You’re reading The (100th!) 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.
In the Background
Your employees probably don’t know what they’re working alongside. Agentic AI is already routing cases, scheduling interviews, and drafting renewal-risk reports under people’s names. In fact, 82% of enterprises found unknown agents running in their own IT environments last year. Disclosure has badly lagged deployment. David Rice’s piece lands on an argument executives can’t wave away: every governance framework – NIST, the EU AI Act, your own AI policy – assumes a human stays in the loop to catch errors. And that human is almost always a frontline worker, close enough to the output to know when something’s off. Hide the agent, and that oversight collapses. One Outbuild CSM learned this walking into a call where the customer referenced a meeting the agent had booked in her name – one she’d never set up. Trust took weeks to rebuild. Fixing this is not easy, but we can start with naming what agentic systems do in plain language (drop the word “automation”), make agent actions 100% visible, and require a named human approver above a threshold of confidence. Your staff aren’t just the audience for disclosure – they’re the error-detection layer the whole system depends on. Keep them blind and governance is only theoretical.
Editors note: we have big ideas on how to fix this!
— People Managing People, 8m, #ai, #management, #trust
Flat Out
While the industry torched headcount in the name of going “AI-native,” Bloomberg quietly doubled its engineering org from 5,000 to 10,000. The signal to employees: we won’t discard you when the market turns. In what many calla retention strategy of key people, Bloomberg’s engineering side is structured with exactly two titles: engineer and people leader. No VP turf wars, no level-number status games, no approval chains. Michael Bloomberg built it that way on purpose, believing titles breed bureaucracy and distract from the work. What you own matters more than where you sit on the org chart, and engineers own products soup to nuts – from idea to production to how clients access it. The catch: flatness that fuels junior and mid-career growth can stall senior progression, and some who left say so. But the trade-off is deliberate. When everyone else optimizes for leverage and levels, Bloomberg optimizes for ownership and tenure – rare in an industry defined by churn nowadays.
— LeadDev, 6m, #engineering, #leadership, #retention
Can’t Fake It
A decade ago, Google’s Project Aristotle tracked 180 teams across every variable it could measure – tenure, intelligence, friendship, personality mix, even whether teammates ate lunch together – trying to explain why some teams hummed while others, stacked with equally smart people, sputtered. Nothing fit until an analyst stumbled onto Amy Edmondson’s 1999 paper on psychological safety: the shared belief that you can speak up, admit mistakes, and challenge ideas without fear. That was the answer, “far and away.” Rebecca Hinds revisits it to ask the obvious 2026 question: with AI pressure-testing every assumption about how work gets done, has it changed the recipe for great leadership? Workhuman CEO Eric Mosley’s answer: “Not yet.” The old rules still apply, which is exactly why they’re getting harder to fake. As AI absorbs the optimized, automatable layer of work, what’s left exposed is the human part: trust, candor, the safety to say “this is broken.” AI can polish output, but it can’t manufacture the conditions that make a team members tell the truth about the output.
— Inc., 6m, #ai, #leadership, #culture
Vibe Slop
The two engineers who built the core of OpenClaw’s AI agent have a warning for everyone betting their codebase on AI: you’re trading near-term speed for long-term wreckage. They call it “vibe slop” – the bastard child of vibe coding and AI slop – and it’s the buggy, brittle, insecure code that piles up when you swap the hard work of designing and testing a system for a quick prompt. The original pitch was seductive: make senior engineers so productive you can stop hiring juniors. The reality is buggy software, outages, security holes, and mounting technical debt – plus a talent pipeline quietly drying up. Pichai says 75% of new code at Google is AI-generated; the skeptics counter that AI is far better at spinning up new code than safely evolving the massive, tacit-knowledge-laden systems real companies actually run. Even OpenAI keeps humans ultimately responsible for anything touching millions of users. The throughline for anyone leading a team: AI is leverage, not a license to skip the thinking.
— WSJ, 8m, #ai, #engineering, #technical-debt
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 treat blockers as one-offs: someone flags it, someone clears it, everyone moves on. But the same obstacles keep resurfacing – design feedback that’s always late, the dependency that’s never ready, the approval that always stalls. This Echo analyzes every blocker filed over the past 30 days and surfaces the top 5 recurring themes, complete with occurrence totals and real examples. Instead of perpetually reacting to individual fires, you get the pattern – so you can fix the root cause once instead of the symptom forever. Runs monthly, lands ready for your retro.
The lightweight 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.
Problem is, status quo approaches to running those loops are an incomplete, inconsistent, and inefficient tangle of meetings, emails, chat threads, dashboards, and manual toil.
Steady is the teamwork OS that runs both loops for you. Purpose-built agents continuously distill updates and activity into personalized intelligence that keeps everyone aligned and informed automatically.
The outcome: high-performing teams that deliver better work, 3X faster.
Learn more at runsteady.com.