The developer stories enterprise IT leaders actually read.
Volume 341, Issue 3 · 10 articles from 2 issues
Cloud Native Now compares Docker Hub and private registries, arguing public sources still have a validated role.
Author Sean Roth examines the security tradeoffs between Docker Hub and private container registries. Docker Hub has become the default starting point for developers and organizations racing to accelerate software delivery.
A private registry provides control over image security, provenance and retention that a public source cannot.
The article argues that public sources still have a legitimate role within a secure, validated supply chain rather than being eliminated. It covers how to combine both without losing the control a private registry provides.
Frontier models have exhausted quality public code, so synthetic training data recycles the same flaws.
AI-generated code has plateaued on security because frontier models have consumed most high-quality public code and now train on synthetic output that recycles existing weaknesses.
The recurring defects are familiar ones: missing validation, unsafe queries, and incomplete access controls.
The structural problem is that the most secure enterprise code is private and never enters training data, so models never learn from the practices that actually hold up in production.
Meanwhile AI generates vulnerabilities at a higher rate than human developers while security remains unoptimized in model design. The recommendation is to treat AI-generated code as untrusted input requiring independent verification, the same way a third-party component would be treated.
AI is shifting frontend engineering focus from routine coding to designing intelligent user experiences.
AI-powered development tools are accelerating routine frontend tasks such as code completion, testing, and bug detection, freeing engineers to concentrate on higher-value creative work.
Modern applications are becoming increasingly adaptive, using machine learning to personalize content and predict user behavior.
AI cannot replace sound engineering principles, thoughtful architecture, and the human judgment needed for maintainable systems. Frontend engineers must now collaborate across disciplines and weigh privacy, transparency, and ethical implementation.
Success increasingly depends on understanding architecture, user behavior, and AI-assisted workflows rather than memorizing frameworks.
Six persistent myths about Agile delivery leaders obscure a role centred on facilitation and coaching.
NTT DATA's Jim Dorney argues that Agile has been so overwhelmed by buzzwords that many organizations believing they are Agile have not actually moved away from command-and-control structures.
He rebuts six myths: that Agile leaders are administrators, that they are scrum police, that the role must be technical, that certification equals capability, that they are cheerleaders, and that they must remove every blocker themselves.
In reality their value lies in facilitation, coaching, protecting focus, and representing teams at governance layers when a problem is too big or political for the team. High-performing teams depend on trust, accountability, and continuous learning, and the mark of a great Agile leader is a team that runs well without them.
The constraint has moved from writing code to validating it, making test infrastructure critical.
AI-assisted coding now produces software faster than enterprises can validate it, inverting the traditional development bottleneck from generation to review.
Companies such as CodeRabbit and Blacksmith have attracted hundreds of millions in funding to attack the validation problem, a sign that testing infrastructure has become critical infrastructure.
The article warns that tests must be anchored to business requirements rather than AI-generated assumptions, or teams end up testing whether the AI agrees with itself.
Continuous testing integrated throughout development, rather than applied as a final gate, is presented as the practical requirement for keeping quality up as code volume rises.
AI-generated code is write-only and disposable, making tests the documentation and automation the reviewer.
Phillip Mortimer argues that AI makes all code write-only because it is too dense to read comprehensively. Tests define behavior and serve as the real documentation.
Code becomes disposable, easier to rewrite than to debug. Human code review does not scale, so automated review and self-healing agents listening to observability platforms become essential.
Developers should focus on creativity and intent rather than implementation detail.
AI coding agents now handle complete development tasks while developers focus on direction and review.
In 2026, AI coding agents have evolved beyond simple suggestions to manage full software workflows including planning, coding, testing and review.
Research shows 90% of professional developers use AI agents weekly and 68% use them daily. Key capabilities include handling complete tasks across multiple files, running parallel agents, understanding entire codebases, self-testing and creating custom agents.
However, only 29% of developers trust AI output for accuracy, making human oversight and verification critical.
CSO Online examines a widening asymmetry between AI's offensive vulnerability discovery and its defensive code quality.
CSO Online examines a growing cyber asymmetry as AI models' offensive and defensive capability gaps diverge.
In recent months LLMs have moved from flooding open-source projects and bug bounty programs with questionable security reports that wasted developer time, to routinely finding zero-day flaws that humans and traditional audit tools had missed for years.
That rapid evolution alarms even the models' own creators. Despite the advances in vulnerability discovery, the same models still cannot reliably write secure code. The article argues this leaves enterprises facing attackers who benefit more from AI than defenders currently do.
InfoQ argues AI-generated code is too dense to read, so tests define behavior and rewriting beats maintaining.
InfoQ examines an argument that AI makes essentially all code write-only. The claim is that AI-generated code becomes too dense for humans to read productively, so reading it stops being the primary way developers understand a system.
Tests instead define the behavior and function as the documentation of record. The second half of the argument is that code becomes disposable: when a change is needed, regenerating the component becomes easier than modifying it.
The piece examines what that shift means for engineering culture and methods.
A practitioner guide to what AI code review is, how it differs from generation, and how to evaluate tools for production.
This HackerNoon piece from Qodo addresses the review problem created when AI writes a large share of the code. It defines what AI code review is and distinguishes it clearly from AI code generation, which are frequently conflated.
It maps where review fits in the software development lifecycle rather than treating it as a bolt-on step. The piece then covers how to evaluate AI review tools for production use. The framing is that review capacity, not generation capacity, is now the constraint.
More from Developer · All IT section newsletters · This week