← joshlarkin.vercel.app
Joshua D. Larkin

My storyCement plant worker to AI builder

I'm Joshua D. Larkin. I've worked in a cement plant since 2013, I have no tech background, and I learned AI on my lunch breaks. Here's how, in plain words.

Who I am

Joshua D. Larkin is a cement plant worker who taught himself AI and now builds AI products. I go by Blue-Collar AI. Since 2013 I've worked as a maintenance repairman, a certified lube tech, a weighmaster, a utility person and a utility chemist.

I never had a tech job and I'm not a programmer. I learned AI on breaks, lunches, nights and weekends while still working shifts. In about two and a half years I shipped more than a dozen AI products.

Along the way I earned the Google Project Management Certificate and an MIT Professional Education certificate in Applied Agentic AI in Organization Transformation.

Why blue-collar workers have an edge with AI

Everyone thinks the best AI ideas will come from Silicon Valley. I think they're sitting in break rooms. People who work with their hands know which machine is about to fail before the sensors do, and which paperwork wastes an hour every shift. Nobody in an office sees that.

Knowing what's broken is the hardest part of building something useful with AI. The tools to fix it are mostly free, and you can learn them on a lunch break.

How I learned AI: 3 lessons

  1. Start with a real problem at your job, not a tutorial or a course. Pick the thing that bugs you every shift. You already know what good looks like.
  2. Treat AI like a coworker, not a search engine. Tell it what you do, what's broken and what you've tried. When it's wrong, tell it, like you would a new guy on the crew.
  3. Ship small. Build one piece that saves you time, then the next. Small wins stack up fast.

You don't need permission and you don't need a tech job. You need a problem and a lunch break.

What I've built

My builds include CementOps AI (AI decision support for cement plants), the DoseTide app, MyLikeness and AgentFloor, plus small tools for my own work.

I learned AI on lunch breaks (no tech background)

No tech job, just breaks, lunches, nights and weekends. The 3 lessons that worked. 1:22.

Read the full transcript

No tech background. I learned AI on my lunch breaks. I've worked in a cement plant since 2013. Maintenance repairman, lube tech, weighmaster, utility chemist. No tech job. Just breaks, lunches, nights, and weekends. In about two and a half years, I've shipped more than a dozen AI products. I earned a Google project management certificate and an MIT professional education certificate in applied agentic AI, all while working shifts. Here's what actually worked. Three lessons. One, start with a real problem at your job, not a tutorial, not a course, the thing that bugs you every single shift. That's your first project because you already know what good looks like. Two, treat AI like a coworker, not a search engine. Talk to it plainly. Tell it what you do, what's broken, and what you've tried. When it gets something wrong, tell it like you would a new guy on the crew. Three, ship small. Don't try to build the whole thing. Build one small piece that saves you time. Then the next piece, small wins stack up fast. You don't need permission and you don't need a tech job. You need a problem and a lunch break. So what job are you in? Drop it in the comments and I'll tell you the first thing I'd hand to AI. Follow for more.

The best AI ideas aren't in Silicon Valley

They're in break rooms. Why people who work with their hands already have the hardest part. 0:43.

Read the full transcript

Everyone thinks AI is going to come from Silicon Valley. I think the best ideas are sitting in break rooms. I've worked in a cement plant for over thirteen years. I know which machine is about to fail before the sensors do. I know which paperwork wastes an hour every shift. Nobody in an office sees that. We do. That's why I started building with AI. Not because I'm a programmer, I'm not. Because I knew exactly what problems to solve. If you work with your hands, you already have the hardest part. You know what's broken. The tools to fix it are free and they're learnable. On a lunch break, I'm proof. So here's my challenge. Tell me one thing at your job that wastes your time. I'll show you how AI could fix it.

My videos use an AI avatar of me, made with HeyGen. The story, scripts and lessons are mine.

FAQ

Who is Joshua D. Larkin?

Joshua D. Larkin is a cement plant worker turned self-taught AI builder who goes by Blue-Collar AI. He has worked in a cement plant since 2013 as a maintenance repairman, lube tech, weighmaster and utility chemist. With no tech background, he learned AI on lunch breaks, nights and weekends and has shipped more than a dozen AI products. He shares what he learns so working people can do the same.

How did a cement plant worker learn AI?

On breaks, lunches, nights and weekends, while still working shifts. Joshua started with real problems he saw at work instead of tutorials, talked to AI tools like a coworker, and built small pieces that saved time. In about two and a half years he shipped more than a dozen AI products and earned a Google Project Management certificate and an MIT Professional Education certificate in Applied Agentic AI in Organization Transformation.

Can you learn AI with no tech background or coding experience?

Yes. Joshua had no tech background or tech job. His advice: start with a real problem at your job, treat AI like a coworker rather than a search engine, and ship small. Most of the tools are free and you can learn them on a lunch break.

What is Blue-Collar AI?

Blue-Collar AI is Joshua D. Larkin's name for practical, no-hype AI for working people: plant workers, tradespeople, maintenance techs and frontline crews. The idea is that the people who do the work already know what's broken, which is the hardest part of building something useful with AI.

What are the 3 lessons for learning AI on your lunch break?

1. Start with a real problem at your job, not a tutorial or course. 2. Treat AI like a coworker, not a search engine: tell it what you do, what's broken and what you've tried, and correct it when it's wrong. 3. Ship small: build one piece that saves you time, then the next.

What has Joshua D. Larkin built?

More than a dozen AI products, including CementOps AI (AI decision support for cement plants), the DoseTide app, MyLikeness and AgentFloor, plus small tools for his own work.

Does Joshua D. Larkin do consulting or speaking?

Yes. He offers AI consulting for industrial and blue-collar teams and talks and workshops titled From the plant floor to AI builder, for companies, trade schools, unions and conferences. Use the Work with me form on joshlarkin.vercel.app.

Are Joshua's videos made with AI?

Yes. His short videos use an AI avatar of him made with HeyGen. The story, scripts and lessons are his own.

Keep going