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 keynote to engineering leaders describes AI as rapidly changing software development, with many developers using multiple coding agents rather than writing every line by hand. The report also flags weaker code quality and more performative reviews as concerns, while stressing that teams and planning remain important. Its observations are a snapshot, not a representative industry-wide survey.
A report on a keynote delivered to more than 2,000 engineering leaders at New York’s LDX3 conference describes AI coding agents reshaping software development in 2026, as developers increasingly delegate coding tasks to several agents at once. The account also warns that code quality and reliability may be falling, while noting that its observations draw on interviews, company access and unpublished data rather than a representative survey of the entire tech industry.
The Pragmatic Engineer report says the speed and scale of AI’s impact have outpaced earlier changes familiar to software teams, including the spread of the internet, smartphones and cloud computing. It describes a shift away from writing code line by line toward directing and checking work produced by AI tools. The report characterizes this as a broad trend, but does not quantify what share of engineers or companies work this way.
Several practitioners cited in the report describe running five to 10 agent sessions in parallel. Claude Code creator Boris Cherny said he uses multiple terminal checkouts alongside agents running on Claude Web. Cockroach Labs co-founder Peter Mattis described a similar workload, sometimes with subagents. Dima Zaytsev, a software engineer at Linear and former Uber colleague of the report’s author, said he rotates among several local worktrees while agents work on separate tasks.
The report’s author also lists problems emerging alongside adoption: assumptions about AI-generated code have broken down, code reviews can become “theatrical,” and software quality and reliability are under pressure. These are the author’s observations and interpretations, not industry-wide measurements in the material provided. The report also says some fundamentals endure: teams and planning still matter, and it does not suggest that non-engineers are broadly shipping production code.
How Agent Use Changes Engineering Work
Using several coding agents at once could let developers advance multiple tasks in parallel, changing how engineering time is spent. The examples in the report suggest that a developer’s work may involve setting direction, managing agent sessions and evaluating their output as much as composing code directly. That shift affects how teams organize work and what skills they value, although the source does not establish how widespread the practice is.
The possible trade-off is between faster output and confidence in the result. If teams generate code more quickly but have less effective review or weaker reliability, they may face more work finding and correcting defects. The report’s warning is a practitioner assessment, not a measured estimate of defect rates or business impact. For organizations adopting these tools, the central question is not simply how much code agents produce, but whether teams can validate and maintain it.
As an affiliate, we earn on qualifying purchases.
From AI Coding to Parallel Agents
The report is based on a keynote by The Pragmatic Engineer’s author at LDX3, a conference attended by engineering leaders, CTOs and senior individual contributors. In preparing the talk, the author says they visited OpenAI and Anthropic, spoke with startups and other technology companies, and received unpublished data from GitHub, Factory AI and Linear. The supplied material does not include that data, so its findings cannot be independently assessed here.
Martin Fowler, an industry veteran quoted in the report from The Pragmatic Summit, compared AI’s impact with earlier shifts in software development and said it was on a different scale. The report links the current acceleration to improvements in models’ coding abilities late in 2025. Its wider point is that tools and practices may change quickly even as the need for coordinated teams and planning persists.
“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 an affiliate, we earn on qualifying purchases.
How Widespread Are These Practices?
The report offers examples from experienced engineers and access to selected technology companies, but the supplied material does not provide survey methods, sample sizes or the unpublished company data mentioned by its author. It is not clear how many engineers have stopped writing code by hand, how common multi-agent workflows are across different company sizes, or whether they improve delivery outcomes overall.
The reported concerns about code review, quality and reliability are not accompanied here by measurements or comparisons with earlier periods. The material also does not specify which coding tools or model versions were evaluated, or how the observations vary between teams. These limits make the report a useful snapshot of practices and concerns, but not a definitive census of the industry.
As an affiliate, we earn on qualifying purchases.
What Teams Will Watch Next
The report expects cloud-based coding agents and the systems used to coordinate them—often called harnesses—to become more prominent. It also anticipates changes to AI infrastructure and to the way engineers inspect code. These are forecasts in the report, not confirmed outcomes or a schedule for particular product releases.
For engineering teams, the next test is whether agent-heavy workflows can produce software that remains dependable and maintainable. More evidence would be needed to compare quality, review effectiveness, productivity and costs across teams using different approaches. The report’s snapshot points to a fast-moving shift, but its longer-term effects remain unsettled.
As an affiliate, we earn on qualifying purchases.
Key Questions
What is the main development described in the report?
The report says AI coding agents are changing developers’ work, with some engineers directing several agents at once instead of writing every line by hand.
Does the report show that most engineers have stopped coding by hand?
No. It describes this as a trend and cites individual practitioners, but the supplied material includes no representative survey or industry-wide percentage.
What risks does the report identify?
Its author raises concerns about code quality, reliability and less substantive code reviews. The source material does not provide measured defect rates or other data to quantify those concerns.
Why does the report cite engineers using five to 10 agents?
The examples show how some developers split work across concurrent sessions while checking or testing outputs. They are individual accounts, not a recommended standard or proof that the method works for every team.
What is expected to change next?
The report anticipates wider use of cloud coding agents and coordination tools, alongside changes to AI infrastructure and code inspection. These are expectations, not confirmed timelines.
Source: rss
Halloween Picks
halloween
As an affiliate, we earn on qualifying purchases.
