The agent writes the code. The engineer owns the outcome

Pini Shvartsman — Technology Leader · AI Transformation & Autonomous Engineering

I redesign engineering organizations around AI-powered execution and human accountability — and I build the autonomous systems that make it real.

Scroll

Execution moved.

I lead engineering organizations through this shift. Here’s what actually changed.

“Writing code was never the bottleneck.”
The full argument →

“AI agents aren’t tools you use. They’re workers you manage.”
Agents on the org chart →

The development sequence: agents take every step; the engineer directs, judges, and owns from above

The development sequence, before and after agents. Every execution step moves to agents; direction, judgment, and ownership of the outcome stay with the engineer — one seat above the whole flow.
Step	Executor before	Executor now
Plan	Human	Agent
Scaffold	Human	Agent
Write	Human	Agent
Validate	Human	Agent
Ship	Human	Agent
Operate	Human	Agent

The sequence changed. The responsibility didn’t →

466 million lines of government code audited by agents in 20 hours — one Canadian province’s entire codebase. What Alberta’s weekend audit means →

My vocabulary

Agents on the org chart

Next · 02 The Cost ↓

Cheap to produce. Not cheap to own.

The bill lands on whoever owns production. That seat is mine.

“The cost doesn’t show up at generation time. It shows up at ownership time.”
The economics →

“Engineering discipline became a career risk. That’s the shift.”
What actually broke →

The dashboard stays green

“Throughput is legible. Judgment is invisible.”

Why the seniors are losing →

My vocabulary

Cheap to produce, not cheap to own

Next · 03 Judgment ↓

The scarce asset.

I came up through every rung — support, NOC, full-stack, architecture. The engineers I lead won’t climb that ladder, so I’m designing the new one.

“The first group learned with AI as a teacher; the second learned with AI as a tool.”
Why I’m worried →

Marcus

247 commits last month. 23 features shipped. Velocity charts trending up. Then one question — “why did you structure the caching layer this way?” — and silence.

Sarah

Same company. Nearly identical velocity. Asked the same question, she walked through the trade-offs, the two AI suggestions she rejected, and the monitoring she added for the failure modes she predicted.

“The agent has all the answers. You are the only one who knows which questions to ask.”
The hopeful version →

My vocabulary

Operators vs. AI-augmented engineers Kolboynik Scaffolding vs. substitution

Next · 04 Overwatch ↓

Supervision is architecture.

Agentic Overwatch is the discipline I named — and the one I run.

“You cannot govern a workforce that runs flat out, around the clock, with a team that logs off at 5 PM.”
Agentic Overwatch →

The overwatch tier model

T1 · Detect — agents

Watch everything, all the time. Surface what matters.

T2 · Remediate — agents

Reproduce, propose, fix — inside the guardrails.

T3 · Judge — humans

Authorize what ships. Own what happens next.

Your presence lights the human seat Tap the third lane to take the seat

The overwatch tier model
Tier	Who runs it	What it does
T1 Detect	Agents	Watch everything, all the time; surface what matters
T2 Remediate	Agents	Reproduce, propose, fix inside the guardrails
T3 Judge	Humans	Authorize what ships; own what happens next

“The diff is the claim. The evidence is the proof. The human is the judge, not the fact-checker.”
The evidence gate →

“They’re debating tools while we’re deploying workers.”
How my org runs →

Not a proposal — the operating model behind these notes, running in production.

Take the Tier 3 seat → Six agent changes arrive with their evidence. You rule on each one.

My vocabulary

Agentic Overwatch The Agent Operations Center The Overwatch Maturity Model The evidence gate

Next · 05 The Operator ↓

Written from the seat.

Technology leader, hands still on the system. Everything above runs in production — I write down what holds.

The builder

Taking things apart since age eight. IRC channels, scripts, communities that didn’t care how old you were as long as you kept up.

