Scan time: 2-4 minutes / Read time: 5-7 minutes
⠀
Hello rebel ideapreneurs 🦸♂️🦸♀️
⠀
🙋♂️ Wes’s first issue, still not sent
Wes is a data analyst who decided in March to start a newsletter about money numbers.
Then he started polishing the first issue, and he is still polishing it.
It is September now, and not one issue has gone out.
🧩 His struggle:
He never gave the first issue a deadline, so there is always time for one more fix.
One evening, a coworker sends him the true story of a tech boss
whose company also waited almost a year to show the public its finished AI,
and then ChatGPT got a million users in five days.
Even his own team thought only a few thousand people would care.
⠀
💬 The chatbot that went out with a tweet
Sam Altman runs OpenAI, the company that made ChatGPT.
Before that, he started a phone app called Loopt, and sold it in 2012 for $43.4 million.
Then he ran Y Combinator, which invests money into brand-new startups and also mentors the founders.
In 2022, OpenAI was, in Sam’s own words, “a quiet research lab.”
Get this — the AI that powers ChatGPT was already finished, almost a year before ChatGPT came out.
OpenAI had been renting it out to other companies, to use in their own software.
Jan Leike, one of the researchers, put it plainly: the same thing had been available “for almost a year before ChatGPT came out.”
OpenAI also had a website where programmers could test the AI for free, to see if it was worth paying for.
They called that website the playground.
You typed something in, and the AI wrote something back.
The team noticed something odd in the playground.
Programmers were not just testing it.
They were chatting with it, for fun, like it was a person.
So the team had a simple idea.
Take the AI that already existed, put a chat box in front of it, and let anyone in the world type to it.
The team took the idea to Sam, and Sam said yes.
He did not ask for a six-month plan or a big marketing campaign.
Two weeks later, ChatGPT went out to the world.
Everybody remembers ChatGPT as a huge launch, because the whole world was talking about it within days.
But OpenAI did not do any of that talking.
There was no press event, no ads, and no big announcement.
Inside the company, they called it a “research preview,” a sneak peek rather than a finished product.
The day before, Greg Brockman, OpenAI’s president, told the staff not to expect much.
Maybe one post on Twitter, he said, that a few thousand people would like.
Late on the night of November 29, Sam was in the San Francisco office with the team, putting the finishing touches on it.
The next morning, November 30, Sam posted one message on Twitter saying ChatGPT was live.
Five days later, he posted again.
One million people had already used it.
Nobody at OpenAI had expected that.
Jan Leike, the researcher from earlier, admitted they were caught off guard.
“It’s been overwhelming, honestly,” he said.
“We’ve been surprised, and we’ve been trying to catch up.”
The AI behind ChatGPT sat there for almost a year, and the world barely noticed.
The hard part was never building it.
The hard part was deciding to put it out there,
and then shipping it in two weeks instead of polishing it for six months.
⠀
🔥 The recipe
➡️ Speed After the Decision
ChatGPT went out two weeks after Sam said yes, with one tweet, and had a million users five days later.
This strategy is called Speed After the Decision.
Sam put it this way in his essay How to Be Successful:
“Once you have figured out what to do, be unstoppable about getting your small handful of priorities accomplished quickly.”
Sam’s team wasn’t doing sloppy work.
They moved fast because every decision had already been made, and building was the only thing left to do.
So the day you decide to start a newsletter, you send the first issue that same week.
You don’t send the perfect issue.
You send the first one.
The polishing happens on issue two, and issue three, while real readers are already reading.
AI made building fast, so the slow part is now the deciding.
⠀
🛠️ Wes’s turn
Now, back to Wes, who decided in March to start his newsletter, and still hasn’t sent the first issue.
That night, he runs the prompt at the bottom of this email, and tells it every fix he has made to the issue since March.
His AI sidekick doesn’t waste words:
Your first issue is ready, so send it this week, and save every new fix for issue two.
He stops fixing it that night.
On Thursday morning, he sends it to the forty people on his list.
⠀
🌅 Two months later
Wes has sent an issue every Thursday since.
Readers write back to tell him which numbers confused them, and he fixes those in the next issue.
A hundred and twenty people read it now.
You may have decided on something months ago, and still be polishing it.
Once you’ve decided what to make, put it out in two weeks, instead of polishing it for six months.
That’s it, my fellow outliers!
Yours ‘helping you build profitable businesses 25x faster’ Vijay Peduru 🦸♂️
⠀
⠀
⠀
⠀
⠀
🏄♀️ The prompt
The prompt below takes the thing you already decided on and plans the fastest way to get it out.
🧰 What you’ll need
What you already decided to do, and when you decided it
- Type it, or say it out loud with a voice tool.
- e.g. start a newsletter that explains money numbers in plain words, back in MarchWhat has happened since, and what you keep doing instead of the next step
- e.g. issue one is written, I have rewritten it six times, and I keep fixing one more chart instead of sending itThe next real step
- e.g. sending issue one, about 1,500 words, to the forty people on my list
🚀 How to run it
Open your AI sidekick, ChatGPT, Claude, or whichever one you use.
Swap the three examples, the text inside the curly brackets, for your own, then run it. If your AI sidekick can already see your notes, you can run it as is.
🧪 Want to test it first? Run it without changing anything, and it will use the examples.
CONTEXT:
- Use my own material first. If you can see my drafts, my sent issues, or my subscriber dashboard, pull the real details from there instead of the examples below.
- If you know how I write, use my voice. Otherwise, write in a clear, warm, no-jargon voice — first-person, plain words, no startup-speak.
- If you know who my readers are, write for them. Otherwise assume people who want money numbers explained in plain words.
- If I have replaced the examples, use only mine. If an example is still there, fill it in from what you already know about me, and mark it as your assumption. If you don't know anything about me yet, use the example exactly as written.
- Do not revisit whether this is the right thing to do. That decision is made. Every output is about compressing the time between now and it existing.
=============== INPUTS — edit these, or just run as is ===============
What I already decided to do, and when I decided it:
{e.g. start a newsletter that explains money numbers in plain words, back in March}
What has actually happened since, and what I keep doing instead of the next step:
{e.g. issue one is written, I have rewritten it six times, and I keep fixing one more chart instead of sending it}
What the next physical step actually is:
{e.g. sending issue one, about 1,500 words, to the forty people on my list}
======================== END OF INPUTS ========================
Outputs:
1. The elapsed cost — how long since the decision, and what has been produced in that time, stated plainly.
2. The critical path — the shortest chain of steps between now and the thing existing, with nothing optional on it.
3. What comes off — everything I have been doing that is not on that path, named.
4. The compressed schedule — those steps across the next fourteen days, with what happens on which day.
5. The quality floor — what "good enough to ship" means here, decided now so it cannot move later.
Then tell me which step I have been avoiding, and what the smallest version of it is.