Tag
Vincent Weisser announces the release of over 365,000 open and agentic reinforcement learning environments for software engineering, terminal, and search agents.
A reflective essay on the lack of accountability in software engineering and LLM development, drawing from a keynote at ICST 2024 that called for responsibility akin to other engineering fields.
In the age of AI-assisted coding, the author argues that deep understanding of programming language concepts remains essential for engineering, listing key ideas from various languages that expand mental models for problem solving.
Explores the concept of technical debt specifically arising from the deployment and maintenance of AI agents, suggesting new challenges for software engineering.
The article argues against blindly adopting LLMs and provides six questions to evaluate whether an LLM is appropriate for a given workflow, emphasizing that LLMs trade determinism for flexibility and should only be used when necessary.
The author shares two years of experience building a platform with AI, identifying six recurring failure modes (Band-Aid, Assumption, Drift, Hallucination, Lack of Common Sense, Path of Least Resistance) and argues that even as models improve, these failure modes persist, becoming harder to detect.
NVIDIA Object-Oriented Agents (NOOA) proposes a Python framework where agents are Python objects, with methods as actions and docstrings as prompts, enabling reliable, testable AI agents using familiar programming paradigms.
Gergely Orosz praises Sirupsen for his deep understanding of networking and latency basics, referencing an interview where Sirupsen used napkin math to build a 10x cheaper or faster product.
A blog post argues that Claude (an AI model) is not a compiler but is better, as it works across abstraction layers like humans do, contrasting with the traditional layered view of software development.
AI tools like Codex are breaking traditional proxies for measuring engineering productivity, forcing leaders to distinguish between activity metrics and actual business value.
Cognition announces the acquisition of TierZero, a team focused on incident detection and system health automation, to enhance their AI software engineer Devin.
This article explains the concept of finite state machines, their formal definition, and why they are a powerful abstraction for building reliable systems, including how they relate to current agent frameworks.
Kent C. Dodds shares Rasmic's insight that old engineering books can help steer AI agents that write code.
Explores the distinction between perfection and over-engineering in software development, arguing that over-engineering is solving the wrong problem while perfection emerges from clear requirements.
This essay examines how AI-generated code has broken the traditional link between the worker's credentials and their output, sparking debates over gatekeeping and quality. It proposes that new verification tools can rebundle contribution with a verifiable receipt, addressing both quality and access concerns.
A comprehensive resource providing guidelines and recommendations for writing dependable, safe C code, covering undefined behavior, memory model, and version-specific advice.
A reflective article questioning whether anyone has successfully implemented multi-agent AI workflow orchestration without resorting to makeshift solutions.
A programmer reflects on AI's rapid improvement in coding, noting that models like Claude now excel at senior-level tasks such as system architecture and refactoring, and predicts that AI will master programming within a few years, with current limitations being temporary harnessing issues.
Google discusses the challenges of monolithic prompts for AI agents and proposes a modular prompt transpilation approach using templates, includes, and build systems to improve maintainability and reliability.
This article examines how AI agents are automating the repetitive tasks that junior developers traditionally used to build taste and judgment, altering the career path from junior to senior and contributing to rising unemployment among recent computer science graduates.