SECOND-HALF MAN

The Dispatch #005  |  March 22, 2026

Weekly Intelligence for Men Navigating Disruption

Enter Agents. The Future Workforce.

Everyone is screaming that AI is coming for your job. Nobody is explaining how.

THE SIGNAL

You have heard the warning a hundred times. AI is coming for your job. AI will replace you. Learn AI or get left behind.

You nodded. You downloaded ChatGPT. You asked it to write an email, maybe summarize a document. You thought: okay, I get it. I am using AI now.

Then last week, Jack Dorsey sent a memo to all 12,000 employees at Block, the company behind Square and Cash App. The message: we will not hire for any role that AI can fill. Managers were told to plan their teams as if software agents were already on the payroll.

Not as tools. As workers.

Dorsey was not talking about the thing you used to rewrite your email. He was talking about something most people have never had explained to them clearly. Something already deployed inside 79% of enterprise companies. Something that Deloitte is now calling a “silicon-based workforce.”

He was talking about agents. And the distance between that word and what most men think AI means is about to become the most expensive misunderstanding in the economy.

THE PATTERN

There are three layers to AI. Most people know the first. The world has moved to the third. Here is what each one actually is.

Layer One: The Chatbot. You type a question. The machine answers. You type another question. It answers again. It is a conversation. Sometimes useful. Sometimes impressive. But it waits for you. It does nothing until you ask. This is where roughly 300 million people interact with AI today. It is a very fast reference desk.

Layer Two: The Copilot. The AI sits inside your workflow. It suggests the next line while you write. It fills your spreadsheet while you set up columns. It drafts while you outline. Faster than working alone, but you are still doing the work. You are still flying the plane. The copilot assists.

Layer Three: The Agent. This is where everything changes. An agent does not answer questions. It does not assist. You give it an outcome, and it figures out how to produce it. It breaks the goal into tasks. It decides which tools to use. It executes across systems. It handles exceptions. It reports back when the work is done.

Not a search engine. Not an assistant. A worker that operates without you in the loop.

The scale. Gartner projects 40% of all enterprise applications will embed task-specific agents by year-end. Salesforce replaced 4,000 customer support positions with agents. Block is restructuring 12,000 roles around them. A recent survey found 37% of companies expect to have replaced jobs with AI by the end of 2026. These are not pilot programs. These are operating decisions already made.

THE TRAP

Here is what nobody in the AI conversation is telling you.

The people screaming “AI is coming for your job” are also stuck at Layer One. The career coaches. The LinkedIn influencers. The corporate trainers running “AI readiness” workshops. Their advice is: learn to write better prompts. Ask better questions. Get more out of the chatbot.

They are teaching you to be a better passenger while the economy is looking for pilots.

The entire mainstream AI conversation, the headlines, the TED talks, the five-step LinkedIn posts, is oriented around asking. How to ask AI the right question. How to phrase the right prompt. How to get the best answer.

Agents do not take questions. They take orders. The difference is not semantic. It is structural. And it is the difference between a man who uses AI and a man who is being replaced by it.

Here is what that difference looks like.

Asking:

Commanding:

Read those again.

Same man. Same technology. Completely different result. The first gives you a tool you operate. The second gives you a worker that operates while you are in another meeting, on a job site, or sitting down to dinner with your family.

One is a question. The other is a command. And that gap is where the entire future of work lives right now. The man who learns to write the command will never worry about AI taking his job. He is the one deploying it.

THE CODE

Writing that command is not a technical skill. It is an operational one. And it favors the man who has spent decades doing the work over the man who just learned to talk about it.

Think about what the command requires. You have to know the process. You have to know what triggers the task, what information is needed, what the steps are, what done looks like, and what exceptions to watch for. That is not programming. That is management. It is operations. It is the accumulated understanding of how work actually gets done that you have been building for thirty years.

A 25-year-old can code an agent. But can he tell it what to do in your business? Can he define the outcome for a client follow-up sequence that actually converts? Can he set the boundary conditions for a procurement workflow that does not blow up the vendor relationship? Can he specify what “done” looks like for a job estimate that accounts for the fifteen things that go wrong on site?

He cannot. You can. Because you have done the work.

This is where Element 13 of The Sovereign’s Code, Force Multiplication, becomes operational.

The sovereign does not work harder. He multiplies the force of every hour through systems, agents, and leverage that extend his reach without exhausting his body.

Force Multiplication is not about learning to code. It is about translating what you already know into language clear enough that a system can execute it. The knowledge is yours. The judgment is yours. The agent just gives it leverage. One man, one command, work done across systems while he moves on to the next thing that actually needs his attention.

The companies deploying agents right now are learning this the hard way. A procurement team sets up an agent to auto-reorder supplies when inventory runs low. Efficient. Except nobody told the agent about the relationship with the steel supplier you have used for twelve years. The agent sees a price discrepancy, fires off a demand for a discount at 3am on a Saturday, and a vendor relationship that took a decade to build is damaged before Monday morning. The agent did exactly what it was told. It just was not told by someone who understood what mattered.

That is the pattern everywhere agents are failing. Not bad technology. Missing judgment. The agent needs the man who knows which outcomes matter, which relationships cannot be automated, and which exceptions require a phone call instead of an email. That man is not obsolete. He is the missing piece.

ONE TRUTH

Stop asking AI questions. Start telling it what done looks like. That is the only shift that matters.

YOUR MOVE THIS WEEK

Write Your First Agent Command

Pick one task you do every week. Something repetitive, structured, and time-consuming. Client follow-ups. Invoice processing. Report generation. Scheduling. The task you keep doing because it has to get done, not because it requires your judgment.

Now write the command. Not a question. A set of instructions clear enough that someone who has never done this task could execute it without calling you. Cover five things:

What triggers the task. What information is needed. What steps happen in what order. What the finished output looks like. What exceptions require a human decision.

Here is what that looks like for something as simple as processing an incoming invoice:

Five lines. No code. Just a man who knows how his invoicing works describing it clearly enough that a system could run it.

Now do yours. Pick your task and write the command.

Then take it and hand it to ChatGPT or Claude. Ask one question: “Could an autonomous agent follow these instructions to completion without asking me anything? If not, what is missing?”

It will tell you where your instructions are vague. Where you assumed knowledge a system would not have. Where the done-state is not specific enough. Fix the gaps. Tighten the language. Hand it back.

When the AI says yes, this is executable, you will be holding something most people do not have: a document that could run a piece of your business without you. Not because the technology is magic. Because you described the outcome with enough precision that a system could deliver it.

You did not learn to code. You did not become technical. You learned to command, not prompt. That is Force Multiplication. And you just did it at your kitchen table.

The conversation about AI and work has been dominated by two voices. Evangelists who promise everything will be fine. Alarmists who say everything is ending. Neither is explaining the actual mechanism. Neither is showing you what is under the hood.

That is what this platform exists to do. Not hype. Not panic. Clarity. So you can make your own decisions from a position of understanding rather than fear.

The agents are here. They are not going anywhere. The only question is whether you will be the man who commands them.

Forward,

Russ Borden

Founder, Second-Half Man

Was this forwarded to you? Subscribe here.

About 2HM  |  Discover Your AI Exposure  |  The MANifesto  |  ASCENT

PwC AI Agents Survey, May 2025 • Gartner Enterprise Software Forecast, 2026 • Deloitte State of AI in the Enterprise, 2026 • Resume.org AI Workforce Survey, September 2025 • Block Inc. Internal Memo, March 2026 • KPMG Q4 AI Pulse Survey, January 2026