AI agents replace junior coders
- Junior developer jobs are vanishing as AI agents write, test, and debug code autonomously.
- Engineers now manage AI swarms, making communication - not coding - the core skill.
- Firms are shedding legacy staff to rebuild teams around AI-native builders.
AI coding agents are no longer assistants. They’re replacements. At Google, AI now writes 40-50% of code. Startups move 20 times faster than two years ago. The developer’s job has shifted from typing lines to commanding swarms - where one agent writes code, another tests it, and a third fixes errors in a recursive loop.
Clive Thompson found that most of the 75 developers he surveyed now outsource daily programming to AI, some writing little to no code themselves. The change accelerated sharply in the last three months. At scale, this isn’t just automation - it’s deskilling. Stanford research shows software job postings down 16%. The entry-level grind, once the industry’s training ground, is disappearing.
"The developer's role is shifting from writing code to specifying what the software should do."
- Clive Thompson, The Daily
Scott Chacon sees the same shift from a different angle. Git, built for humans in the 1970s, now fails AI agents that need rich, structured context - not raw grep pipelines. His team at GitButler is designing tools for agents as first-class users, not shell scripts. Parallel, stacked branches let dozens of agents work simultaneously, turning development into a high-concurrency factory.
The implications cut beyond engineering. Nikhyl Singhal warns that product managers who merely move information are already obsolete. AI handles status reports, ticket tracking, and documentation. Judgment - what to build and why - is the only defensible skill left. Companies aren’t just cutting costs; they’re swapping out entire workforces.
"Companies realized they doubled headcount without doubling productivity. Now they’re clearing decks to hire builders."
- Nikhyl Singhal, Lenny's Podcast
The old career ladder - climb at Meta, retire at Google - is broken. Resume logos mean less than fluency with AI tools. Interviews now test whether you can build, not just manage. The threshold isn’t tenure. It’s whether you’ve crossed into the new workflow. Those who haven’t are already behind.
Source Intelligence
- Deep dive into what was said in the episodes
Rethinking Git for the Age of Coding Agents with GitHub Cofounder Scott Chacon • Apr 20
- Matt Bornstein notes that coding agents are now the fastest-growing users of command-line tools, introducing a new persona that struggles with interactive Git commands like `rebase` and frequently runs `status` after every command.
- Scott Chacon reveals that 80% of developers still use the Git command-line interface, finding GUIs generally lack added functionality beyond basic command wrapping.
- Git-Butler offers multi-persona interfaces (GUI, CLI, TUI) operating on the same data structures, allowing optimization for humans or agents through features like hints for humans or `dash-dash JSON` for scripting.
- Scott Chacon notes that agents prioritize different outputs; while `dash-dash JSON` was expected, agents often preferred human-readable output to process with tools like `JQ`, and consistently ran `status` after mutable commands.
- Git-Butler's parallel branch system allows multiple agents or humans to work on the same codebase and commit to different, isolated branches from a single working directory, avoiding the conflicts common with Git worktrees.
- Scott Chacon found that inter-agent chat channels, while 'super cool,' did not improve productivity because agents independently deduce changes and adapt, finding direct communication an unnecessary overhead.
- Scott Chacon believes the current pull request (PR) system is suboptimal, advocating for patch-based local review that agents can automate, rather than the centralized URL-based review that encourages 'commit slop.'
- Scott Chacon asserts that the next 'superpower' for software developers will be communication and writing skills, as agents increasingly handle implementation details, shifting the focus to clear specifications and 'write-ups.'
- Scott Chacon highlights that agent-assisted development shifts the constraint from coding to achieving team consensus on 'what we want,' making the iterative process of writing a spec, building a proof of concept, and refining it more efficient.
- Storing all agent activity - including prompts, tool calls, and 'thinking logs' - creates a significant big data problem that rapidly balloons storage requirements, necessitating advanced Git primitives for metadata management.
Also from this episode: (4)
Coding (3)
- Scott Chacon highlights that Git, the most widely used developer tool, was never designed as a user-friendly product but evolved from Unix plumbing commands for the Linux kernel team, intended to be wrapped by scripts.
- Scott Chacon explains that GitHub had a grudging relationship with the Git core team, who valued Git's speed and reliability but disliked GitHub's PR and issue primitives, contributing to Git's 'Frankenstein' design through committee.
- Git-Butler aims to inject 'taste' into Git's user interface while retaining its robust storage and data transmission layers, which were built using the Unix philosophy of small, pipeable tools.
