Your job might disappear in a few years.
This isn't being said by a hype-monger or an influencer. It's coming from someone who was inside OpenAI, specifically tasked with predicting the future of AI.
His name is Daniel Kokotajlo. He joined OpenAI in 2022, where his job was to forecast what would happen in the coming years. Now, he runs a small non-profit called the AI Futures Project, focusing exclusively on AI forecasting.
In an interview, he stated:
"Everyone should be afraid of losing their job. It's not just taxi drivers. Except for a few jobs that legally must be done by humans, almost everyone is a target."
To be honest, I was skeptical when I first read this. But when you look at his background and the testimony I'm about to share, it becomes hard to laugh off.
# The Man Who Refused to Be Silenced, Risking $2 Million
When Daniel left OpenAI, his termination papers included a specific clause:
Do not criticize the company. And do not tell anyone about the existence of this agreement.
If he didn't sign, he stood to lose equity worth about 80% of his assets at the time—nearly 300 million yen (approx. $2 million USD).
After agonizing for a month or two with his wife and consulting lawyers, he refused to sign. He decided there were things more important than money.
When this story spread online, it caused quite a stir. Employees within the company began questioning management about why such clauses existed, and the company eventually retracted the policy. As a result, Daniel was able to receive his money after all.
But what he wanted to convey by going this far wasn't about money. This was the view he saw:
"Most of the world is sleepwalking. Almost no one realizes what is actually happening with AI."
Publicly, companies only make modest announcements that stir up just enough expectation. According to what he felt inside, they don't really want to talk about what will truly happen in a few years.
# Abnormal Speed: 60x Revenue and 100x Parameters in One Year
The numbers Daniel shared are, frankly, quite abnormal.
A company called Anthropic had about $1 billion in annual revenue a year ago. Now, it's about $60 billion. That's roughly 60x growth in one year. He says this is unprecedented even for a startup and might be the fastest growth for a company of this scale in history.
He doesn't think this growth will continue at this exact rate. However, even if it slows down significantly, his view is that it's on track to reach a scale close to the entire economy by around 2030.
The number of parameters, which make up the core of AI, was around 175 billion in 2020. Now, it's around 10 trillion. In just six years, the scale has increased about 100 times.
Daniel emphasized that the speed of this trend itself is more important than the specific number of years until it arrives. Predictions on years can be wrong. But the pace of this acceleration has not stopped yet.
# Why CEOs Won't Step on the Brakes
It's impossible that the heads of these companies don't know how risky this is. Daniel says they actually understand it perfectly.
So, why don't they stop?
The answer is simple competitive psychology. Sam Altman, Dario Amodei, and Elon Musk each genuinely fear that if the other reaches the goal first, they might hold dictatorial power. Because they cannot trust each other, they feel they have no choice but to keep running to get there first.
And what these companies are aiming for now isn't just the automation of coding. They are trying to automate the entire research process—from generating ideas to conducting experiments and analyzing results. In other words, they are aiming for a state that eventually requires almost no human employees.
Dario of Anthropic called this a "country of geniuses" inside a massive data center. Daniel disagrees, saying it should more accurately be called an "army of geniuses." His reasoning is that it's not a collection of diverse individuals, but an army of copies of the same model, all following the company's instructions.
The corporate side's true feeling is: "It'll probably be fine, and the best way to ensure it's fine is for us to take the lead." Daniel sees this as nothing more than convenient self-justification.
# 70% Chance of a Worst-Case Scenario, But He's Not a Doomsayer
Daniel says that if nothing changes, there is a 70% chance we are headed toward a very terrifying conclusion.
However, this needs careful clarification. He explains that this doesn't mean human extinction, but rather the probability of a major catastrophe where AI escapes human control and takes the lead.
He cited two main risks. One is the loss of control, where AI falls outside human management. The other is the concentration of power, where only a few companies or individuals hold super-high-performance AI. He says these are deep-seated concerns that have been discussed for decades, long before the AI industry was born. This isn't a new conversation.
At the same time, he is not just a pessimist. He says he would be "overjoyed" if all his predictions turned out to be wrong. If we can solve the problems properly, AI has the potential to be wonderful for everyone. His stance is to look correctly at what might be coming, without being purely optimistic or pessimistic.
# Judge by Actions, Not Words: Applying This to Our Own Work
There is a phrase Daniel repeated:
"People should be judged not by what they say, but by what they do."
There is often a huge gap between what corporate leaders say in public and what they are actually doing. He said he felt this strongly when he actually met Sam Altman.
I think this is a perspective we can apply directly to our own work. Don't take the rosy promises of clients or industries, or the hype-filled announcements on social media, at face value. Verify your own position based on the numbers that are actually moving and the speed of that change.
Daniel even revealed that he told his wife they shouldn't have any more children. He feels that what's coming is too uncertain. It's important to remember that this story is being told with that level of gravity.
That said, what we can do now isn't to freeze in fear. Try writing down which parts of your job are likely to change and at what speed due to AI. Then, decide on just one thing you can move on now. By doing that, vague anxiety turns into concrete preparation.
The testimony introduced in this article is based on an interview with Daniel Kokotajlo himself on the channel "The Diary Of A CEO." If you're curious about his expressions and the weight with which he spoke, please check out the original video.