Employee #5

No DevOps, no infrastructure, barely a product. Built all of it — the CI/CD, the platform, the offshore team — and grew from five people to leading engineering at global scale.

The through-line

“I just kept saying yes to whatever scared me most.” The current answer: redesigning how an engineering org runs when agents do the executing.

“We were the failsafe, and the failsafe doesn’t get to sleep through the incident.”
A year inside a NOC →

Model Context Protocol: The Missing Connection Between AI and Your Real Work The Context Problem: Why AI Can't Remember You Across Apps (And Why That's Not an Accident) The Agentic Commerce Protocol: We Just Gave Every LLM the Ability to Buy Things The Magic Behind AI IDEs: How Cursor, Windsurf, and Friends Actually Work Developer Work Did Not Change. The Sequence Did. GitHub's Double CLI Release: How Two AI Tools Are Reshaping Development Workflows So, You've Built a RAG. Now Let's Make It Not Suck. Hiring Developers in the Age of AI: What Actually Matters Now RAG for Developers: A No-BS Introduction The Uneven Reality of AI Adoption — What Anthropic's New Report Tells Us The Context Engine: What Comes After We've Solved Code Generation I'm Pro-AI. That's Exactly Why I'm Worried About Our Next Senior Engineers When CI/CD Speaks Human: A Friendly Nudge to DevOps (and Developers) From "Toys" to "Tools": The Missing Layer Developers Actually Need

Amazon S3 Vectors: When Storage Learns to Think Your AI Browser Can Be Hijacked by a Single Webpage. Here's How Companies Are Fighting Back. Have You Seen All These OpenAI Blueprints? What the Heck Are They Doing, and Why Is (or Isn't) Your Country In? Krakow Offsite: Real Connections, Real Momentum When Nvidia's CEO Says 100% of Engineers Use Cursor, He's Not Exaggerating When AI Writes 90% of Your Code, What Are You Actually Doing? Securing Intelligence: The Complete AI Security Series [Video] AI Security Isn't a Tool Problem, It's a Culture Problem Securing the AI Supply Chain: The Threat Nobody's Talking About Building AI Systems That Don't Break Under Attack Prompt Injection 2.0: The New Frontier of AI Attacks AI's Dual Edge: When to Disrupt and When to Compound Grokipedia and the New Era: When Building a Wikipedia Becomes Trivially Easy Two Weeks with Gemini in Chrome: The Browser That Actually Gets It Google Jules: Always on My Radar, But Never Quite the Star Ship Faster Without Breaking Things: DORA 2025 in Real Life Build Your First AI Agent This Week: A Practical Guide AI Agents for Real Productivity: What Works in 2025 What's Holding You Back from Succeeding in the AI Era?

Web Bot Auth: Giving Bots a Crypto ID Card in a World of Fakes CLI Agent Orchestrator: When One AI Agent Isn't Enough AI Wrapper Companies: Is This Real or Just API Theater?

Org Charts for AI Agents: Mapping Your Human and AI Workforce

Cisco Built an LLM Security Leaderboard. You Should Care Even If You Don't Use Cisco. OpenAI Killed Sora. That Tells You Everything About Where AI Is Actually Heading. Claude Can Now Use Your Computer. Here's What That Actually Means. DeerFlow 2.0: ByteDance Just Open-Sourced What Most Companies Are Trying to Build Internally Zuckerberg Is Building an AI CEO Assistant. The Rest of Us Should Have Started Already. Cursor Automations: Your AI Just Stopped Waiting for Permission AI Is Now Reviewing AI's Code. That Should Make You Think. Glassworm Is Back. Your Code Review Won't Catch It. WordPress Just Let AI Agents Publish to 43% of the Web. Now What? AI Didn't Replace Software Engineering. It Made Bad Engineering Easier to Ship. SaaS Is Dead. We Just Haven't Stopped Paying for It Yet. Google Stitch Just Made UI Design a Developer Skill OpenClaw Is Not a Chatbot. It's a Personal Agent Gateway. Open WebUI Isn't a ChatGPT Clone. It's AI Infrastructure. NotebookLM Is Not a Chatbot. It's a Research Workbench. Your AI Agents Are Flying Blind. Here's How to Fix That.

