MAY 21, 2026 · 5 MIN READ

Your job was never typing

You probably do it every day: type. It would be bizarre if you didn’t, almost regardless of your job.

You log in somewhere, you send a text message.

If you work in an office environment, or run a business, our guess is you type a lot. You email, a lot. You text, a lot. You write slide decks, edit documents, open programs and write stuff in them, a lot.

Typing isn’t your job though. Even the materials you type, the completed ones, are not your job.

What do the smart people say?

Jensen Huang is the CEO of Nvidia. For most people, they never buy anything from Nvidia even though you’ve probably heard the name. Chances are, you’ve used a device of some kind with Nvidia technology, a chip, a processor, inside of it.

Huang has been the CEO of Nvidia since 1993, day 1 of the company.

He’s well-regarded in the tech world, and today one of the most sought after voices on the rapid advancement of artificial intelligence.

Huang said a couple weeks back: his job should be gone. At least, if you view the world through the lens of an AI pessimist or critic.

Kind of, anyway. Here are his exact words:

“What Jensen does for a living is tap on phones and talk. And tapping on phones and talking, AI has done that just fine, and therefore my job should be gone — but I’m busier than ever. So there’s a fundamental difference between the purpose of the job and the task of the job.”

He goes on to make a point most of the AI commentary class is missing. We don’t need a billion lines of code written. We need a trillion. The constraint on what humans build has been typing speed for 50 years. Pull that constraint out and the work expands, it doesn’t disappear.

Then he buries the lede:

“The idea that being human means to hunch over on this little thing, typing all the time. 50 years before that, people didn’t do that.”

Read that twice. The hunched-over-a-keyboard version of work is roughly as old as the personal computer. About as old as our parents. It just happens to be the only state most of us have ever known.

The task is not the job

Every SMB owner we talk to hits the same wall. They started a business to do a thing: install HVAC systems, treat patients, fabricate parts, manage a portfolio, cut hair, fix teeth. Then somewhere around year three they realized 60% of their week is typing. Quotes. Invoices. Scheduling. Insurance forms. Compliance attestations. Email triage. Status updates to a CRM nobody reads.

That’s the task layer. The job is the reason they signed the lease and put their name on the door.

The reason this matters now, and didn’t matter five years ago, is that the task layer is what AI is actually good at. Not the job. The job (judgment, relationships, knowing which customer is about to churn, knowing which technician is two months from quitting) is still yours. The typing is what’s leaving.

The headlines say AI is coming for jobs. AI is actually coming for the worst part of every job. The part you didn’t sign up for. The part that grew like kudzu over the last 50 years until it ate your week.

Rough math on a 30-person HVAC company

Take a real-shaped scenario. Rough math, not from a specific engagement:

A 30-person residential and light-commercial HVAC company in central Virginia. Owner is a master tech who built the business from his truck. His week, broken down honestly, looks like this:

  • 12 hours on quotes, invoicing, and scheduling typing-work
  • 8 hours on email and texts
  • 6 hours on compliance, payroll review, vendor paperwork
  • 4 hours actually on customer sites or with techs
  • The rest scattered across meetings, fires, and the drive between them

Four hours on sites. He started an HVAC company. He spends four hours a week on HVAC.

If you automate 80% of the quote-to-invoice-to-schedule pipeline (a conservative number for what’s possible today with off-the-shelf models wired into his existing tools), he gets roughly 9 hours back per week. Not in theory. Mechanically. The quote draft writes itself from the tech’s voice memo. The invoice fires when the work order closes. The schedule rebalances when a tech calls out.

Nine hours a week is 450 hours a year. At the value of an owner-operator’s site time, that’s a different company. He’s closing more $8K-$15K jobs, catching warranty issues before they escalate, and coaching the green tech who’s about to wire something backwards.

The job didn’t change. The task layer collapsed.

What we keep seeing in the field

Two patterns show up over and over:

  1. The typing-work eats more than time. It’s the part of the week they hate, which means they avoid it, which means it piles up, which means quotes go out three days late and customers go to the competitor. The cost is 12 hours plus the deals that died waiting in the queue.
  2. When the typing comes off their plate, they don’t relax. They don’t take Friday off. They do more of the thing they started the business to do. The HVAC owner gets back on roofs. The dental practice owner sees more patients. The family-office partner gets to actually think about the portfolio instead of reconciling spreadsheets. Huang said it about himself: he’s busier than ever, doing the part of the job that’s actually his job.

The output of the company goes up. The owner’s sanity goes up. The work that’s left is the work that has the owner’s fingerprints on it, which is the only work the owner can do anyway.

The 50-year anomaly

The thing we keep coming back to is Huang’s throwaway line about the 50 years. Before personal computers, the master plumber didn’t spend 60% of his week typing. The country dentist didn’t either. They worked on the thing. Someone else, or nobody, handled the paper.

Then computers showed up, and instead of removing the paper, we just digitized it and handed it back to the operator. The owner became the data-entry clerk. Every SMB owner has, somewhere on their desk, a stack of work they “need to get to” that is functionally a typing queue.

The typing was never supposed to be there in the first place. AI taking it is just the cleanup. We’re exiting a 50-year detour.

For SMB operators, that’s the whole story. The week they thought they’d have when they started the business — that one was real. They’re just getting it back.

Tagged #AI for SMB #HVAC AI #Ironworks AI #Jensen Huang #small business automation #SMB operators #task vs purpose #Virginia AI consulting