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
Software engineer Neil Kakkar reports merging 11 times as many pull requests per month at AI-native startup Tano as he did at PostHog before widespread agentic AI use. His account, published by The Pragmatic Engineer as the winning entry in an essay challenge, describes faster output alongside new costs, including context switching and reduced shared understanding.
Software engineer Neil Kakkar says he merged 11 times as many pull requests per month at AI-native startup Tano as he did at PostHog, where he worked before the current generation of coding agents became available. The comparison is based on his own work data, not a controlled study, and highlights both the speed gains he associates with AI and the workflow problems that came with them.
The Pragmatic Engineer published Kakkar’s essay as the winner of a reader challenge on how technology workplaces and projects have changed. The publication said it received 125 submissions and awarded the winning article a $10,000 grand prize. Kakkar is a London-based engineer with nine years of professional experience, according to the publication.
Kakkar says he worked at PostHog for four years, through late 2024, and then spent eight months at Tano, where he joined as its first engineer. He reports that his monthly pull-request output at Tano was 11 times his PostHog rate; at PostHog, he says, he shipped 2.4 times as many as at Bloomberg, where he began his career. He cautions that pull-request counts can be gamed and says his figures are an honest record of his work, but they do not independently measure quality or impact.
His account links the change to both AI tools and workplace culture. At Bloomberg and PostHog, he describes peer review as a requirement and says senior engineers devoted substantial time to reviewing code. At Tano, he says, AI agents let him work across broad projects at a faster pace, including building an affiliate platform, improving a content-monitoring system and migrating the company’s agent architecture. The source contrasts older AI use, such as code completion and repetitive test generation, with newer agents capable of carrying out larger tasks.
Faster Shipping, New Workflow Costs
Kakkar’s comparison offers a concrete, though individual, account of how coding agents may change the volume and shape of engineering work. If an engineer can produce more code and pull requests in less time, teams may deliver features faster. But the essay also suggests that output volume is not the same as product quality: more changes can increase the need to coordinate, track decisions and catch defects.
Kakkar identifies new pressures at Tano, including context switching, loss of shared context and difficulty maintaining focus. He also worries about AI plan limits and says rapid execution can leave less attention for design. These observations matter to engineering managers weighing adoption: the tools may reduce some bottlenecks, but existing review practices, communication norms and systems may not keep pace automatically.
The report does not establish that AI caused an 11-fold increase on its own. Kakkar moved between companies of different sizes and cultures, took on different roles and compared periods with different technology. His experience is useful as a case study, not a forecast for every developer or organization.
Products Worth Considering
Amazon Kindle Scribe (64GB) with Premium Pen, Metallic Jade
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Black
DJI Osmo Mobile 8 Gimbal with Mic Mini Transmitter, Infinity Black
As an affiliate, we earn on qualifying purchases.
From Code Completion to Agents
Kakkar’s career provides three different reference points. At Bloomberg, a large financial company, he describes more defined processes. PostHog was a smaller, fast-moving startup when he joined and grew from about 10 people to around 200, according to his account. Tano was about 10 people when he wrote the essay, and he joined as its first engineer.
He treats his PostHog work as a pre-AI baseline, noting that AI there was largely limited to tab completion and assistance with repetitive tasks. He says he used contemporary agent tools at Tano. He also cautions that PostHog has since adopted AI extensively and that its output is increasing, so the comparison does not represent a permanent divide between the two companies.
The Pragmatic Engineer said a follow-up issue would highlight runner-up essays. Its publication of Kakkar’s piece is an editorial selection, while the work-rate figures and workplace observations are Kakkar’s own account.
“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, in his essay published by The Pragmatic Engineer
What the Output Figures Cannot Show
The essay does not provide an independent audit of Kakkar’s pull-request counts, a shared definition of a pull request across the companies or a measure of how many changes reached customers. It also does not compare defect rates, maintenance work, customer outcomes or the time spent prompting and reviewing AI-generated code. The 11-fold figure is a personal comparison, not a measured industry-wide effect.
It is also difficult to separate the contribution of AI from changes in company size, responsibilities, team practices and product scope. Kakkar says AI and culture both helped enable the shift, but the account does not quantify the effect of either factor separately. The essay’s claims about PostHog’s more recent AI-related output are not accompanied here by figures or an external assessment.
More Essays and Broader Evidence
The Pragmatic Engineer said it planned to publish a follow-up highlighting runner-up submissions and other essays that stood out. Those accounts may add perspectives on how AI is changing engineering work, though they would remain individual experiences unless supported by comparable data.
For now, Kakkar’s report raises questions that broader workplace evidence could address: whether higher code output leads to faster customer delivery, how review and design practices change, and whether teams can preserve shared context as agents take on larger tasks. His figures describe his own experience; answers for the wider industry remain open.
Key Questions
What change did Neil Kakkar report?
Kakkar says his monthly pull-request output at AI-native startup Tano was 11 times his rate at PostHog. He attributes the change to AI tools as well as the engineering culture at Tano.
Is the 11-fold increase an industry-wide finding?
No. It is Kakkar’s self-reported comparison across his own jobs. The essay is not a controlled study and does not establish what other engineers or companies should expect.
What new difficulties did he describe?
Kakkar reported context switching, weaker shared context and less focus on design as concerns. He also raised the challenge of pacing work and the possibility of AI plan limits.
Why compare Tano with PostHog?
Kakkar says his PostHog work through late 2024 took place before current agentic tools were in use there. He uses it as a pre-AI baseline, while noting that PostHog has since adopted AI extensively.
Source: newsletter.pragmaticengineer.com