The One-Man Show Company. Don't Let the Monkeys Touch Production. 100 Days to the EU AI Act Deadline. Your Engineering Team Hasn't Started. The Vibe Coding Backlash Is Right. Seniors Are Losing the Argument Anyway. The End of Courses: Learn From AI Like a Toddler, Or Become Obsolete I Don't Put All My Eggs in One Basket. Anthropic Is Making That Hard. The IDE Is Becoming Mission Control Your AI Stack Is Rented Until You Can Run Part of It Yourself AI Makes Code Cheap to Produce. Not Cheap to Own. 'I Only Built a Small Script for Myself.' That Might Be the Most Dangerous Sentence in Your Company. The Claude Code Leak Isn't Dramatic. That's the Point.

Agentic Overwatch: Why Your Next Dev Team Will Look Like a NASA Control Room

Everyone Trained Their Engineers on AI. The Gap Didn't Move. Make the Cheap Path the Default. Make the Expensive Path Prove It. Alberta Scanned 466 Million Lines of Code in 20 Hours. The Architecture Is the Story. Stop Reviewing Code. Start Reviewing Evidence.

The record — every tick a published note. The quiet was deliberate.

How I work

Embed

Inside your team — your codebase, your incidents, your constraints.

Ship one

The first system goes to production, not to a slide.

Prove it

Measured against the work it replaced: real time, real budget.

Make it repeatable

The version that works becomes a capability your org owns.

Enter the archive · 06 Field Notes ↓

The words I had to coin.

New work needed new names. Each one is argued for in a note — take them, use them, tell me where they’re wrong.

Agentic Overwatch

Supervising an agent workforce as a designed, tiered, shift-based discipline — not an afterthought.

The Agent Operations Center

The control room for your fleet: agents run detection and remediation, humans hold judgment.

The Overwatch Maturity Model

Five levels from blind trust (L0) to an orchestrated, supervised fleet (L4).

The evidence gate

Machines produce the proof — repro, tests, validation. The human judges the case, not the diff.

Cheap to produce, not cheap to own

Generation is nearly free. Ownership — comprehension, maintenance, incidents — is where the cost lands.

Agents on the org chart

Agents doing real work are workers: a named owner, boundaries, onboarding, KPIs.

Operators vs. AI-augmented engineers

One group learned with AI as a teacher, the other as a tool. Only one compounds.

Scaffolding vs. substitution

Every prompt is one or the other: building your understanding, or replacing it.

The three buckets

Every task sorted before automation: AI alone, AI with human approval, or forbidden.

Securing Intelligence

My series on defending AI systems: prompt injection, supply chain, and the culture underneath.

Field notes

Everything I’ve published from the seat — 68 field notes on running engineering when agents do the executing and I stay accountable for the outcome.

Latest

Start here

The Shift

Execution moved to agents. The work — and who owns it — did not.

The Cost

Cheap to produce is not cheap to own. The bill arrives downstream.

Judgment

The scarce asset — and how engineers grow it when AI does the typing.

Overwatch

Supervision as architecture: control rooms, evidence gates, agents on the org chart.

The Perimeter

Agents are an insider workforce. Secure them like one.

The Playbook

Hands-on: agents, RAG, MCP, and the tooling underneath — no hype.

Dispatches

News through an operator's lens — what launches mean for how orgs run.

By date 68 notes · newest first

202631

202537

Not advice from a deck. Advice from the seat — backed by a system in production.

I still lead this work every day: building AI systems that investigate bugs, review code, and run operational workflows. The engagements I take come from that seat — the org-level practices that actually hold up, not tool training. A few at a time.

See how we can work together →