Open Textbook

Productivity

Managers and AI: The Most Frequent Users – But Not for Managing

14 min read

Of all the professions that most frequently open Claude, managers came in first. In Anthropic’s survey, they made up 23% of respondents – while accounting for roughly 7% of U.S. employment. Managers are overrepresented among AI users by a factor of three. Now the second number: management tasks make up just about 4% of all sessions. The people who manage are using AI for everything except managing.

Behind these two numbers lies the most precise description of how managers actually work with AI. And why the fear of “it will take my job” works differently in this profession than you might expect.

Read more
Managers and AI: The Most Frequent Users – But Not for Managing
5 Layers of an AI Agent: What Every Manager Needs to Know
19 min

5 Layers of an AI Agent: What Every Manager Needs to Know

“Most people still think an AI agent is just ChatGPT with a good prompt.” That is the opening line of Sunil Ramlochan’s article “The AI Agent Stack Is Not a Prompt. It’s a Production System”, and the author calls this belief “a comforting myth.” The useful truth, he argues, is different: a real agent is closer to a small operating system for getting work done. It has a brain, hands, memory, rules, logs, recovery plans, and someone accountable when the agent does the wrong thing.

The article’s thesis fits in a single line: an agent is an entire stack. Reliability comes from the architecture around it, while the model itself – or a clever prompt – is just one ingredient. The picture is an engineering one, so let us approach it from the other side: what in this stack actually concerns the manager who does not write code but decides whether to put an agent to work.

AI Doesn't Make You Dumber. It's About How You Use It
9 min

AI Doesn't Make You Dumber. It's About How You Use It

A year and a half ago, I wrote a note on my personal blog about something I was noticing in my colleagues’ work and in my own: the more you trust AI, the less often you ask yourself “is this actually right?” I was drawing on a Microsoft study at the time – it showed that trust in AI suppresses critical evaluation of the answers it produces. The argument felt strong to me, but it had an obvious flaw: correlation, not causation.

In February 2026, Anthropic researchers Judy Shen and Alex Tamkin published an experiment that closed that gap. Randomized control. Concrete data. And a conclusion that, I think, most people who’ve read about it have misunderstood.

Because this isn’t a story about AI making us dumber. It’s a story about how exactly we use it.

AI Doesn't Save Time – It Compresses It: 8 Months of Observations
11 min

AI Doesn't Save Time – It Compresses It: 8 Months of Observations

Companies are worried about getting employees to use AI. The promise is seductive: AI will handle the drudgery – document drafts, information summarisation, code debugging – freeing up time for higher-value work.

But are companies ready for what happens if they actually succeed?

Researchers at Stanford conducted an 8-month observational study of roughly 200 employees at an American tech company that had rolled out generative AI. The company didn’t mandate AI use – it simply provided corporate subscriptions to commercial tools. Employees decided for themselves whether to adopt them.

The result was paradoxical. AI didn’t reduce work. It intensified it. Workers moved faster, took on more tasks, spread their work across more hours in the day – often without any explicit external pressure. AI made “doing more” possible, accessible, and in many cases internally rewarding.

Strikingly, the same pattern shows up in other research. Microsoft found that 62% of product managers use Gen AI daily, yet while 81% say AI saves time, 56% deny that effort has decreased. A paradox? No – a pattern.

6,600 Commits in a Month: Workflow Lessons from the Creator of OpenClaw
16 min

6,600 Commits in a Month: Workflow Lessons from the Creator of OpenClaw

One developer. 6,600 commits. One month.

More than most teams ship in a quarter. More than many startups produce in half a year. This is not a marketing metric – it is the real-world productivity of Peter Steinberger, creator of OpenClaw (formerly known as clawdbot), one of the most viral AI projects of January 2026.

Steinberger describes the project plainly: “It’s not a company – it’s one guy sitting at home enjoying the process.” After a successful exit from PSPDFKit, he could have taken a break. Instead, he is building an AI assistant that manages his calendar, sends emails, and checks him in for flights. “AI that actually gets things done” – that is how he articulates the project’s mission.

How can one person work like an entire company? What skills are critical when working with AI agents? Why does experience managing a team of 70+ people turn out to be the key to AI-driven productivity? And how does an engineer’s focus shift – from writing code to designing architecture?

Let us examine the actionable lessons from Peter Steinberger’s workflow – applicable to any AI-assisted project, even if you never install OpenClaw itself.

The Hidden AI Tax: Why 40% of Time Savings Disappear
10 min

The Hidden AI Tax: Why 40% of Time Savings Disappear

87% of employees use AI at work. 77% say they’ve become more productive. Sounds like a digital transformation success story. But ask the uncomfortable question: if AI boosts productivity so effectively, why do only 14% actually see net gains? Where do the other 86% go?

Workday conducted a global study with 3,200 respondents – and found that behind the impressive numbers lies an inconvenient reality. Nearly 40% of time saved with AI is spent fixing its mistakes. This isn’t a bug in specific tools, nor a problem of “incorrect usage.” It’s a systemic failure in how organisations deploy artificial intelligence.

At mysummit.school, we’ve previously covered Dejan.AI (AI as a junior intern), Gallup (usage doubling in a year), and MIT (the gap between leaders and laggards). Workday adds a new angle: the measurable cost of low-quality AI output. And that cost turns out to be higher than many are willing to admit.