TL;DR
Get the latest gadgets 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 coding agents as a fast-growing force in software development, with many engineers delegating coding tasks to several agents at once. The account also flags concerns about code quality and reviews, while stressing that teams and planning still matter. It is a snapshot based on interviews and industry observations, not a comprehensive measure of the entire tech sector.
AI coding agents are changing how software engineers work, with some experienced developers now managing several agent sessions in parallel rather than writing code line by line, according to a 2026 industry report based on a keynote at the LDX3 engineering leadership conference in New York. The shift is unfolding quickly, but the report also identifies weaker code quality and increasingly performative reviews as risks that companies have yet to resolve.
The Pragmatic Engineer report says its author assembled a snapshot from the conference, visits to OpenAI and Anthropic, conversations with startups and technology companies, and data provided by GitHub, Factory AI and Linear. The account focuses on AI labs, venture-backed startups and large technology companies; it is an informed industry survey, not a published census of developer practices.
Several engineers described coordinating five to 10 AI agent sessions at the same time, switching between tasks while agents generate or test code. Claude Code creator Boris Cherny said he works across multiple terminal checkouts and also runs agents through Claude Web. Linear engineer Dima Zaytsev described rotating between local worktrees, prompting one agent and reviewing another’s output. These are individual accounts, not evidence that all developers work this way.
The report’s author says that hand-writing code is becoming less common among engineers he has encountered, and that the traditional IDE may be losing its central role. At the same time, the report highlights problems: assumptions about how much usable code agents produce no longer hold, code reviews can become “theatrical,” and quality and reliability have declined in some observed workflows. The report does not quantify those changes or provide a sector-wide baseline.
AI Changes the Engineer’s Daily Work
If engineers spend more time directing agents and checking their output, companies may need to change how they assign work, evaluate productivity and maintain software. The shift affects more than coding speed: testing, review and accountability become central when code is produced faster than people can inspect it.
The report warns that the tools do not remove the need for people or sound engineering practices. It says teams and planning remain important, even as individual workflows change. For technology leaders, the near-term challenge is to capture potential gains from agents without treating generated code as reliable by default.
As an affiliate, we earn on qualifying purchases.
From Better Models to Agent Workflows
The report places the current changes after a period of improvement in AI models’ coding abilities that it says gathered pace toward the end of 2025. It compares the resulting shift with earlier changes such as the spread of the internet, smartphones and cloud computing, but argues that AI is arriving on a different scale and at a faster pace.
At the Pragmatic Summit, software engineer Martin Fowler said AI’s impact was larger than other major changes he had experienced in software development. His assessment is a professional judgment, not a measured comparison across technology eras. The report also argues that established parts of engineering have not disappeared: planning and collaboration still matter, and it says non-engineers are not generally shipping code themselves.
“Nothing has hit with the magnitude of AI. This is a whole size difference from anything that we’ve faced before.”
— Martin Fowler, software engineer, speaking at The Pragmatic Summit
integrated development environment (IDE) for AI coding
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
How Broad Are These Practices?
The report does not establish what share of engineers use multiple agents, how typical the cited workflows are, or how practices differ across industries and company sizes. Its observations and interviews offer a snapshot, not representative statistics.
It also gives no numerical measure of changes in code quality, reliability or review effectiveness, and does not specify a common baseline for comparison. The extent to which agent-assisted coding improves delivery time or business results therefore remains unclear. The report describes further changes, including cloud-based coding agents and new AI infrastructure, as trends that may accelerate, rather than settled outcomes.
code review tools for AI-generated code
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Quality Controls and Agent Infrastructure
The report anticipates wider use of cloud coding agents and the systems that coordinate them, alongside investment in new infrastructure for AI-assisted software development. Those directions are presented as emerging trends, not confirmed predictions about adoption across the industry.
For companies adopting these tools, the next practical test is whether they can keep software dependable as more code is generated by agents. That will require organizations to watch how review, testing and responsibility change in their own teams. The report offers no timeline or industry-wide standard for that work; the effects on engineering roles and established practices remain in development.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is the main change described in the report?
The report says some software engineers are delegating coding tasks to multiple AI agents at once and spending more time directing and reviewing their output.
Does the report show that most engineers use AI agents?
No. It presents interviews and observations from particular technology companies and engineering leaders. It does not provide a representative industry survey or a percentage of engineers using agents.
What concerns does the report raise?
It points to concerns about code quality, reliability and reviews, which may not keep pace with faster code generation. It does not quantify these problems or establish how widespread they are.
What does the report say still matters?
It says teams and planning remain important, even as AI changes engineers’ day-to-day work. The report does not claim that agents eliminate the need for engineering oversight.
Source: rss
Fall Picks
fall essentials
As an affiliate, we earn on qualifying purchases.
