You’re reading The Steady Beat, a weekly pulse of must-reads for anyone orchestrating teams, people, and agents across the modern digital workplace – whether you’re managing sprints, driving roadmaps, leading departments, or just making sure the right work gets done. Curated by the team at Steady.
Caretakers
Alex Pliutau thinks AI made software engineering boring, and he lists the reasons why by what’s different. First, learning. The old path from problem to solution ran through Stack Overflow threads and books like Designing Data-Intensive Applications, and we learned a ton along the way. AI skips all of that. Second, the codebase. Agents add code instead of removing it, reviews get skipped or handed to another AI, and the result is slow builds, slow tests, and features that contradict each other. “It’s like a [bad] drug for the codebase,” he writes. Third, noise. AI-written emails and docs make the reader do the decoding, and managers who expect faster delivery push teams toward more of both. He quotes an old reviewer’s rule seems apt for the moment: give someone a 10-line pull request and they’ll find 10 issues, give them 500 lines and they’ll say LGTM. The conclusion here is somewhat bleak: engineers are inevitably becoming caretakers of code nobody fully understands.
— Package Main, 4m, #engineering, #ai, #quality
Find the constraint
Tarik Guney manages a team that keeps critical Microsoft Teams services running, and he focuses on finding where that team gets stuck. Usually, it’s communication that’s missing, late, or garbled. Gregor Ojstersek profiles him in a case study from his book The Multiplier Mindset. Guney’s goal is to raise each person’s throughput without adding hours or stress. He’s wary of the reflex to add a process for every problem, because it makes work more complex and keeps the gains small. He watches leading indicators most closely. If those improve and outcomes don’t, the team is optimizing the wrong thing or missing a constraint. AI sits on top of the workflows his team already has: meeting summaries, catching up on long threads, drafting plans and pull requests, sending updates. He still treats it as an aid to judgment. “AI does not replace judgment,” he says. “It increases the speed and quality of how that judgment is applied.”
— Engineering Leadership, 6m, #management, #leadership, #coordination
Help wanted
Lenny Rachitsky’s late-2026 job market report has one clear winner: product managers. Open PM roles are at a nearly four-year high, up 75% from the March 2023 low, and PM is the only core role whose share of hiring grew. Engineering demand has flattened. U.S. companies hired about 113,000 software engineers in the year through April, under half the 2022 peak. Design and user research took the hardest hit. Open design roles are at a record low, and U.S. companies hired 66% fewer user researchers than in 2022. Rachitsky’s theory is that faster cycles leave less time for research, and AI tools let anyone run it. “LLM” is now the most-mentioned skill in PM job descriptions, and “Claude” passed CSS in designer listings. The jobs are also moving to the Bay Area, which holds a record 15% of open tech roles, while only 14% offer remote work.
— Lenny’s Newsletter, 7m, #hiring, #careers, #ai
Syntax loyalty
Marcos Lobo’s advice to junior engineers works for everyone else too: stop collecting languages and start solving problems. “Companies don’t hire you for your loyalty to a syntax; they hire you to build solutions.” He treats AI as the latest layer of abstraction, the same kind of jump as assembly to Python. Let agents write the repetitive code. Spend your time on architecture, debugging, and trade-offs, and on how the systems fit together and what business problem they solve. He also says to take on work in a language or framework you’ve never used, because untangling the confusion is where you grow. A commenter adds a practical trick: ask an agent to explain the new language by comparing it to one you know.
— The Optimist Engineer, 3m, #careers, #engineering, #ai
Bottlenecked on judgment
Christoph Nakazawa says coding agents now write better code than he would by hand, in minutes instead of weeks. His GitHub numbers back it up: 770 commits in the past 30 days, more than double the same stretch a year earlier. Agents wrote nearly all of it, including features for Vite+ in a language he doesn’t even know (Rust). Writing code simply stopped being his constraint. “I was bottlenecked on writing code, now I’m bottlenecked on exercising judgement.” So he wrote down the development values that still matter. Owners with deep context get faster, and people without it generate noise. Teams need taste about what’s worth building at all. Lint rules and fast tests decide whether an agent finishes in one minute or 60. Context belongs in the repo, because agents behave like new hires. Overall, he expects engineering managers to get more technical, and small teams of two or three with clear boundaries to do the best work.
— Christoph Nakazawa, 14m, #engineering, #ai, #management
Echo of the week
Echoes are AI agents in Steady that automatically gather and deliver work context to teams on a schedule – answering recurring questions about progress, capacity, and coordination so you stop burning hours assembling the same information manually.
Blocker trends. Most teams handle blockers one at a time. Someone’s stuck, someone unsticks them, and the same kind of blocker shows up again two weeks later. This Echo looks at every blocker your team filed in the last 30 days and pulls out the top five recurring themes, each with a count and up to five examples. It runs once a month. Engineering managers get a reason to fix the process instead of the symptom, retro facilitators get real numbers instead of vibes, and ops leads can see which handoff or dependency keeps holding everyone up.
The human-agent teamwork OS
Teams rely on two coordination loops to function: a big-picture loop connecting plans to progress, and a ground-level loop keeping teammates in sync.
Running those loops was already a marathon of meetings, chat threads, dashboards, and manual toil. Pile on flatter orgs, exponential output, and AI agents shipping 24/7 — the old way can’t keep up.
Steady removes the coordination bottleneck by running both loops for you. Working in the background, Steady distills updates and activity into targeted context for everyone on the team — human and agent alike. Full visibility, tight alignment, zero overhead.
The outcome: high-performing teams that deliver the right work, not just more of it.
Learn more at runsteady.com.