Tech and case studies on organizational management, team management, and engineering efficiency
The article explains Codeium’s two new metrics—Characters per Opportunity (CPO) and Percentage of Code Written (PCW)—how they are calculated, why they better capture the true value of AI‑powered code completion compared to traditional benchmarks, and how they guide product decisions.
Generative AI tools such as GitHub Copilot, ChatGPT, and Amazon CodeWhisperer are rapidly reshaping software engineering, with surveys showing over 90% of U.S. developers using them, and Seattle tech leaders reporting doubled productivity, faster code migration, and new workflows across startups and enterprises.
The article introduces DIDACT, a large‑scale multitask machine‑learning framework that trains on the full software‑development workflow—including edits, builds, reviews, and tool interactions—to create AI assistants that can predict and suggest developer actions throughout the coding process.
A recent Nature paper by MIT and collaborators argues that language serves mainly for communication rather than thought, presenting neuroscientific evidence that reasoning can occur without linguistic processing and sparking debate about the implications for large language models and artificial intelligence.
After more than a year of using LangChain for LLM‑driven applications, we found its heavy abstractions and inflexibility hindered production development, leading us to abandon it in 2024 and reconsider whether a dedicated AI framework is truly necessary.
This article explains how Meta applies AI, specifically a fine‑tuned Llama2 model, to improve AIOps by automating incident monitoring, providing real‑time summaries, assisting responders with contextual information, and efficiently narrowing down root‑cause changes, ultimately reducing incident resolution time from hours to minutes.
The article describes how Google has integrated AI, particularly large language models, into its internal software development tools to improve developer productivity, outlines the challenges faced, shares lessons learned, and outlines future directions for AI‑driven engineering assistance.
The article presents a comprehensive framework for evaluating engineering productivity by separating efficiency (speed and quality) from effectiveness (usability), reviewing academic literature, interviewing Google engineers, and proposing four interrelated types of software quality—process, code, system, and product—along with practical measurement challenges and recommendations for technical leaders.
The article explains how Google tackles the costly problem of dead code in its massive monorepo by using the Sesenmann automated deletion system, which leverages build‑system dependency graphs, activity signals, and graph‑analysis algorithms to safely identify and remove unused C++ code while addressing cultural resistance among engineers.
The article reviews typical code‑coverage goals such as 70‑80% for system testing, explains how factors like failure cost, resource constraints, testability, and development stage influence target selection, discusses why full coverage is often impractical, and summarizes major industry standards and coverage metrics used in safety‑critical domains.
