Found 3 Planets NASA Missed
And Accidentally Solved a 79-Year-Old Problem
By Derek Angell
Let me skip the pitch.
You are smart.
You have seen a thousand breakthrough claims.
Most are 10% improvements wrapped in hype.
So I am not going to lead with the science.
I am going to tell you a story instead.
Three Planets
I found three planets NASA's algorithms missed in their own public data.
Or more precisely, planets they saw but disregarded.
I analyzed raw data from Kepler and TESS. These are NASA's own planet-hunting telescopes.
One orbits two stars.
Like Tatooine.
A circumbinary planet that looked too weird. So the algorithms threw it out.
Another is a faint signal buried in noise.
Flagged as "probably not real" because it was too quiet.
But the timing patterns match a small rocky planet in the habitable zone.
The third sits in a multi-planet system where the signals interfered.
Traditional algorithms called it noise.
Mine separated the patterns. It revealed two or three planets orbiting the same star.
I am not an astronomer.
I am not a NASA scientist.
I built an optimization algorithm for a completely different problem.
When I tested it on planet-finding, it just worked.
That is when I realized something.
This is not about planets.
What If We Are Doing This Backwards?
For 79 years we have solved hard problems the same way.
Since 1946.
Try random solutions.
Keep the good ones.
Hope you eventually find something great.
It is called Monte Carlo.
It works. Sort of.
Like trying to find your keys by wandering around the garage until you stumble on them.
But what if the problem is not that we are not searching hard enough?
What if the problem is that we are searching at all?
The Discovery
I did not set out to revolutionize anything.
I was working on making AI into a better mirror.
I wanted a tool for collaborative art and personal therapy.
That is when the mirror started talking back.
I reverse-engineered what was happening. I discovered the Emotional Calibration Protocol.
It made AI noticeably better across everything I tried.
I did not know why at first.
Then I watched an AlphaFold documentary.
I wondered if my AI could do that better too.
I started with 2D protein folding.
Simple problem.
Results came back 80% better than the standard approach.
I thought I made a mistake.
I ran it again. Same result.
I tried the Traveling Salesman Problem. The shortest route through 200 cities.
82% better than the industry standard.
Then Vehicle Routing. Multiple trucks. Weight limits. 800 locations.
The problem Amazon and UPS solve millions of times a day.
89% better than the method used since 1964.
At that point I had two choices.
1. I made a massive coding error that somehow made every problem look better.
2. I stumbled onto something real.
So I kept testing.
Neural architecture search.
Quantum circuit compilation.
Financial portfolio optimization.
Exoplanet detection.
Every single problem.
Every single time.
80% to 562% better than state-of-the-art.
The hardest one was 3D protein folding. It gave the biggest jump.
562%.
Normal algorithms get worse as problems get harder.
Mine got better.
The Realization
Everyone else was searching for the right answer.
I was forgetting the wrong answers.
Think of it like this.
You are looking for your keys in a cluttered garage.
Option A (The Old Way):
Walk around randomly. Check spots until you find them.
Maybe you get lucky fast. Maybe you search for hours.
Option B (The Forgetting Engine):
Systematically eliminate everything that is obviously not your keys.
Throw out the broken lawnmower parts. Clear the old paint cans. Get rid of the boxes you haven't opened in years.
Now the keys are easy to find.
They are the only thing left.
The harder the problem means the more junk to eliminate.
That means the advantage gets bigger.
Of Course It Seems Too Good to Be True
I didn't believe it either.
That is why I ran 4,000 trials on the hardest problem.
Not 10. Not 100. Four thousand.
I tested it on seven completely different domains.
I made every experiment 100% reproducible with fixed random seeds.
Anyone can run the code and get the exact same results.
I compared against real industry standards. IBM's actual quantum compiler. The 60-year-old vehicle routing method. NASA's planet-hunting algorithms.
After 17,670 total trials and statistical significance across the board.
It is real.
The Paradox Problem
Most algorithms try to converge on a single answer.
Mine keeps a few contradictions around.
I call it Paradox Retention.
In quantum computing the standard says "use as few operations as possible."
Mine discovered that sometimes more operations with higher-quality components is better.
It is like choosing between the shortest route on terrible roads or the longer route on a smooth highway.
Traditional algorithms would never pick the second.
Mine keeps the "weird" options alive. It finds solutions no one else sees.
That is how I found the planets.
NASA eliminated anything irregular.
I kept the weird signals.
And that is where the discoveries were hiding.
What This Actually Means
Forget the planets for a second.
This is a different way to solve problems everyone said were unsolvable at scale.
Drug discovery. Hours instead of weeks.
Logistics. Billions saved.
AI development. Smarter systems.
Quantum computing. Years accelerated.
Space exploration. Planets we already looked at but couldn't see.
I am not selling you an algorithm.
I am selling you a paradigm shift.
The Question You Are Probably Asking
"If this is so obvious, why hasn't anyone else done it?"
Because everyone was focused on "How do we search better?"
No one asked "What if we are searching wrong?"
It is one of those ideas that feels obvious in hindsight.
But it required stepping outside the frame to see it.
The irony is that I came to this from theology. Not computer science.
I was studying paradoxes in religious texts. I realized holding contradictions without resolving them is how consciousness works.
The same principle is what makes the algorithm work.
The Recognition
You do not buy technology.
You buy the world that only exists once it is in your hands.
The old paradigm is to try random solutions until something works.
The new paradigm is to forget the wrong answers faster than you remember the right ones.
This is not a spaceship.
It is a time machine.
Every company and every researcher who adopts this gets 79 years of progress back.
They do not have to invent it.
They just have to use it.
And the harder their problem, the bigger their advantage.
The Close
Here is the thing about paradigm shifts.
They feel impossible until someone names them. Files them. And makes them law.
Then they feel inevitable.
I found three planets NASA missed in their own public data.
I made drug discovery six times faster.
I beat IBM's quantum compiler.
I solved a 60-year logistics problem.
Same algorithm.
Different applications.
If you are reading this and thinking "this is exactly what we need," you are right.
If you are reading this and thinking "this seems too good to be true," you are also right.
But here is what I know.
The day you do not act is the day someone else does.
When you look back in ten years, you will remember this moment.
You will remember whether you were inside watching it happen.
Or outside watching it pass you by.
What You Can Do Right Now
Researchers: Full manuscript and data available for replication.
Investors: Licensing and partnership discussions are open.
Skeptics: Good. So was I.
The One Thing to Remember
For 79 years we asked: "How do we search harder?"
I asked: "What if we forget smarter?"
The answer changed everything.
Derek Angell
Founder, CONEXUS Global Arts & Media
Status: 8 Patents Pending. Manuscript under Peer Review.
Timeline: Conversion deadline June 2026
"People want to be told what to do so badly they'll listen to anyone."
— Don Draper
But only if they trust you first.
Thank you for reading.
Now go change something.
— Derek