Sasha Thackaberry-Voinovich

The old guard of Silicon Valley used to have a saying: "Software is eating the world." For two decades, that meant if you couldn't write Python or Java, you were just a passenger. You had the ideas, but the engineers had the keys to the kingdom. If you wanted to build a solution for the legal industry or the medical field, you had to spend six months trying to explain the nuance of your profession to a developer who had never stepped foot in a courtroom or a clinic. It was enormously expensive as a domain expert to stand something up.

Subject Matter Experts were hampered.

In the old world, we never could have launched a new university.

Enter 2026, and the script has officially been flipped. We all know this: the "coding bottleneck" is dead.

But the implications of this may be a surprise to many. Pattern matching is dead now. VCs and investors, watch out. (Side note: if you invest based on old patterns, you increase your risk.)

We’re entering the era of the Leader-Doer. These are individuals who possess deep, "un-Googleable" domain expertise and have mastered AI tools to build their own solutions. They aren't just "writing things" with AI; they are choreographers within the system, configuring, designing, and directing bots to solve niche pain points that generic apps completely ignore.

The Proof is in the Prototype: The Claude Code Hackathon

If you want to see what this looks like in the wild, check out the results of the February 2026 Claude Code Hackathon. Focused on building with Opus 4.6, the event saw over 500 participants.

Spoiler alert: the winners weren’t just "traditional" software engineers. They were doctors, lawyers, and engineers—people with dirt under their fingernails in their respective industries.

These experts used the Claude Code terminal tool to bypass the months of development hell that usually kill a good idea. Instead of begging for a dev budget, they just built the things.

Because you can now. It’s just that only some people believe it.

The Winners’ Circle:

• 1st Place: CrossBeam (Mike Brown): Mike didn't build another "to-do list" app. He tackled the nightmare of California permitting for builders and municipalities. That’s a niche, painful, "I-want-to-pull-my-hair-out" problem only a domain expert would think to solve.

• 2nd Place: Elisa (Jon McBee): A visual, block-based programming environment for kids that generates AI agents. It’s about building the next generation of AI choreographers.

• 3rd Place: postvisit.ai (Michal Nedoszytko): This transforms doctor visit transcripts into personalized health guidance. It solves the "what did my doctor just say?" problem with medical-grade nuance.

• "Keep Thinking" Prize: TARA (Kyeyune Kazibwe): A system that turns road footage into economic and infrastructure investment insights.

• Creative Exploration: Conductr (Asep Bagja Priandana): A MIDI-controlled AI generative music band.

How We Built Newstate U from the Ground Up

This isn’t just a trend we’re watching; it’s the DNA of Newstate University. We are a startup university born from the realization that there is a massive, unmet need for business, AI-focused education that results in actual university credentials—at a radically affordable cost.

We didn't hire a massive team of 200 people to launch. Instead, our entire team consists of Leader-Doers. We are "hands-on keyboard," using AI to build the systems, analyze the data, and create the bots. We don't just "manage" the work; we choreograph it.

And we’re not a team of administrators - we’re also faculty. We teach in the subjects that we’re domain experts in, and we’re teaching what we’re learning every day.

Our focus is on impact, not the size of our headcount. By removing the friction of traditional bureaucracy and using AI to handle the "repeatable gunk," we can focus on what matters most: student outcomes and sustainable growth for the middle class.

Now go do your thing!

Side note - how did I write this? A simple process.

1. I noted the results of the Claude competition based on one of the many newsletters I get and have summarized.

2. I asked Gemini for an overview of the results.

3. I used a pre-created tone and writing sample document along with prompts to write the initial draft (context stacking, we teach it).

4. I rewrote and edited sections

5. I plugged it into Perplexity to check for final factual accuracy

6. I had Nano Banana create an infographic based on the existing style and Brand Book we have.

Estimate of the time commitment? (Including posting to Substack, LI, etc.) About 20 minutes.