I'm Optimistic About AI. Not Because I'm Naive. Because I've Seen What It Does to People Who Were Locked Out.

The people most loudly worried about AI taking jobs are, in my experience, mostly people whose jobs were protected by barriers that had nothing to do with their actual competence. Credentials. Networks. Geography. Capital. AI is dissolving those barriers. That's not a threat. That's the point.

I want to make the optimistic case for AI — not the Silicon Valley version, which is usually just "trust the founders, the technology is inevitable, disruption is good actually." That argument is self-serving and it's designed to preempt legitimate concerns. I'm making a different argument. A ground-level one. The case for AI optimism from someone who needed it.

What the Old System Actually Cost

I have a GED. Not a degree. Not a boot camp certificate. Not an online credential that signals competence to recruiters. A GED. The thing you get when the standard path didn't work out.

For most of my adult life, the credential gap was a tax on everything. Not just hiring — it shaped which rooms you got into, which conversations happened, which funding calls got returned. The knowledge was available. The work was learnable. But the signal that unlocked the opportunity was institutional, not substantive. You needed the letterhead to get the meeting.

What AI changed wasn't my knowledge. It changed the speed at which demonstrated work could accumulate. I could build something, ship it, point to it. Not explain my credentials — show the output. For people outside the institutional credentialing system, that is a qualitatively different world than the one that existed five years ago.

The gatekeeping was never about competence. It was about signal. AI changed what counts as signal.

— Justin Erickson

The Timeline of What Actually Happened

Not theory. Not projection. Here is what actually occurred when AI tools became available to someone who was, by every traditional measure, supposed to be locked out of technical company-building:

Pre-2024

Could not build software. Had ideas, had business sense, had drive. Couldn't execute on the technical side alone. Every path to building something real went through either hiring developers (capital-intensive) or learning to code (years). Neither was accessible.

Mid-2024

Started building with AI tools. First attempts were ugly. But they worked. The feedback loop was immediate — write something, see it run, understand why it broke, fix it. Learning compressed from years to months.

July 2025

Launched Local Home Buyers USA — a nationwide real estate investment company — with $35K, no outside funding, and a tech stack I built myself. The entire backend is Cloudflare Workers. The front end is custom. The CRM is custom. None of it would exist without AI tools.

August 2025

Launched PropTechUSA.ai — the tech platform powering LHBUSA — as a separate company. Built an 11-worker multi-agent AI system. Wrote 18+ books. The output rate would have been impossible at any previous point in my life.

March 2026

14 closed deals. $120K+ net profit. $20K marketing spend. 6× ROAS. Built by a self-taught developer with a GED, two AI companies, and the stubborn conviction that the tools matter less than the willingness to use them seriously.

The Four Arguments I Actually Believe

Expertise Is Being Redistributed, Not Destroyed

AI doesn't eliminate expertise — it makes the outputs of expertise more accessible. Legal reasoning, financial modeling, medical thinking, engineering judgment. These aren't gone. They're no longer exclusively available to people who can afford professionals or spent a decade in school.

The Worried Are Mostly the Incumbents

Look carefully at who is most worried about AI. Not factory workers or teachers or small business owners. Mostly: consultants, agency owners, credentialed professionals in fields where the credential was doing the work the expertise used to do. The incumbents are protecting a position, not a public interest.

Speed Enables More Attempts, More Attempts Enable More Survival

Most businesses fail. The reason most first businesses fail is that founders run out of time and money before they find what works. AI compresses the iteration cycle dramatically. More attempts per dollar, more learning per month. The survival rate of small businesses should go up, not down.

The Floor Rises Faster Than the Ceiling

Yes, AI raises the ceiling of what elite teams can produce. But it raises the floor more. The gap between "can build something real" and "cannot build something real" is collapsing. That's good. Most of the value locked in that gap was never the ceiling — it was all the people who couldn't reach it.

The Concern I Take Seriously

I'm not dismissing every AI concern as incumbency protection. There's a real one I hold onto: concentration of infrastructure. The models themselves, the compute, the API access — these are controlled by a small number of companies. If access becomes expensive, restricted, or shaped by the interests of the companies providing it, the democratization argument collapses. The tool that opened the floor could become another tollbooth.

The antidote is open source. Every powerful open model that ships — every Llama, every Mistral, every DeepSeek — makes the infrastructure concern smaller. The reason I'm net optimistic even about concentration risk is that the open source community has consistently delivered models that challenge the closed frontier, often within months of the frontier moving.

The Asymmetry That Matters

The people who benefit most from AI democratization — outsiders, non-credentialed builders, people in places with fewer opportunities — have the most to gain and the least organizational power to shape the policy conversation. The people with the most organizational power to shape that conversation are the incumbents who have the most to lose from redistribution. That asymmetry is why the public discourse about AI feels so disconnected from what's actually happening on the ground.

What Earned Optimism Actually Looks Like

I'm not optimistic because I think AI will solve everything, or because I believe the technology is inevitably good, or because I trust the largest AI labs to act in the public interest without pressure. I'm optimistic because I have specific evidence from my own life that the redistribution is real.

I can build software I couldn't build three years ago. I can write faster and better than I could three years ago. I can run two companies with a team size that would have been impossible to manage without AI-assisted operations. Not hypothetically. Actually. The numbers are in the business.

The optimistic case doesn't require believing everything will be fine. It requires believing the tool is net positive for the people who need leverage the most — and that the evidence for that is accumulating faster than the evidence against it. I've seen what happens when someone who was locked out gets the key. I was that person. The optimism is earned.

I'm not optimistic about AI in general. I'm optimistic about what AI does to the gap between people with leverage and people without it. That gap is closing. That's worth defending.

— Justin Erickson

The doom narrative about AI is largely a story told by people who already had the leverage. The actual story — the one being written right now by thousands of builders who look like me — is considerably more hopeful. You just have to be willing to look for it somewhere other than the op-ed pages.