Martin Fowler
PulseAugur coverage of Martin Fowler — every cluster mentioning Martin Fowler across labs, papers, and developer communities, ranked by signal.
7 day(s) with sentiment data
-
LLM evaluations shift from subjective checks to structured metrics
An article discusses the importance of structured evaluations (evals) for assessing the performance of large language models (LLMs), moving beyond subjective judgments. The author details a project where a deterministic…
-
Claude Fable 5.1 deciphers ancient cipher; AI regulation debated · 7 sources tracked
Claude Fable 5.1 has demonstrated its capability by deciphering a 370-year-old cipher, the Cyphral Distich, in under an hour using approximately 176,000 tokens. This feat highlights the model's advanced autonomous reaso…
-
Consort framework enables test-driven development on live database branches
A new agent framework called Consort has been introduced, designed for test-driven development on live database branches. This framework aims to enforce engineering discipline by allowing agents to run tests and control…
-
Martin Fowler questions TDD's value in AI agent loops
Martin Fowler explores the practical application of test-driven development (TDD) within an agent loop, questioning whether it offers genuine value or is merely performative. He examines how TDD principles might be inte…
-
Thoughtworks engineers build airline IROps system in 4 days using emergent agent coordination
Thoughtworks technologists conducted a four-day experiment in Barcelona, Spain, involving 10 engineers to explore "hyper-agentic" software development. The team successfully built a simulated airline IROps system, a com…
-
Harness Engineering: Guiding AI Coding Agents for Reliable Software Development
Harness engineering is a concept for effectively integrating AI coding agents into software development workflows. It involves creating a surrounding system, or "harness," that guides the AI with project-specific rules …
-
Martin Fowler: Data preparation is key for agentic AI
Martin Fowler argues that for AI agents to be effective, the data they interact with must be meticulously prepared. He emphasizes that current data practices are often insufficient for the demands of agentic AI, which r…
-
Martin Fowler details "The Orchestrator's Tax" in AI systems
Martin Fowler's "The Orchestrator's Tax" article explores the hidden costs associated with using AI orchestrators. These systems, while powerful, introduce complexities and overhead that can outweigh their benefits if n…
-
Martin Fowler advocates for AI tools that augment, not replace, human input
Martin Fowler discusses the concept of "harness engineering" in software development, emphasizing that AI tools should augment rather than replace human input. He suggests that these tools should guide developers toward…
-
Understanding the 'Agentic Harness': The Infrastructure Behind AI Agents
This post explores the concept of an "agentic harness," which is the surrounding infrastructure that enables a language model to function as an agent. While the model provides the intelligence, the harness offers reliab…
-
New IRC-A Architecture Enforces Agent Boundaries via Infrastructure
A new decentralized architecture pattern called IRC-A (Internet Relay Chat for Agents) is proposed as a solution to the open problems in agentic engineering, such as trust, boundaries, and control. The architecture emph…
-
AI's impact on software development: Refactoring economics and agent isolation
Two distinct articles discuss the implications of generative AI for software development workflows. One article, by Martin Fowler, explores the economic advantages of refactoring code in the context of generative AI. Th…
-
Martin Fowler explores AI code refactoring economics
Martin Fowler's article explores the economic advantages of refactoring code generated by AI agents. He details an experiment where a large, 17,000-line Rust file, part of a 150,000-line codebase primarily written by AI…
-
Martin Fowler explores local AI models for coding; MDN launches new server
Martin Fowler's latest article explores the feasibility of using local AI models for coding tasks, contrasting them with cloud-based solutions. The piece delves into factors influencing the viability of local models, su…
-
Hacker News highlights Apertus open foundation model and reliable agentic AI
Hacker News featured two prominent articles: one on Apertus, an open foundation model for sovereign AI, and another on building reliable agentic AI systems. The Apertus model aims to provide a foundation for AI developm…
-
Developer's prompt tames Claude's over-engineering reflex
A developer has found a specific instruction to be highly effective in guiding Claude's coding behavior, preventing it from over-engineering solutions. By adding the line "Default to the smallest change that fits. Add a…
-
AI text impact on web studied; VibeSec security approach questioned
A new study explores the pervasive influence of AI-generated text across the internet, examining its impact on web content and metrics. Concurrently, a separate analysis delves into the concept of 'VibeSec,' a security …
-
Video reframes code's role in the age of LLM assistants
A new video discusses how Large Language Models (LLMs) are changing the nature of code. The video reframes code's role, suggesting that as LLMs commoditize the machine-readable aspect, the human-readable, conceptual mod…
-
VS Code shifts focus from AI models to agent harness engineering
The VS Code team's recent documentation on their Copilot agent harness reframes the focus from solely improving AI models to enhancing the surrounding infrastructure. Their internal benchmark, VSC-Bench, revealed that i…
-
Spec-driven development gains traction with AI coding tools
Spec-driven development (SDD) is emerging as a more structured approach to software creation, contrasting with traditional "vibe coding" methods. This methodology emphasizes defining a complete specification upfront, wh…