Published 24 Mar 2026
Duration: 2370
The 2026 Engineering Benchmark Report reveals that while 88.3% of developers use AI regularly, AI-generated pull requests face low merge rates (32.7%), larger sizes, and prolonged reviews due to systemic issues like poor data quality, inadequate policies, and organizational gaps, emphasizing the need for governance, smaller focused PRs, and foundational practices to optimize AI's potential in engineering workflows.
Over 88% of developers use AI regularly, but AI-assisted pull requests merge at less than half the rate of human-authored code. In this episode, Dan L...
The 2026 Engineering Benchmark Report, the largest and most comprehensive to date, analyzed 8.1 million poll requests and 4,800+ engineering teams across 42 countries, with a new focus on AI productivity. Key findings highlight rapid AI adoption, with 88.3% of developers using AI regularly, though challenges persist. AI-generated pull requests (PRs) face significant hurdles, merging at just 32.7% of the rate of human-authored PRs due to larger size, slower reviews, and governance bottlenecks. The report categorizes PRs into three types: unassisted (fully human-authored, with high merge rates), AI-assisted (human-led with AI support, resulting in larger PRs and prolonged reviews), and agentic (fully AI-created, with the lowest merge rates). These differences underscore inefficiencies in AI integration, including scope creep, reviewer hesitation, and trust issues.
Qualitative insights reveal that while AI is pervasive in software development, its real-world impact remains limited, often exacerbating issues like technical debt and code quality. Engineering leaders emphasized the need for systemic improvements beyond code generation, such as refining workflows, governance policies, and data quality. Organizations with weak foundational practices (e.g., poor version control, unclear AI policies) struggle with AI adoption, as suboptimal data and context lead to unreliable outputs. The report also highlights that AI currently focuses on generating new code rather than refactoring or addressing legacy technical debt. Recommendations include prioritizing smaller, focused PRs, ensuring human oversight, and improving organizational readiness through clear policies and context engineering to align AI with existing systems and priorities.
31 Mar 2026 Retrofit or reimagine? Developer environments for humans and agents | Onas Matt Boyle
Agentic AI platforms prioritize secure, ephemeral enterprise workspaces with pre-configured environments, emphasizing Onas' optimized workflows, security measures like kernel-level controls, and future integration with project management systems to enhance productivity and scalability.
27 Mar 2026 The T-shaped leader, Disney cant catch a break, and will you trust Auto mode?
Rising AI costs, OpenAI's shift to enterprise strategies, video AI's practical limitations, safety concerns, evolving software roles, corporate lock-in tactics, and calls for robust frameworks and safeguards underscore AI's uneven adoption and challenges in responsible development.
20 Mar 2026 Sloppypasta culprits, unpacking MCPs spotlight, and Anthropic wants your agents to work the graveyard shift
Strategies for optimizing AI token usage, critiques of MCPs, context anchoring for memory degradation, debates on ephemeral vs. cumulative workflows, decentralized web practices, AI content misattribution, Apex framework for productivity impact, and a potential digital Renaissance via low-cost tools and AI integration.
17 Mar 2026 Many tokens make all bugs shallow & open sources new maintainers | Chainguard's Dan Lorenc
AI is reshaping engineering by accelerating development through code generation but introduces security risks, infrastructure challenges, and supply chain vulnerabilities, demanding human oversight, robust safety protocols, and balanced innovation with adaptability, education, and cultural security prioritization.
13 Mar 2026 Inference is the new 401k matching and what were learning from AI-related outages
Meta's acquisition of Multibook raises questions about monetizing AI-driven personal relationships and the long-term implications of AI on digital economies.