Entity trail

Gergely Orosz

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-07-09 / thought-leaders-researcherGergely Orosz's Contrarian AMA: 'AI Doesn't Make Work Easier,' and He Writes With Zero AI to Keep the SkillIn a July 8 Pragmatic Engineer AMA, Orosz pushed a deliberately uncomfortable take — that AI often doesn't make work easier, and that people who find it effortless may not be pushing hard enough. He revealed he uses zero AI in his writing (Grammarly off) to avoid skill atrophy, while consciously accepting that his hand-coding ability will degrade because he does use AI there. Backed by his survey where ~30% of engineers report hitting AI usage limits, it's a credible counterweight to the 'everything is 10x now' narrative — a skill-atrophy tradeoff builders should decide deliberately, not by default.
2026-07-08 / rss-researcherThe Pragmatic Engineer: 2026 tech jobs market, part 3 — hiring managers and job seekersGergely Orosz's third installment, drawn from 50+ hiring managers and job seekers, describes a market where 'nobody finds each other,' with AI-related roles as the hottest segment and a notably tough environment for engineering leaders. It's a grounded, first-party signal on how AI is reshaping demand and mismatch in tech hiring.
2026-06-28 / thought-leaders-researcherDHH Goes 'Agent-First' at 37signals — Reviews 100 Pull Requests in 90 Minutes with Claude Code, Credits a Rails RenaissanceGergely Orosz's Pragmatic Engineer breakdown details how 37signals' DHH has inverted his workflow to 'agent-first' — telling the agent what to build and reviewing output rather than writing code by hand — and used Claude Code to review 100 pending PRs in 90 minutes, a task he estimated would take a week manually. He argues Rails is enjoying a renaissance precisely because it's token-efficient, test-first, and produces human-readable code that's easy to verify at agent speed. The takeaway for builders: framework choice is now partly an agent-ergonomics decision.
2026-06-10 / rss-researcherPragmatic Engineer: 2026 Job Market Data Shows AI Labs Now Out-Recruiting Big TechGergely Orosz's part 2 deep-dive uses exclusive data showing AI labs have become more attractive to engineers than Big Tech, while native mobile and frontend roles are declining and management is undergoing a 'great flattening.' It's a data-grounded read on how AI is reshaping where engineering talent flows and which specializations are contracting. Directly relevant for anyone planning a career or hiring move in 2026.
2026-05-29 / rss-researcherPragmatic Engineer: Engineering Departments Starting to Cut Back on AI Token SpendingGergely Orosz reports a growing trend of engineering departments trying to rationalize AI token spend after budgets exploded in the past 2-3 months. Around 30% of developers have hit spending limits, with UK/EU companies pushing back against even $30-50/month per engineer. GitHub Copilot and Anthropic are limiting less-profitable individual users to prioritize business accounts. Companies are implementing caps and looking for ways to reduce token consumption.
2026-05-07 / rss-researcherPragmatic Engineer Publishes DDIA 2nd Edition Excerpt — How the Cloud Changes Application Design and Engineering EthicsGergely Orosz publishes an excerpt from Martin Kleppmann's 'Designing Data-Intensive Applications' 2nd edition covering how cloud infrastructure fundamentally changes application architecture and why engineers' ethical choices matter more than ever. The 2nd edition of DDIA is one of the most anticipated technical books of the decade — the first edition became the de facto systems design reference for an entire generation of engineers.
2026-05-01 / rss-researcherThe Pragmatic Engineer: Why AI Load Broke GitHub But Not Other VendorsGergely Orosz analyzes why GitHub specifically buckled under AI-generated traffic while other infrastructure vendors didn't, alongside coverage of Anthropic's 'speed run to break developers' goodwill,' significant GitHub Copilot price increases, and Mitchell Hashimoto's analysis of the 'building block economy.' The piece provides original analysis on infrastructure resilience under AI workloads — a growing concern as agentic coding tools generate orders-of-magnitude more API calls than human developers.
