I Don't Use AI Like a Chatbot. I Use It Like a Crew.

Instantiate a temp workforce.

10:34

Why the next generation of workers won't be judged by what they can do alone — but by what they can orchestrate.

A friend called me a few weeks ago looking for job advice. He's a programmer. Smart, genuinely AI-savvy — probably more capable than 90% of people in roles adjacent to his. He'd been leaning hard on LinkedIn, putting in applications, waiting for callbacks.

I told him LinkedIn isn't a job place. He was surprised. But here's the reality: the moment you put "Open to Work" on your profile, you've already surrendered your leverage. LinkedIn is a network surface, not a hiring mechanism. If you don't have a position, you don't have leverage. That's just physics. But the conversation went somewhere more interesting. He asked how to stand out. How to walk into a room and make employers understand what he could actually do.

I told him: pick the five companies you want to work for. Research them. Then walk in and show them — not tell them — what you can bring.

And then he asked me the question that broke the whole thing open.

"Okay," he said. "What would you actually do if someone asked you to explain a complex technical issue to a room full of executives — on short notice?" I didn't hesitate.

I'd instantiate a temporary workforce.

What That Actually Means

That phrase sounds insane for about half a second. Then your brain catches up. Because that is exactly what's happening when an AI-native operator sits down with a problem. They're not "using AI." They're assembling flash labor capacity around an objective.

Here's what it looks like in practice. The moment I know the audience and the topic, I'm doing several things simultaneously. I send a message to my research agent: do a dossier on these people — their roles, their incentives, their likely level of technical fluency, what they care about, what they're afraid of. That's a 90-second text message. It's off and running in minutes.

While that's happening, I open NotebookLM. I drop in the source material. I generate a mind map, expand every node, run the synthesis reports. I ask it to produce a podcast briefing targeted at this specific audience. I ask it to build a six-slide deck — not thirty slides of bullet points. Six slides. High impact.

Then I listen to that podcast on the way to the room. I'm absorbing the material while the deck loads onto whatever device I'm presenting from.

I walk in prepared. Not because I spent eight hours manually building every artifact. Because I directed a temporary synthetic workforce toward a clear outcome.

That is not cheating. That is modern execution.

The Three Eras of Labor Leverage

To understand why this matters, you have to understand where we are in the sequence.

For most of human history, if you wanted more work done, you needed more people. Output scaled with headcount. That was the basic relationship.

Then software changed the equation. Spreadsheets replaced clerks. Scripts replaced repetitive tasks. Cloud platforms replaced server rooms. But automation was mostly rigid — you had to define the process in advance. The machine could repeat the workflow. It could not interpret intent. Agentic AI changes that.

Now systems can interpret goals, break them into steps, use tools, search, summarize, draft, build, test, revise, and report back. That is a categorically different thing.

Automation says: When X happens, do Y.

Synthetic labor says: Here is the outcome. Figure out the path.

That difference is enormous. It's the difference between a conveyor belt and a junior analyst. One repeats. The other adapts.

The industrial revolution mechanized muscle. The information age digitized knowledge. The AI age programs labor. And when execution becomes programmable, the entire economy reorganizes around whoever can aim it.

The Bottleneck Moved

Here is the load-bearing sentence in all of this: When labor becomes abundant, clarity becomes scarce.

The old economy rewarded execution skill. Could you write the report? Build the spreadsheet? Code the feature? Research the market? Draft the memo?

The new economy still values those skills. But it increasingly rewards the person who can coordinate ten versions of those skills simultaneously — because the bottleneck moved. The bottleneck used to be labor capacity. Now the bottleneck is clarity of intent.

One AI-native orchestrator directing five specialized agents can execute what used to require a small team. Not perfectly. Not without human judgment at every critical junction. But the math is real, and it's already visible in the market.

Eighty-two days — that's how long it took for a solo-built AI agent framework to go from nonexistent to acquired by one of the most powerful AI labs in the world. Six months — that's how long it took a single founder to build and exit a platform for eight figures in cash.

That's not irrational. That's leverage pricing. Capital is repricing fluency.

The New Professional Stack

The AI-native operator has a different skill set. Not just technical. Not just soft. A new blend.

Intent clarity. Can you define what you actually want? Most people cannot. They vaguely gesture at an outcome and hope someone turns it into execution. AI punishes vague intent — so does every good employee, but AI does it faster and without the facial expressions.

Task decomposition. The operator doesn't ask for the final thing. They ask for the system behind the thing. Not "make me a pitch deck" — but: analyze this audience, identify their incentives, extract the business problem, generate three strategic framings, build an outline, surface the objections a skeptical CFO would raise, then create the deck.

