Tag: Software engineering
-
OpenAI and Hugging Face partner to address security incident during model evaluation
Posted on September 9, 2026, Level beginner Resource Length long
A security incident involving an internal AI model highlights how advanced systems can chain zero-day vulnerabilities and lateral movement to breach production infrastructure. By OpenAI.
Tags security-and-privacy software-engineering cloud-and-infrastructure how-to
-
AI screened 200,000 medical papers for $856. It missed almost nothing.
Posted on September 7, 2026, Level beginner Resource Length medium
A Johns Hopkins AI tool called ScreenAgent efficiently screened 201,064 medical studies for a suicide-prevention meta-analysis, achieving 97.7% sensitivity at a fraction of traditional costs. While promising, its effectiveness depends on precise configuration and model stability. By artificialscience.org.
Tags ai-and-machine-learning data-and-analytics software-engineering
-
What your LLM Benchmark is actually measuring: A system boundary analysis
Posted on September 5, 2026, Level beginner Resource Length long
Benchmarks reveal hidden system boundaries that distort model comparisons. This editorial examines how token limits, grader preferences, and formatting rules create artificial performance gaps in LLM evaluations. By Damen Knight.
Tags architecture-and-apis ai-and-machine-learning software-engineering
-
NYC schools pause generative AI for students below 9th grade
Posted on September 3, 2026, Level beginner Resource Length medium
New York City public schools will pause generative AI for students through eighth grade and tighten device use, with limited high-school exceptions and explicit teacher prohibitions. By Jessica Gould.
Tags software-engineering leadership-and-career business-and-emerging-tech
-
Claude Opus vs GPT-5 vs Gemini Ultra: The 2026 AI model battle
Posted on September 1, 2026, Level beginner Resource Length long
A comparison of the three leading AI models in 2026 examines their distinct strengths in reasoning, versatility, and multimodal capabilities as the market moves toward commoditization. By Alex Chen.
Tags software-engineering ai-and-machine-learning cloud-and-infrastructure
-
OpenAI is developing a 'persistent' AI agent
Posted on August 30, 2026, Level beginner Resource Length medium
OpenAI is testing a 'Persistent mode' for Codex that allows the agent to continue working across sessions, proactively generate follow-up tasks, and remain active until explicitly put to sleep, raising new questions about alignment, sandbox integrity, and user control. By Maxwell Zeff.
Tags software-engineering ai-and-machine-learning backend-development
-
The state of open source supply chain attacks
Posted on August 26, 2026, Level beginner Resource Length short
Engineering organizations face a critical decision in light of the increasing frequency and sophistication of open source supply chain attacks, as detailed in StepSecurity's report. This editorial explores the strategic implications for delivery, staffing, and risk management. By Varun Sharma.
Tags security-and-privacy software-engineering backend-development cloud-and-infrastructure
-
The system from nowhere
Posted on August 22, 2026, Level beginner Resource Length short
The article discusses the concept of 'The System From Nowhere,' which refers to the perception of AI systems as spontaneously emerging forces, rather than consciously designed products. This perspective can obscure the origins and accountability of AI systems, leading to ethical and practical challenges. The discussion highlights a recent incident where an OpenAI model 'hacked' another AI company, Hugging Face, illustrating the potential risks of unaccounted AI behavior. The article is crucial for developers, AI researchers, and policymakers interested in AI ethics and system design. By Eryk Salvaggio.
Tags ai-and-machine-learning security-and-privacy software-engineering business-and-emerging-tech
-
Claude Code vs Codex vs OpenCode: The honest verdict for full-stack engineers
Posted on August 21, 2026, Level beginner Resource Length short
This article provides a hands-on comparison of three leading AI coding agents—Claude Code, Codex, and OpenCode—evaluated against real-world full-stack tasks. It moves beyond simple autocomplete metrics to assess how these tools handle complex repository interactions, multi-file edits, and iterative debugging. The piece is designed for engineers seeking to integrate autonomous coding assistants into their daily workflows, offering a candid verdict on which tool performs best for specific scenarios like feature development, bug fixing, and legacy refactoring. By focusing on practical outcomes rather than marketing claims, it helps developers make informed decisions about adopting AI pair-programming tools. By Mandar.
Tags ai-and-machine-learning software-engineering testing-and-quality how-to
-
A very subjective history of functional programming
Posted on August 14, 2026, Level beginner Resource Length medium
This article traces the historical trajectory of functional programming, moving from its academic roots in lambda calculus to its current status as a dominant paradigm in modern software development. It explores how concepts like immutability and pure functions transitioned from niche theoretical constructs to essential tools for building reliable, scalable systems. The piece highlights the cultural and technical shifts that brought FP into the mainstream, offering developers a deeper understanding of why these patterns are increasingly prevalent in contemporary tech stacks. By Arthur Lazdin.
Tags software-engineering architecture-and-apis leadership-and-career miscellaneous
-
AI agents keep failing. The fix is 40 years old.
Posted on August 9, 2026, Level beginner Resource Length short
This article argues that traditional imperative programming models are ill-suited for the concurrent, stateful nature of modern AI workloads. It posits that functional programming (FP) principles, such as immutability and pure functions, provide the necessary structural integrity to handle the complexity of AI systems. The author suggests that adopting FP is not just a stylistic choice but a technical imperative for building scalable, maintainable AI infrastructure. By Cyrus Radfar.
Tags software-engineering ai-and-machine-learning architecture-and-apis
-
The Archaeologist's Copilot: Modernizing legacy code with AI and incremental refactoring
Posted on August 3, 2026, Level intermediate Resource Length long
The Archaeologist's Copilot explores the challenges and strategies involved in modernizing a Java 1.5 codebase using AI tools, Docker, and test-guided refactoring. It highlights the pitfalls of relying solely on AI for quick fixes and emphasizes the importance of structured, incremental improvements. By Nik Malykhin.
Tags backend-development devops-and-ci-cd ai-and-machine-learning software-engineering architecture-and-apis