TL;DR
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
In an essay published by The Pragmatic Engineer, software engineer Neil Kakkar says he merged about 11 times as many pull requests per month at AI-native startup Tano as at PostHog. The figures come from one engineer’s work history, not a controlled study, and he says faster output brings new costs, including context switching and weaker shared understanding.
Software engineer Neil Kakkar says he merged about 11 times as many pull requests per month after moving from PostHog to AI-native startup Tano, according to an essay published by The Pragmatic Engineer. His account offers an individual comparison of software work before and after the arrival of coding agents, while also describing new pressures that can accompany faster delivery.
Kakkar’s comparison draws on his work at PostHog, where he spent four years, and Tano, where he joined as a founding engineer. He reports shipping 11 times more pull requests per month at Tano than at PostHog, and says his PostHog rate was 2.4 times his rate at Bloomberg, where he began his career. The newsletter’s introduction gives a broader rounded estimate of more than 200 pull requests per month at Tano, compared with eight to 20 per month before. These are his personal figures, not an industry-wide measurement.
The jobs involved substantial product work in both settings. At PostHog, Kakkar worked on analytics infrastructure, experimentation, surveys and feature flags over four years. At Tano, he says he covered comparable breadth in eight months, including building an affiliate platform, improving a content-monitoring system and migrating the company’s agent architecture. He credits AI as a major factor in the increased output, alongside differences in team culture and working practices.
His account contrasts PostHog’s pre-AI workflow, where tools mainly offered code completion, with Tano’s use of agents capable of taking on larger feature-building tasks. At Bloomberg and PostHog, he says, peer review was required and senior engineers devoted substantial time to reviewing code. At Tano, less review and communication overhead helped work move more quickly, but the tradeoff was a greater need to manage quality and coordination as the team changed how it built software.
Faster Shipping, New Work Pressures
Kakkar’s experience points to a change in what can constrain software teams. If coding agents reduce the time needed to produce implementation work, then coordination, product judgment and review may become more prominent bottlenecks. His essay describes context switching, loss of shared understanding, less focus on design and concern about AI usage limits as emerging costs of the faster workflow.
The account matters to engineers and managers weighing AI adoption because it does not present output as a simple measure of success. Pull-request counts can be increased without equivalent gains in quality or customer value. Kakkar acknowledges that limitation and says he did not set out to game the metric, but his figures still reflect one person’s work across different companies, roles and periods. They illustrate a possible shift, rather than proving how much AI raises productivity across the industry.
Products Worth Considering
LG UltraGear 34GX90SA-W 34-inch WQHD Curved OLED Gaming Monitor
LG UltraGear 32GS60QC-B 32-inch Curved QHD Gaming Monitor (180Hz, 1ms, 1000R, AMD FreeSync, HDR10)
Logitech MX Keys Combo for Business Keyboard and Mouse – Graphite
As an affiliate, we earn on qualifying purchases.
From Code Completion to Agents
Kakkar worked at Bloomberg for three years, then spent four years at PostHog before joining Tano. The companies differed in size and structure: he describes Bloomberg as a company of about 20,000 people with defined processes, PostHog as growing from roughly 10 people to about 200 during his tenure, and Tano as a startup that had about 10 people when he wrote the essay. Those differences matter when comparing pace as well as tools.
He treats PostHog as a pre-AI baseline because, until late 2024, he says AI use there largely meant tab completion and help with repetitive tasks such as writing test cases. The newsletter notes that PostHog has since adopted AI more extensively and that its output is increasing, so Kakkar’s comparison is not a direct account of the same team before and after an AI rollout. He says Tano was founded after AI agents emerged and that agent support was already part of his workflow when he joined.
The essay was selected as the winning entry in The Pragmatic Engineer’s challenge on changes in technology workplaces and projects. The publication says it received 125 submissions and awarded Kakkar’s entry the grand prize. It says a later issue will highlight runner-up essays.
“I’m shipping 11x more PRs at Tano than at PostHog, where I shipped 2.4x as many as I did at Bloomberg.”
— Neil Kakkar, software engineer and essay author
What the Output Figures Cannot Show
The figures do not establish that AI alone caused the change. Kakkar moved between companies with different sizes, cultures, responsibilities and stages of development, and he says he was more experienced at Tano than when he started at Bloomberg. The essay offers no controlled comparison isolating the effect of AI from those factors.
It is also unclear how the reported increase in pull requests relates to measures such as defects, time to deliver customer value, maintenance costs or long-term product quality. The source does not provide a common measurement method across the companies or a fuller account of how Tano’s review and testing practices compare with PostHog’s. Kakkar’s experience should not be read as a representative estimate for software engineers generally.
More Essays, Wider Questions
The Pragmatic Engineer says a later issue will feature runner-up essays and other submissions it found notable. Those accounts may provide further examples of how teams are adapting, though they would still be individual perspectives rather than a formal study.
For now, the practical questions raised by Kakkar’s account are whether faster agent-assisted coding produces durable improvements in delivery and how teams preserve review, design attention and shared context as output grows. The essay describes those issues but does not report a measured industry-wide answer.
Key Questions
What is the main development in the report?
The Pragmatic Engineer published Neil Kakkar’s winning essay comparing his work before and after joining AI-native startup Tano. He reports merging about 11 times more pull requests per month at Tano than at PostHog.
Does the comparison prove AI makes engineers 11 times more productive?
No. It is Kakkar’s personal comparison across different companies, roles and periods. It does not isolate AI’s effect or measure productivity across the software industry.
What work did Kakkar do at Tano?
He says he built an affiliate platform, improved a content-monitoring system and helped migrate the company’s agent architecture. He describes completing broad product work in eight months that he compares with work done over four years at PostHog.
What problems can come with faster AI-assisted development?
Kakkar describes context switching, weaker shared context, less attention to design and worries about AI plan limits. He also says reducing review and communication overhead can make it harder to maintain coordination and quality.
What happens next?
The Pragmatic Engineer says it plans to publish highlights from runner-up essays in a later issue. The source does not announce a follow-up study measuring AI’s effects across teams.
Source: newsletter.pragmaticengineer.com
