Aktagon Signals AI-generated & human-reviewed
categories

Architecture

rikiyat.github.io 5 min read

Sparks of Weight-Space Programming: LLMs Can Natively Read & Write Neural Network Weights

Researchers at MIT CSAIL demonstrate that frontier LLMs like GPT-6 Astra can natively read and write neural network weights without gradient descent, identifying trained behaviors from raw weights and writing …

AI · Architecture Editorial Team
youtube.com 6 min read

Building Always-On Persistent AI Agents: Architecture, Context Engineering, and the Future of Autonomous Systems

A deep dive into the architecture and engineering principles behind always-on, persistent AI agents that can run autonomously for extended periods, covering context management, durable workflows, sub-agent orchestration, …

AI · Architecture · Development Editorial Team
arxiv.org 4 min read

Harness Learning: Training LLMs to Adapt Agent Programs at Test Time via Meta-Learning

This paper introduces harness learning, a method that trains a proposer model to revise executable agent programs using execution feedback, framed as meta-learning over code. Experiments show the approach generalizes to …

AI · Architecture Editorial Team
typesafe.ai 3 min read

Introducing System One Models & Jev: A New Frontier for AI-Powered Automation

TypeSafe AI introduces System One Models and Jev, a new class of frontier AI models optimized for fast, structured, type-safe decisions rather than text generation. Jev achieves comparable intelligence to existing LLMs …

AI · Architecture Editorial Team
arxiv.org 4 min read

Tensor Logic: A Unified Language for Neural and Symbolic AI

This paper proposes tensor logic, a programming language that unifies neural and symbolic AI through tensor equations based on the equivalence between logical rules and Einstein summation. The language elegantly …

AI · Architecture Editorial Team
arxiv.org 0 min read

The Log is the Agent: Event-Sourced Reactive Graphs for Auditable, Forkable Agentic Systems

ActiveGraph introduces an event-sourced runtime where an append-only log serves as the primary source of truth for AI agents, with graph state as a deterministic projection of that log. This architecture enables …

AI · Architecture Editorial Team
arxiv.org 3 min read

Zep: A Temporal Knowledge Graph Architecture for Agent Memory

Zep is a memory layer service for AI agents powered by Graphiti, a temporally-aware knowledge graph engine that dynamically integrates conversational and structured business data. It outperforms MemGPT on the Deep Memory …

AI · Architecture Editorial Team
arxiv.org 5 min read

Dive into Claude Code: The Design Space of Today's and Future AI Agent Systems

A comprehensive architectural analysis of Claude Code’s agentic coding tool, examining its design principles, permission systems, context management, and extensibility mechanisms through source code analysis. The …

AI · Development · Architecture Editorial Team
arxiv.org 3 min read

Scaling Latent Reasoning via Looped Language Models

This paper introduces Ouro, a family of Looped Language Models (LoopLM) that achieve 2-3× parameter efficiency by implementing iterative computation through shared parameters and adaptive depth allocation, demonstrating …

AI · Architecture Editorial Team
block.xyz 4 min read

From Hierarchy to Intelligence: How AI is Reshaping Organizational Design

Block is pioneering a fundamental shift from traditional hierarchical management to AI-driven organizational intelligence, replacing human coordination layers with world models and intelligence systems. This approach …

AI · Architecture Editorial Team