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, where his work preceded the current generation of coding agents. The figures are a single engineer’s account, not an industry-wide measure; Kakkar also describes new costs, including more context switching and weaker 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, in an account published by The Pragmatic Engineer as its winning essay on changes in tech workplaces. The comparison offers one engineer’s measure of how AI tools and company practices may affect software output, while also describing new costs that raw counts do not capture.
Kakkar reports that his work at Tano covered a similar breadth of product development to his four years at PostHog, but was compressed into eight months rather than 45. At Tano, where he joined as a founding engineer, he says he worked on an affiliate platform, improved a content-monitoring system and moved the company’s agent architecture to a system with persistent memory, different tools and improved observability.
His reported pull-request rate at PostHog was 2.4 times his rate at Bloomberg, where he began his career. Kakkar describes Bloomberg as a large organization with defined processes, and PostHog as a faster-moving startup where engineers were encouraged to take responsibility and respond to customers. In both workplaces, he says, code required peer review and senior engineers spent substantial time reviewing others’ work.
The comparison is based on Kakkar’s personal work records, not an independently audited study. He cautions that pull-request counts can be gamed, though he says he did not manipulate his figures. The essay says AI at PostHog until late 2024 was mainly used for code completion and repetitive tasks; by the time he joined Tano, he had access to coding agents capable of handling larger pieces of work.
Higher Output, New Engineering Costs
Kakkar’s account points to a change in how a small engineering team can divide and complete work. Coding agents can take on implementation tasks, while a company willing to experiment may be able to move projects forward with fewer conventional handoffs. His experience suggests that engineering culture and tools both shape output; access to AI alone is not the full explanation.
But more pull requests do not necessarily mean better software or greater business value. Kakkar says the faster pace brings more context switching, less shared context and reduced attention to design. He also describes concern about AI plan limits. For teams adopting similar tools, the account highlights a practical tradeoff: speed may rise while coordination, review and product judgment become harder to sustain.
Products Worth Considering
Jackery HomePower 3600 Plus Portable Power Station, 3584Wh, 3600W
Bose S1 Pro+ All-in-One PA Speaker
Bose SoundLink Flex Bluetooth Speaker (2nd Gen), Waterproof, Dustproof, USB‑C, Up to 12 hr Battery, Black
As an affiliate, we earn on qualifying purchases.
From Peer Review to Coding Agents
Kakkar’s career comparison spans three different workplaces: Bloomberg, where he spent three years; PostHog, where he worked for four years; and Tano, where he joined as its first engineer. Those settings differ in size and maturity, so the change in his output cannot be attributed to AI alone. He credits both the tools and a culture designed around faster experimentation.
The essay’s editor says the publication received 125 submissions for its fifth-anniversary challenge and selected Kakkar’s entry as the winner. The report presents PostHog as a pre-AI reference point for Kakkar’s own work, not as a current assessment of the company: the editor notes that PostHog has since adopted AI heavily and that its output is increasing, citing the company’s own direction.
“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 for The Pragmatic Engineer
What the Pull-Request Count Cannot Show
The figures cover one engineer and three workplaces with different sizes, stages and expectations. The account does not establish how much of the reported increase came from AI agents, company culture, the type of work, or the way pull requests were divided. No independent measurement or common productivity baseline is provided, and pull-request volume alone does not show code quality, customer impact or maintenance costs.
The source also does not provide a controlled comparison of the two companies’ overall engineering output. Kakkar’s experience should therefore be read as a detailed personal account, not evidence that all teams will see an 11-fold increase from adopting coding agents.
More Essays, Broader Questions
The Pragmatic Engineer said it plans to publish a follow-up highlighting runner-up essays and other entries it found notable. Those accounts may provide additional workplace perspectives, but they will not by themselves settle how AI changes engineering productivity across companies.
For now, the next test for teams adopting coding agents is whether increased implementation speed can be matched with sound review, design and coordination. Kakkar’s account identifies those pressures but does not quantify their effects or establish which practices can address them.
Key Questions
Who made the 11-fold productivity claim?
Neil Kakkar, a software engineer, reported the difference in an essay published by The Pragmatic Engineer. It describes his own work, not a company-wide or industry-wide result.
What does the 11-fold figure measure?
Kakkar says he merged 11 times as many pull requests per month at Tano as at PostHog. Pull-request volume is a measure of submitted code changes, not a direct measure of software quality or business impact.
What role did AI play in the comparison?
Kakkar says his PostHog work through late 2024 mostly involved AI code completion and routine tasks, while Tano used coding agents that could handle larger tasks. His account does not isolate AI’s effect from differences in company culture, role or project mix.
What problems did Kakkar report alongside faster output?
He describes context switching, loss of shared understanding, less focus on design and concerns about AI plan limits. The essay does not quantify how these issues affected outcomes.
Source: newsletter.pragmaticengineer.com