Startups (1)
- Scott Chacon, co-founder of GitHub and CEO of Git-Butler, returned to a startup in the version control space because existing Git tooling has not significantly changed since 2005, presenting an opportunity to rethink its user interface for modern needs.
Why half of product managers are in trouble | Nikhyl Singhal (Meta, Google) • Apr 19
- Nikuel Singal notes that product leaders previously focused on moving information, a role he believes will become obsolete as AI transforms product development and requires a shift towards direct building.
- Lenny's recent report indicates that open product manager roles globally are at their highest level in over three years, showing continued demand despite industry stress and rapid change.
- Nikuel Singal predicts a massive shedding of staffs and subsequent rehiring within the next 12 to 24 months, with new hires being primarily 'AI-first' builders focused on new skill sets.
- The industry is experiencing a 'renaissance' for product builders who are having fun, seeing record-high compensation, and more opportunities, including founding companies or moving into C-level roles outside product.
- Product managers are increasingly asked to drive judgment and obsolete mechanical tasks through AI, focusing on evaluating changes rather than traditional information flow or backlog management.
- Nikuel Singal believes bad software will largely disappear in two years because AI tools like Claude and CodeX enable rapid bug fixes, security enhancements, and overall product improvement.
- Nikuel Singal encourages product professionals to find joy in using new AI tools for building, noting that personal projects or internal efficiency tools can transition fear into an infectious enthusiasm.
- Nikuel Singal uses Claude extensively for his web properties and community management, building AI agents for member matching, job placement, and content generation to obsolete his own manual tasks.
Also from this episode: (5)
Labor (1)
- The importance of company logos on resumes is diminishing; instead, modern product professionals must demonstrate current skills and a willingness to embrace new ways of building products with AI.
Mental Health (1)
- Nikuel Singal finds that the intense pace of change in the AI era is causing significant stress and may disproportionately affect diversity in tech, as some demographics struggle to keep up.
Psychology (1)
- To thrive, PMs must overcome psychological blocks to reinvention, increase their work pace, be willing to swallow their ego by taking smaller roles, and maintain a long-term career focus on 'skip opportunities'.
Autonomous Vehicles (1)
- Lenny and Nikuel Singal both find Tesla's full self-driving software (specifically version 14.2) to be highly effective, with Nikuel using it 95% of the time and experiencing reduced driving anxiety.
AI & Tech (1)
- Nikuel Singal's life motto, 'Genius is 1% inspiration, 99% perspiration,' takes on new meaning with AI, as inspiration becomes paramount while AI handles much of the perspiration.
The Workers Letting A.I. Do Their Jobs • Apr 14
- At large firms like Google, AI writes 40-50% of code, increasing overall development speed by about 10%, which is considered a huge win at scale.
- Developers now work with AI agents in a swarm, where a main agent spawns sub-agents to write code, test it, and fix errors in an automated loop before presenting the final product.
- To control AI agents, developers write stern, repetitive command files with emotional language, which appears effective because large language models understand the contextual weight of words like 'embarrassing' or 'unacceptable'.
- Thompson argues that historically 'hard' technical skills like coding are easier to automate than 'soft' skills like strategy, prioritization, and understanding human needs, which may become the core of future white-collar work.
- Full economic impact will be slow because companies must reorganize workflows around AI, similar to the decades-long lag between personal computer adoption and measurable productivity gains.
- A potential upside is that cheaper, faster software development could serve mid-sized industries currently underserved by technology, like a $50M concrete company running on outdated spreadsheets.
Also from this episode: (8)
Coding (3)
- Clive Thompson found a majority of the 75 software developers he surveyed were outsourcing significant day-to-day programming to AI, with some writing very little to no code themselves.
- This shift accelerated heavily in the last six months and dramatically in the last three months as AI coding tools improved and gained developer trust.
- Small startup developers report moving up to 20 times faster with AI, completing feature requests that took a full day in about 30 minutes.
AI & Tech (5)
- The developer's role is shifting from writing code to specifying what the software should do, becoming more like an architect or a product manager who iterates through AI-generated options.
- Developers are having constant conversations with AI, prompting them to become clearer communicators, which some report improves their overall human communication skills.
- A primary concern is deskilling, where developers worry they and the next generation will lose 'code sense' - the deep understanding needed to debug, maintain, and foresee subtle interactions in complex systems.
- Stanford researcher Eric Benjolson found job postings and hirings for software developers were down by 16% recently, indicating early AI impact on labor demand.
- Thompson compares the AI coding revolution to the proliferation of paper or word processors, predicting software will become a ubiquitous, trivial-to-summon tool that catalyzes unpredictable social and creative behaviors.