There's a formula that can predict whether you'll become successful.
But it has a few blind spots, and I wanted to know what happens to your chances once AI gets involved.
A few weeks ago I came across this Substack article with the title “The Mathematical Reason Most People Never “Make It” written by Kaguura Gichuru. If you read the article, you can detect the hidden message. That if you can be part of the reasons why people "make it," you will be a success too. And I found myself wondering right away what AI does to those factors.
Within the article the author mentioned Price’s Law. A mathematical observation from the 1960s, stating that the square root of the people in a field accounts for half the output. In a team of 100, that’s 10 people. In a field of 10,000 researchers, it’s 100. The physicist Derek de Solla Price discovered it in 1963 while studying scientific citations.
You can also apply Price’s Law to your own 40 hour workweek. Only 6 hours of your work produce half the output, and 34 hours of your work produce the other half. So logically it would make sense that if we can maximize the impact of those 6 hours, maybe even with AI, the probability of becoming successful increases. But this is where it starts getting fuzzy, because what does becoming successful actually mean?
In the article, Gichuru says that a few people in the named square root share four things. Skill, consistency, opportunity, luck. Whoever has all four makes it. Whoever has one or two doesn’t.
And even though it sounds plausible at first read, Price’s Law says nothing about who ends up in the square root. It only describes that a square root exists. The four components were invented by the author.
So I asked myself this question. What does research really say about who becomes successful, and what would a more honest formula look like, especially now that AI enters as a new variable.
What is success?
Self-determination theory, one of the most researched frameworks on human motivation, describes three basic needs that have to be met for people to experience themselves as fulfilled and successful. Autonomy, competence, relatedness. What immediately stands out is that these components describe inner success, and they’re hard to plot on a growth chart.
Outer success is the other half, and this is the part that can actually be measured, which helps us when building the formula. Angela Duckworth studied this in a longitudinal study of over 11,000 cadets and found that neither intelligence nor perseverance alone is enough. Both contribute in their own way to the outcome, alongside other factors like physical capacity. From this comes grit, the combination of passion and perseverance toward long term goals, a factor that stands on its own and doesn’t simply fold into consistency. Self-determination theory adds autonomy to this, meaning how much someone can set their own goals rather than have them handed down, also its own factor that stands apart.
In other words, the real predictors of success are more specific, harder to generalize, and less convenient, because they can’t be solved with more hustle alone.
Where AI actually intervenes
If we now take these predictors, competence, autonomy, relatedness, grit, AI doesn’t affect them equally.
Competence improves with AI for nearly everyone, regardless of starting point. A good tool makes beginners better faster, and it makes experts better faster, by a similar margin. Here I do have to say that I have heard someone say before, that AI makes brilliant people more average, but I would argue that this only happens if brilliant people use AI the wrong way. But that’s a different story.
Back to the topic …
Autonomy and relatedness work differently. They depend on structures that already exist or don’t. Someone who already has a network that grants them autonomy, because they can afford to hire help or because their voice already gets heard, can use AI to expand that autonomy further. Someone inside a system without those structures gets more output from AI, but not automatically more autonomy.
A formula that actually shows this
I tried putting this into a formula that captures the difference rather than just asserting it.
The probability of success depends on your competence, multiplied by its own AI factor, plus your structural access, multiplied by a separate AI factor that grows with the access you already have, all of it shaped by a randomness term that deliberately never disappears, because the future can’t be predicted.
Written out, P equals a sigmoid function of, a times K times AI factor of K, plus b times S times AI factor of S, all multiplied by epsilon. K stands for competence, S for structural access, and the AI factor for each variable is treated separately on purpose. Epsilon is the randomness term that stays open.
The key difference sits in the two AI factors. For competence, the AI factor is roughly similar for everyone, regardless of where K starts. For structural access, the AI factor grows with S itself, the more access you already have, the more AI amplifies it. That shows why the gap is widening instead of closing.
What you’d actually need to track
The formula only becomes useful if you can apply it to yourself, which means tracking the right things.
For competence, that means watching how feedback changes over time, whether clients or collaborators return, how your output per hour shifts once you start using AI compared to before, and how your error or correction rate develops.
For structural access, that means being honest with yourself about how much reach, time, capital, and access to decision makers you already had before your current work, independent of how good that work actually is.
And the randomness term stays the randomness term. There’s nothing to track there by definition.
What this means if you take it seriously
If you take this formula seriously, the question you should be asking changes. You are not looking at if you have assembled skill, consistency, opportunity, and luck. But where you stand on structural access, and whether the AI you’re using right now is expanding that access or just amplifying the head start you already had.
For some, that means looking honestly at whether their AI advantage is a genuinely new opportunity, or just an old head start growing faster. For others, it means the lack of progress isn’t automatically their own fault, it’s a structural gap that AI has so far widened more than closed.
I put together a cheat sheet with the exact things to track. You can find it at my digital garden, vely.ai.