Gergely Orosz

A long interview with Addy Osmani on the path from building browser tooling to AI developer experience, and on what changes in how engineers work when the assistant is in the loop.

Ramp's engineering leadership walks through Inspect, the in-house coding agent it built rather than adopting a third-party harness: remote sandboxes with access to internal data sources, many agents in parallel, backend and frontend changes verified remotely, and bugfixing driven from Slack.

Gergely Orosz interviews seven OpenAI engineering leaders on how Codex became the backbone of the company's work, with IDE usage declining and PR review being rebuilt around an agentic 'software factory' of automated production and performance monitoring loops.

One team migrated off a testing framework in two weeks with AI, a rewrite it had put off for years, and it is not the only one finding that framework migrations are suddenly tractable. The issue also covers the engineering leader career break and what AI startups do to analyst firms.

Gergely Orosz on napkin math, the practice of reasoning about system limits from a handful of memorised latency and throughput numbers, following an interview with turbopuffer's Simon Eskildsen.

A Pragmatic Engineer episode walks through Uber's internal developer and AI infrastructure, including DevPods, Code Inbox, Slate and Wisdom, and how engineers use them day to day.

Dex Horthy on context engineering: what actually goes into an agent's window, and why managing that is the discipline rather than prompt wording.

Dax Raad on building OpenCode, on the Pragmatic Engineer podcast.

Hillel Wayne pushes back on the theory that machine-written code will finally mainstream formal verification, using TLA+ war stories like Amazon's 35-step distributed bug to show where specs pay off and where property-based testing is the realistic ceiling.

Impressions from visiting the offices of OpenAI, Anthropic and Cursor in San Francisco, and what the engineers there say about how they actually work day to day.

A conversation with the OpenAI engineer who helped build Codex on why the CLI runs on Rust, why OpenAI open-sourced it, and how the harness and models underneath it have changed since release.

Gergely Orosz examines why Shopify reversed its React Native bet for native mobile development, arguing that stronger AI coding agents changed the tradeoff. He covers the 2020 rationale, the shift, and what it means for the wider mobile ecosystem.

Gergely Orosz covers DHH's Rails World keynote declaring hand-written code over at 37signals as it shifts to agent-generated code, alongside Amazon and Meta's engineering hiring struggles and the likely decline of manual code review.

Charity Majors on the Pragmatic Engineer podcast, arguing against continued skepticism about using AI in software development.

Spending $165,000 to migrate Bun from Zig to Rust looks expensive until you note it compressed a one-to-two-year migration into 11 days, which changes what migrations are worth attempting at all.

Gergely Orosz interviews Microsoft's Windows team about native agent-identity primitives and a local MCP server registry planned for Windows, and examines whether developers who left the OS will return.

The Pragmatic Engineer interviews Matt Pocock, creator of the popular 'grill-me' Claude skill and the AI Hero course, on why fundamentals still matter for engineers leaning on coding agents and how he builds skills people actually keep using.

Kent Beck revisits TDD, Extreme Programming and the Agile Manifesto for the agent era, with a sharper claim underneath the history: teams now pile up generated code much faster than they build trust in it, and the practices have to close that gap.

The Pragmatic Engineer interviews Cockroach Labs CTO Peter Mattis on how a veteran who shipped database-grade code pre-AI now feels more productive with AI tooling while keeping quality.

An editorial article from The Pragmatic Engineer exploring how AI coding tools are dramatically increasing the volume of code developers generate, and why deliberately slowing down can improve quality, reliability, and reduce tech debt.

A keynote argument that slowing down is how teams speed up, revisited after six months in which almost everything about the tooling changed.

A deep-dive report on Anthropic's internal engineering practices, based on interviews with four engineers, showing how AI has reshaped work: a six-month project to build Claude Managed Agents (a production agent harness) required upfront architecture and PRDs much like pre-AI planning, while Bun creator Jarred Sumner used 64 parallel AI agents and $165K in tokens to complete a 500K+ line Zig-to-Rust rewrite in 11 days instead of an estimated year. It also documents broader shifts at the AI lab, including near-unlimited token usage, faster prototyping, and verification now taking longer than implementation.

Gergely Orosz catalogs how top engineering orgs, including OpenAI and Anthropic, are adapting code review to AI-generated PR volume: reviewing the AI review, blast-radius triage to skip low-risk review, and reviewing plans/schemas instead of full diffs.

Gergely Orosz on what has been happening inside Meta's engineering organisation, and why the changes look self-inflicted from the outside.