Topic
All digests tagged Software Engineering
Suggested route
Move from earlier context through the available material to the newest signals. This route is a reading aid, not a difficulty ranking.
Arize's agent Alex evolved from a 50% hallucination prototype to a robust system. Key lessons include moving planning into custom harness code, treating large tool results as files, and using small markdown files for long-term memory.
Read article
Code Mode is an alternative architecture for AI agents that uses a programming sandbox instead of a large tool catalog. This approach drastically reduces context bloat, lowers token usage, and enables emergent behavior by allowing agents to compose tools in code.
Read article
AI success is an ecosystem problem, not a model problem. The Full-Stack AI Engineer must orchestrate intelligence across data platforms, agents, and low-code tools, while prioritizing governance and data quality.
Read article
Learn about meta-tooling: building AI agents that write, load, and use their own tools at runtime. Discover how to implement self-healing, self-modifying agents using the Strands Agents SDK.
Read article
As AI coding systems become capable, writing code is the easy part. The real challenge is planning, coordinating, and validating massive changes across multiple repositories and pipelines.
Read article
A quarter of PRs reviewed by Greptile are now AI-generated. The talk compares agent code (Codex, Claude, Devin) to human code on bug rates and review cycles, finding unique failure modes for each agent.
Read article
AI agents are not the revolution; they are the catalyst. The real impact is accelerating foundational changes in data, systems, and connectivity, forcing standardization and interoperability.
Read article
Jev introduces System One Models, a new class of AI designed for deep software integration. Learn why reliability, structured state, and decomposing tasks are the future of programmatic AI.
Read article
Coding agents succeeded because their infrastructure was built for automation. Knowledge work lacks these primitives—Centralization, History, Context, Verification, Governance, and Reversibility—creating a massive gap in reliability.
Read article
AI is shifting the focus of software reviews from line-by-line implementation details to validating business outcomes, requirements, and overall intent. This marks a major evolution in development workflows.
Read article
Translating messy operational knowledge into executable agent workflows requires building a robust 'refining loop' around the AI agents. The system's true asset is the SOP corpus, which captures global process variations and drives systematic improvement.
Read article
The talk explores how LLMs impact software expertise. While adoption is high, researchers warn that over-reliance can degrade skills, impair critical thinking, and lead to 'cognitive surrender,' posing a risk to junior engineers.
Read article
Scaling agentic software engineering requires moving beyond simple tool calls to advanced primitives like Skills and Hooks, ensuring that knowledge is injected into the limited context window only when absolutely necessary.
Read article
Software engineering is shifting from deterministic instructions to defining goals for AI agents. Agentic Engineering emphasizes human oversight in managing probabilistic workflows, making the developer an orchestrator rather than a manual coder.
Read article
Diogo Almeida argues that current LLMs, trained via RLHF, optimize for human preference rather than reliable automation. The next frontier requires optimizing AI for verifiable rewards and autonomous decision-making.
Read article
To train autonomous AI agents, Emulated simulates entire companies in high-fidelity environments. This moves beyond simple code diffs to model complex failures across distributed clusters and cloud resources.
Read article
The industry is shifting focus from training bigger models to optimizing cost and speed. Key techniques include the 'sidekick' agent architecture, which uses parallel processing of multiple LLMs for better price performance.
Read article
Lucas Palma discusses how Nubank built Skill Vector to vet 2,000+ AI skills, treating them like supply chain dependencies in a regulated environment. The system uses hybrid scanning combining deterministic checks and LLM review.
Read article