TL;DR
Get wellness gear delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
Software engineer Neil Kakkar says he merged about 11 times as many pull requests per month at AI-native startup Tano as he did at PostHog, based on his personal work records. His account describes how AI tools and a permissive shipping culture changed his workflow, while creating new pressures around context, focus and quality.
Software engineer Neil Kakkar says he merged about 11 times as many pull requests per month after joining AI-native startup Tano than he had at PostHog, according to a first-person essay published by The Pragmatic Engineer. The comparison offers a detailed account of how AI tools and startup practices can change an individual engineer’s output, while also bringing new trade-offs in attention, coordination and code quality.
Kakkar’s figures compare work at PostHog, where he spent four years, with eight months at Tano, where he joined as a founding engineer. He says his work at PostHog amounted to about 45 months of output, while the Tano period covered a similar breadth of product work in a much shorter time. He reports shipping 11 times more pull requests per month at Tano than at PostHog; at PostHog, he says, he shipped 2.4 times as many as he had at Bloomberg, where he began his career.
The engineer describes PostHog as a fast-moving company before the current generation of AI coding agents became available to him. There, he worked on analytics infrastructure, experimentation, surveys and feature flags. At Tano, he says, he built an affiliate platform, improved a content-monitoring system and helped migrate the company’s agent architecture to a system with persistent memory, new tools and improved observability.
Kakkar attributes much of the change to AI, but also to the company’s engineering culture. At Bloomberg and PostHog, he says, code review and peer review were standard parts of the process. At Tano, he describes less review and communication overhead, allowing work to move faster. He also warns that pull-request counts are an imperfect measure and says his figures are a personal record, not a controlled comparison of engineering productivity across companies.
AI Changes the Engineering Workflow
The account illustrates why the spread of coding agents is prompting questions beyond whether the tools can generate code. If an engineer can open and merge more changes, teams may be able to test ideas and deliver features faster. But more pull requests do not automatically mean more valuable or reliable software, and the essay does not establish that AI alone caused the difference.
Kakkar says the faster pace introduced its own constraints: frequent context switching, weaker shared understanding, less time for design and concerns about limits on AI plans. Those costs matter to companies considering how to change review and release practices. The experience suggests that teams may need to adapt their coordination and quality controls as well as adopt new tools, though the essay does not measure the effects across a wider workforce.
Products Worth Considering
Oral-B iO Series 5 Rechargeable Electric Toothbrush with 3 Brush Heads and Travel Case, Black
Therabody SmartGoggles 2nd Gen Heated Eye and Temple Massager
Fitbit Versa 2 Health and Fitness Smartwatch with Heart Rate, Alexa, Music, Sleep and Swim Tracking
As an affiliate, we earn on qualifying purchases.
Three Companies, Different Baselines
Kakkar’s career spans workplaces with different sizes and operating styles. He describes Bloomberg as a large financial technology company with roughly 20,000 employees and defined processes. PostHog was a much smaller startup when he joined, growing from about 10 people to around 200 during his time there. Tano had around 10 people in total when he wrote the essay, and Kakkar was its first engineer.
He treats his PostHog experience as a pre-AI baseline, explaining that AI use there was largely limited to code completion and repetitive tasks such as producing test cases. He says that by the time he joined Tano, AI agents could take on substantially larger coding tasks. He also notes that PostHog’s approach has since changed and says the company is now actively using AI; the comparison is not a claim about PostHog’s current output.
The essay was published as the winning entry in an essay challenge marking The Pragmatic Engineer’s fifth anniversary. The publication said it received 125 submissions and awarded Kakkar’s piece the grand prize. It said it planned to feature runner-up essays in a later issue.
“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 reported multiplier is based on one engineer’s experience and the author’s own records. The source does not provide a full methodology for calculating the monthly averages, nor does it independently verify the figures. Pull-request counts also do not show the size, complexity or lasting value of each change, or how much work was required to maintain it.
It remains unclear how much of the difference came from AI tools, company size, job responsibilities, product scope or Tano’s operating practices. Kakkar himself points to culture as a factor, and the companies represent different stages and periods. The essay also does not provide team-wide data, defect rates, customer outcomes or a direct comparison of software quality.
More Essays, Broader Evidence Needed
The Pragmatic Engineer said it would publish or link to other strong submissions in a follow-up issue. Those accounts may add perspectives, but they would not by themselves settle how much AI changes engineering productivity across different companies.
For now, Kakkar’s essay is best read as a case study, not a general benchmark. A broader assessment would need comparable data across engineers and teams, along with measures of quality, maintenance and outcomes—not just the number of pull requests merged.
Key Questions
What change in output does Neil Kakkar report?
Kakkar says he merged about 11 times as many pull requests per month at Tano as he did at PostHog. He also reports that his PostHog rate was 2.4 times his rate at Bloomberg.
Does the comparison prove AI caused the increase?
No. Kakkar attributes much of the change to AI, but his comparison is personal and also involves different companies, roles and working practices. It does not isolate AI as the sole cause.
What downsides does Kakkar describe?
He points to context switching, loss of shared context, less focus on design and concerns about AI plan limits. He also describes faster work with less review and communication overhead.
Are pull-request counts a reliable productivity measure?
They show how many changes were merged, but not their size, quality, maintenance burden or impact. Kakkar acknowledges that pull requests can be gamed and presents his figures as a personal record rather than a definitive productivity measure.
Source: newsletter.pragmaticengineer.com