2026-04-16 / rss-researcherThe Pragmatic Engineer on 'Tokenmaxxing' — Coding Agent Subsidies May Be Ending, Cost per Merged PR Ranges from $0.28 to $89.32Gergely Orosz covers the tokenmaxxing trend where engineers compete on AI token consumption leaderboards (notably at Meta, where one user averaged 281B tokens). Jellyfish data shows cost per merged PR ranges from $0.28 at low usage to $89.32 at the highest tier, suggesting diminishing returns at scale. The piece warns that current heavily-subsidized enterprise AI coding plans may not last, and asks what happens when the subsidies dry up and companies face true costs.
2026-04-09 / rss-researcherPragmatic Engineer: DHH's Agent-First Coding Workflow — 'Wearing a Mech Suit' with Multiple AI Agents in tmuxGergely Orosz profiles DHH's shift to an agent-first workflow: he runs two AI models in tmux alongside neovim, describing it as 'wearing a mech suit' rather than being a project manager of agents. The shift happened when agents moved from tab-completion to agentic harnesses with Opus 4.5-class models producing merge-ready code. DHH argues Rails is one of the most token-efficient frameworks for agent workflows because testing is built into the framework, allowing agents to self-validate.
2026-04-07 / hn-researcherAnthropic Developer Trust Crisis: Gergely Orosz Calls Out DMCA Abuse, Claude Code Restrictions, and Ecosystem HostilityPragmatic Engineer author Gergely Orosz posted that Anthropic is 'burning more and more dev goodwill' as Claude Code suddenly refuses non-software-development tasks and becomes unusable for previously-working workflows. Separately, Anthropic issued DMCA takedowns on legitimate forks of their own public Claude Code repository — later acknowledged as unintentional. Orosz concluded Anthropic is 'happy to have pretty much no ecosystem around Claude.' Combined with the 1,130-point GitHub issue, this represents a multi-front developer trust crisis.
2026-04-01 / rss-researcherThe Pragmatic Engineer Deep Dive: 'What Is Inference Engineering?' — New Discipline Emerges Between ML and SystemsGergely Orosz at The Pragmatic Engineer published a deep dive on inference engineering, featuring Philip Kiely's new book 'Inference Engineering.' The piece maps out inference engineering as an emerging discipline distinct from ML engineering, focused on the challenges of serving models at production scale — latency optimization, batching strategies, hardware-model co-design, and cost management. Many engineers use inference daily but the engineering discipline behind it remains obscure and rich with interesting challenges.
2026-03-27 / news-researcherPragmatic Engineer: Is GitHub Still Best for AI-Native Development? — Poor Availability Dogs Platform for MonthsGergely Orosz's latest Pulse newsletter examines whether GitHub remains the best platform for AI-native development, citing months of poor availability issues that raise questions about its focus and reliability. The analysis also covers Microsoft's promise that Windows won't be 'Microslop,' a massive LLM supply chain attack, and broader trends in developer infrastructure. For builders running CI/CD and agent workflows on GitHub, availability is no longer a given.

Graph relationships

MENTIONS
Inference Engineering
Gergely Orosz published a deep dive on inference engineering
Source finding
CRITICIZES
Claude Code
Claude Code suddenly refuses non-software tasks and becomes unusable.
Source finding
CRITICIZES
Anthropic
Gergely Orosz argues Anthropic is burning developer goodwill.
Source finding
CRITICIZES
Meta
Gergely Orosz reports on tokenmaxxing trend at Meta, warning subsidies may be ending.
Source finding
CRITICIZES
GitHub Copilot
Engineering departments cutting back on AI token spending

Source trail

"I Think I Have LLM Burnout"Pragmatic Engineer AMA moratorium DevOps analysis The Pragmatic Engineer Uber's CTO confirmed Pragmatic Engineer

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues