TL;DR
Get monitors, keyboards and dev 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 keynote to engineering leaders describes AI coding tools as reshaping software development in 2026, with some engineers managing several agents at once instead of writing code by hand. The report also flags unresolved problems, including weaker reliability and code reviews that may not adequately check AI-generated work; its observations are not a comprehensive industry survey.
A 2026 report on the tech industry says AI coding agents are changing how software engineers work, with some developers coordinating five to 10 agent sessions at a time instead of writing code line by line. The account, based on a keynote to more than 2,000 engineering leaders and visits to technology companies, also identifies concerns about code quality, reliability and reviews as organizations adapt.
The report was written by The Pragmatic Engineer’s author after a keynote at the LDX3 engineering leadership conference in New York. The author says the research drew on visits to OpenAI and Anthropic, conversations with startups including Ramp and Uber, and unpublished data from GitHub, Factory AI and Linear. The material is a snapshot assembled from those sources, not a published, representative survey of the entire industry.
Its clearest reported change is the spread of parallel AI-assisted coding. Boris Cherny, identified in the report as the creator of Claude Code, described using five terminal sessions and five to 10 Claude sessions on the web at once. Software engineer Dima Zaytsev, now at Linear, said he rotates among multiple local worktrees, prompting one agent while checking another’s output. The report says such workflows are appearing among productive engineers, while cautioning that working practices at AI labs may be ahead of wider adoption.
The author also lists problems emerging alongside the change: assumptions about code output no longer hold, reviews can become “theatrical,” and software quality and reliability are down. Those are the report author’s observations; the supplied material does not quantify these effects or provide a comparative dataset. The report argues that teams and planning remain important even as tools and engineering practices change.
AI Agents Reshape Engineering Work
The shift matters because software development is a core activity across technology companies, and AI-assisted workflows can alter how teams divide work, review changes and judge productivity. If developers routinely supervise multiple agents, the job may place more emphasis on task selection, coordination and verification than on typing code directly. The report describes this as an emerging practice, not a settled description of every engineering team.
Its quality concerns point to a practical challenge for companies adopting these tools: producing code faster does not by itself show that software is dependable or adequately reviewed. The report offers no measured industry-wide effect on output or defect rates, so readers should distinguish its firsthand observations and attributed comments from broader claims about the whole sector.
As an affiliate, we earn on qualifying purchases.
A Faster Shift in Software Development
The report places the current AI shift against earlier changes that technology workers have experienced, including the growth of the internet, smartphones, cloud computing and new programming languages and frameworks. Martin Fowler, an industry veteran quoted in the report, said AI’s impact was substantially larger in scale than previous changes he had seen.
The author links the acceleration to improvements in models’ coding ability late in 2025, but the source gives no specific benchmark or model comparison to quantify that change. The report’s broader point is that software practices are changing quickly, while some fundamentals—such as the need for teams and planning—remain.
“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, as quoted in The Pragmatic Engineer report
code review tools for AI-generated code
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
How Widely the Practices Have Spread
The report does not establish what proportion of engineers or companies use coding agents, how often developers work with five to 10 sessions, or whether those practices improve productivity overall. Its claims about declining quality and reliability are not accompanied in the supplied material by rates, time periods or a comparison baseline. It is also unclear how much the observations from AI labs and technology companies reflect practices in smaller firms and other industries.
The report describes several changes as emerging and predicts they may accelerate, but it does not give a timeline or measured forecast. The balance between faster code generation and the work needed to test, review and maintain that code remains an open question.
reliable SSD for software developers
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Quality Controls Will Shape Adoption
The report expects cloud-based coding agents and supporting “harnesses”—the surrounding tools and processes used to run agents—to become more common. It also anticipates changes in how engineers interact with code and the development of new AI infrastructure. These are the author’s expectations, not confirmed outcomes or scheduled industry milestones.
For technology teams, the next test will be whether they can integrate agent-assisted work while maintaining effective review, reliability and clear accountability. The report points to these as active areas of change; the supplied material does not identify a single next event or publication date that will settle how widespread the practices are.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is changing in software engineering in 2026?
The report describes more engineers using AI coding agents to generate and check code, sometimes running several sessions in parallel instead of writing every line by hand. It does not claim this describes all engineers or companies.
How many AI agents do engineers use at once?
Several people quoted in the report describe working across roughly five to 10 agent sessions. These are individual examples, not an industry-wide average.
What problems does the report identify?
The author flags concerns about code quality, reliability and reviews, including reviews that may not meaningfully assess generated code. The supplied material does not quantify these problems.
Does the report show that AI has improved developer productivity?
It includes accounts of engineers using agents to handle work in parallel, but it does not provide a measured, industry-wide productivity comparison. Individual workflows do not establish an overall effect.
What is expected to happen next?
The author expects wider use of cloud coding agents and supporting infrastructure. These are forecasts in the report, and their pace and impact remain uncertain.
Source: rss
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