Agent selection. Which tool does which job? NotebookLM for source-grounded synthesis. A coding agent for implementation. A research agent for market scans. A design tool for visuals. The operator knows the bench and knows which synthetic worker to put in which seat.

Sequencing. Order matters. Research before writing. Audience analysis before design. Risk review before recommendation. Bad sequencing creates expensive nonsense at machine speed.

Verification. This is where the human stays very much in the loop. AI hallucinates. It overstates. It produces beautiful nonsense with the confidence of a TED speaker who just discovered stoicism. The operator catches it. Corrects the machine. This is not optional — it's the whole job.

Taste. Taste is knowing what good looks like. What sounds credible. What slide should be cut. What sentence is trying too hard. AI generates options. Taste chooses.

Strategic judgment. Not everything that can be produced should be produced. The operator decides what matters. Without judgment, you get infinite output and zero wisdom.

Why This Changes Hiring

The old hiring system is credential-forward. Where did you work? What title did you have? What degree did you earn? How many years of experience do you claim?

That model assumes past affiliation is the best proxy for future output. But synthetic labor weakens that assumption — because a single AI-native operator with strong judgment can now produce what used to require a team.

The better hiring question becomes: What can this person orchestrate?

Not just "what have they done?" — but "what can they cause to happen?"

The best candidates won't say "I'm good with AI." They'll say: Here's the temporary workforce I assembled around your problem. Here's what it produced. Here's where it was wrong. Here's how I corrected it. Here's what I'd do next. That is a monster interview answer.

The Corporate Implication

Inside organizations, this gets even bigger. Every department has hidden temporary workforces trapped inside overloaded employees. Marketing needs research, copy, segmentation, and reporting. Sales needs account research, call prep, proposal drafting, and pipeline analysis. Operations needs SOPs, dashboards, and process maps.

The modern company is full of work that doesn't necessarily require permanent headcount. It requires temporary capability. The company that figures this out stops asking "where can we add AI?" and starts asking: Where are we using permanent human labor for temporary cognitive work?

That question is dangerous. Useful, but dangerous. Because once you see it, you can't unsee it. And if you don't have a real AI-native operator — not someone who took a weekend prompt course, but someone who thinks in systems and orchestrates synthetic labor at the strategic level — you're structurally disadvantaged. That person isn't a nice-to-have. They're a C-suite-level asset.

Data Centers Are Labor Infrastructure

Most people are still getting this wrong. People think data centers store data. That's old language. Increasingly, data centers produce labor. Synthetic labor. Cognitive labor. Execution capacity.

A GPU cluster is not just infrastructure. It's a labor engine. A model is not just software. It's a workforce substrate. An agent platform is not just an app. It's a management layer for synthetic execution. That's why the capital intensity makes sense. If labor becomes programmable, compute becomes labor capacity. And if compute becomes labor capacity, data centers become the new industrial base. Not metaphorically. Economically.

The people moving hundreds of millions into AI infrastructure aren't building a fancier search box. They're building factories. Factories that produce task execution.

The Warning Label

Temporary synthetic workforces are powerful. Powerful things create blast radius. Bad data in, bad decisions out. Private information leaked into tools that should never see it. Hallucinated citations presented as fact. Output volume mistaken for value.

AI can produce infinite nonsense. That doesn't make nonsense strategic.

The operator's job is not to make more. It's to make better — faster, with less waste. That requires governance, privacy boundaries, and human accountability at every critical decision point. The temporary workforce may be synthetic. The responsibility is not.

The Honest Answer

So when someone asks how you'd solve a complex problem — how you'd prepare a high-stakes presentation, run a market analysis, build a strategic brief, research a room full of executives you've never met — the honest answer in 2026 is increasingly this: I'd instantiate a temporary workforce.

I'd spin up research. Synthesis. Translation. Design. Critique. Rehearsal. I'd direct those functions toward the outcome. I'd verify the outputs. I'd apply judgment. I'd correct the hallucinations. I'd walk in with the artifact. Not because I'm lazy. Because I understand what leverage looks like now.

The workers who win the next decade won't be the ones who do the most tasks manually. They'll be the ones who can assemble capability, direct execution, verify output, apply taste, and carry accountability — around any objective, on demand. The ones who understand that sentence are going to run circles around the ones still asking whether AI is cheating.

---

Rich Washburn is a technologist and strategist working at the intersection of AI, infrastructure, and capital. He is Managing Partner and Chief AI Officer at Eliakim Capital and CIO of Data Power Supply.