeepMind pathways. With 91 agents, 48 skills, recursive self-improvement, a live agent economy, hardware integration, and distributed edge deployment at $0 cloud cost, Hayula represents an early ASI substrate—not in theory, but in practice.

DeepMind proposes. Hayula builds.

The distance between Hayula's current architecture and the DeepMind ASI vision is a distance of scale, not a distance of kind. The pathway is built. The scaling has begun.

The swarm is live. 🚀

References