TL;DR
Get travel and outdoor gear delivered free — and shop member deals
- Fast, free delivery on millions of items
- Access to Prime Big Deal Days deals on October 6–7
- Prime Video, Amazon Music and more included
A report based on a 2026 engineering leadership keynote describes AI as rapidly changing software development, with many engineers using multiple coding agents instead of writing code by hand. The account also flags concerns about code review, quality and reliability, while saying teams and planning remain important. It is a snapshot of selected companies and practitioners, not an industry-wide survey.
A 2026 snapshot presented to more than 2,000 engineering leaders describes AI tools reshaping software development, with some engineers coordinating five to 10 coding-agent sessions rather than writing every line themselves. The account, published by The Pragmatic Engineer after a keynote at the LDX3 conference in New York, also identifies concerns about code quality, reliability and reviews; it is a report of observed practices, not a representative survey of the whole tech industry.
The keynote drew on the author’s visits to OpenAI and Anthropic, conversations with startups and technology companies including Ramp and Uber, and unpublished data from GitHub, Factory AI and Linear. The source does not publish the underlying datasets or describe their methods, so those references provide reporting context rather than independently verifiable industry-wide measurements.
One prominent change is the move toward parallel work with coding agents. The report quotes Claude Code creator Boris Cherny describing five terminal sessions and five to 10 Claude sessions on the web running at once. Cockroach Labs co-founder Peter Mattis and Linear software engineer Dima Zaytsev also describe managing several agent sessions or worktrees concurrently. These accounts illustrate individual workflows; they do not establish how common the practice is across all engineers.
The report says that hand-written coding and the traditional integrated development environment are becoming less central for some practitioners. At the same time, it lists problems the shift has exposed: assumptions about generated code no longer hold, code reviews can become performative, and quality and reliability have declined. Those are the author’s reported observations, not quantified findings in the material provided. The account also argues that teams and planning still matter, and says non-engineers are not broadly shipping code on their own.
AI Changes the Engineering Workflow
The change matters because coding tools affect more than how quickly a developer types. If engineers delegate implementation to several agents, their work can shift toward defining tasks, managing parallel sessions, checking outputs and deciding whether software is safe to release. That raises practical questions for engineering teams, managers and companies about training, staffing, review processes and accountability.
The source’s concerns about reliability and code review point to a potential cost of adopting tools faster than teams can adapt their safeguards. More generated code does not by itself show that products are better, cheaper or delivered faster. The report offers no broad performance measurements, so the scale of any productivity gains or quality decline remains unknown.
As an affiliate, we earn on qualifying purchases.
From Coding Models to Agent Work
The report places the current shift alongside earlier changes such as the spread of the internet, smartphones and cloud computing. It argues that AI is arriving at a different scale and pace, while noting that software development practices will change without every part of engineering disappearing. Its author says coding models improved substantially near the end of 2025, helping make agent-assisted work more visible in 2026.
At The Pragmatic Summit, software engineer and author Martin Fowler compared AI’s impact with earlier changes in software. His remarks support the idea that the industry is facing a major shift, but they are an expert assessment, not a measurement of adoption. The report anticipates cloud-based coding agents and new AI infrastructure, while presenting these as developments to watch rather than settled outcomes.
“Nothing has hit with the magnitude of AI. This is a whole size difference from anything that we’ve faced before.”
— Martin Fowler, at The Pragmatic Summit
integrated development environment (IDE)
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
How Broad Is Agent Adoption?
The source does not establish what share of engineers use coding agents, how adoption varies by company or role, or whether the examples are typical. It also gives no published methodology or figures for the referenced unpublished data. The reported declines in quality and reliability are not accompanied by metrics, comparison periods or independent assessments, so their scale and causes are unclear.
It remains uncertain whether multi-agent workflows will become standard, how companies will change their review and testing practices, and whether claimed productivity benefits will persist once maintenance and defect costs are counted. The report’s predictions about cloud agents and AI infrastructure are forward-looking, not confirmed outcomes.
As an affiliate, we earn on qualifying purchases.
Tracking Tools and Quality Measures
The report expects cloud-based coding agents, supporting software systems known as harnesses, and infrastructure designed for AI workloads to develop further. These are the developments identified by its author, not announced milestones with specified release dates. The available source does not provide a schedule for additional data or follow-up findings.
For companies adopting these tools, the next useful evidence will be clear measures of delivery time, defects, maintenance burden and review effectiveness. Until broader data is available, the 2026 picture remains a reported snapshot of fast-changing practices among selected teams and practitioners.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is changing in software engineering in 2026?
The report describes engineers using coding agents to produce and modify code, often running several sessions in parallel. It says this may reduce the role of hand-writing code for some practitioners, but does not quantify industry-wide adoption.
Does the report show that most engineers have stopped writing code by hand?
No. It reports signs of that shift and gives examples from individual engineers, but provides no representative survey or adoption rate. Its examples should not be treated as proof of what most engineers do.
What risks does the report identify?
It flags concerns about generated-code assumptions, code reviews, quality and reliability. The source does not provide figures to measure the scale of those issues or establish that they apply across the industry.
What does the report say is staying the same?
The author says teams and planning remain important and that non-engineers are not broadly shipping software independently. These are observations in the report, not industry-wide statistics.
Source: rss
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
