- Soapbox News
- Posts
- The Supercar in Your Garage
The Supercar in Your Garage
You just got handed the keys to a Ferrari.

Hi reader,
You just got handed the keys to a Ferrari.
Zero to sixty in 2.9 seconds. 710 horsepower. Top speed of 211 mph.
Question: What happens next?
Most people? They stall it at the first light. Or they punch the gas and lose control before they hit third gear. Some wrap it around a tree at the first corner.
A few learn to drive it properly. They practice. They study. They fail. They get better.
And then there are the pros: they know exactly when to brake, which line to take through the corner, how to position the car before the apex. They can push it to the limit lap after lap without breaking a sweat.
That's what using an LLM (AI) is like right now.
Everyone Has the Keys
ChatGPT launched. Claude appeared. Gemini showed up. Suddenly everyone has access to a supercar sitting in their driveway.
The barrier to entry? Zero.
Type words. Get words back. Simple.
But here's what nobody tells you: getting output from an LLM is easy. Getting good output? That takes skill.
Real skill.
The kind you can't pick up from a YouTube video or a blog post. The kind that comes from hours of practice, from understanding how these systems think, from knowing when to push and when to pull back.
The Amateur Hour Problem
I see it constantly. Someone sits down with Claude or GPT-4, types in a vague prompt, gets mediocre results, and declares AI is overhyped.
"I asked it to write a marketing email and it gave me garbage."
Sure it did. Because you asked it to write "a marketing email" the same way you'd tell a Ferrari to "go fast" without knowing which gear you're in.
The AI gave you exactly what you asked for. The problem? You didn't know what to ask for.
You didn't specify tone. You didn't provide context. You didn't give examples. You didn't iterate. You didn't refine.
You just mashed the gas pedal and wondered why you spun out.
What Skill Actually Looks Like
Professional LLM use isn't about knowing tricks. It's about understanding the machine.
You learn what works through repetition. You figure out when to be specific and when to leave room for creativity. You discover that context is everything. You realize that the first output is rarely the best output.
You start to see patterns. This type of prompt gets these types of results. That phrasing unlocks better reasoning. Adding examples changes everything.
You develop intuition. Not magic. Not hacks. Just pattern recognition built through experience.
Can you learn it? Absolutely.
Does it happen overnight? Not a chance.
The Productivity Multiplier
When you know how to drive the car, everything changes.
Tasks that took hours take minutes. Projects that seemed impossible become routine. The quality of your output shoots up while the time invested drops.
I run three different ventures (Vervology, 315ny, TallowSoap (coming soon)). Without LLMs, I'd need a team of 14 people. With them? I can do more with less because I've learned to drive the machine.
Writing. Research. Analysis. Code. Strategy. Customer service. Content creation.
All of it moves faster. All of it gets better.
But only because I've put in the hours. Only because I've learned what works and what doesn't. Only because I've crashed the car enough times to know where the limits are.
The Revolution Is Real
This isn't hype. This isn't a bubble. This is a fundamental shift in how work gets done.
LLMs are tools. Like spreadsheets. Like the internet. Like email.
Except they're more powerful. More flexible. More transformative.
They don't replace humans. They multiply human capability. They take the ceiling off what a person can accomplish.
But you have to learn to use them. Really use them. Not just dabble. Not just prompt and pray.
You have to practice. You have to fail. You have to get better.
And It's Fun
Here's the part people miss: driving a supercar is exhilarating.
When you nail a corner perfectly. When you feel the car respond exactly how you wanted. When you push it harder than you thought possible and it handles beautifully.
That's what it feels like when you get good with LLMs.
You ask a complex question and get a brilliant answer. You iterate on an idea and watch it evolve. You solve a problem in seconds that would have taken days.
It's satisfying. It's exciting. It's genuinely fun.
But you only get there if you learn to drive.
So What Now?
You've got the keys. The supercar is sitting there waiting.
What are you going to do with it?
Are you going to tap the gas nervously and complain it's too complicated?
Or are you going to learn to drive it properly?
The machine doesn't care. It'll sit there whether you use it or not.
But if you put in the time, if you develop the skill, if you learn to push it without crashing?
You'll never look back.
The future belongs to people who can drive fast without wrapping it around a tree.
Start practicing.
Jono