The Real Reason You Can't Find a Job Right Now

追加: 2026-06-02

[00:00:00]In the spring of 2024, a woman named Karen sat down at her desk at a mid-size accounting firm [music] and did something that would have been unthinkable 3 years earlier. She deleted her own job description. Her firm had just deployed an AI platform that [music] could process tax filings, flag anomalies, and generate first draft client reports in under a minute.

[00:00:24][music] Karen had been doing those same tasks for 11 years, and here is the thing about Karen. She was very good, in fact excellent at her job. She was a person other people came to when something felt off about the filing. She has, as her manager once told her, good instincts.

[00:00:42][music] But good instincts don't show up on a software licensing bill. Karen was not fired. She was given what her firm called an expanded [music] scope of responsibilities. She now manages the AI's output. She reviews four times as much many client [music] files as she used to. She makes the same salary.

[00:01:02]Karen kept her job. She just doesn't recognize it >> [music] >> anymore. Hey, welcome to AI Revolution Digest, where bring you powerful insights from CEOs, innovators, industry leaders who are redefining businesses, leadership, and technology. I'm your host today, and we're going to do something I've been wanting to do for a long time. Going to look at what the smartest people in the world said was going to happen and ask with all the evidence now in [music] front of us, what actually has happened. So, here's the puzzle. In 2025, the American economy grew at a rate of nearly 4%.

[00:01:37]Corporate profits hit record highs.

[00:01:40]Customer spending held strong by every traditional measure of economic health.

[00:01:45]Boom times. [music] And yet, payroll employment expanded by only 584,000 jobs, the weakest [music] annual job growth since the pandemic year of 2020.

[00:01:56]Job openings fell to their lowest level since September 2020. Companies were announcing restructuring plans while simultaneously posting record earnings.

[00:02:05]The economy got bigger. It just didn't need as much many people as it used to.

[00:02:10]That is a puzzle and a puzzle has a name. I'm going to call it the quiet replacement. A phenomenon so gradual, so bureaucratic invincible, or should I say visible, that by the time most people notice it, we'll have already reshaped the entire architecture of work as we know it. So, to understand the quiet replacements, we have to go back to the year 2023. In March of 2023, the economists at Goldman Sachs published a report. It was a careful, [music] measured research piece of work and it set off a panic. 300 million jobs. That was the number. 300 million jobs globally at risk of automation. The World Economic Forum said [music] 85 million jobs will be displaced by 2025.

[00:02:55]Kenzie said half of all the current work activities could be automated. The IMF [music] said around 40% of jobs globally would be affected. And then the architects of the technology itself stepped [music] forward. Sam Altman, the CEO of Open AI, said AI would eventually replace most human work. Jensen Huang, the man who built the chips [music] that power the AI revolution, said software engineers would be replaced. Dara Amoodi at Anthropic predicted [music] AI could handle the majority of cognitive tasks within years. Those people were not doomsayers.

[00:03:31]>> [music] >> These are the exact same people who are building AI. Therefore, they had more information than anyone on Earth about what or where they're coming from. So, here's a question I want to sit for with a moment. How did people this smart with this much information or with this much data get some things wrong or did it get it all wrong. But here's the thing, because the thing with big predictions is they're almost never wrong about the destination, almost always wrong about the road it takes [music] to get there.

[00:04:04]So we start with what official numbers say. Christmas is a firm that has been tracking corporate layoff announcements in the United States since 1993. They are meticulous, they are comprehensive. In 2025, they have found that American companies officially attributed around 55,000 job cuts to artificial intelligence. 5,000 [music] out of 160 million workers. That is to be around 0.03% [music] of the American workforce. Now, you might look at the number and feel relief. The experts were wrong. The robots didn't come. And [music] in a certain narrow technical sense, they are correct. But I want you to think about how company announces job [music] cuts in 2025. Companies don't say "We are replacing David with an AI." They say, [music] "We are pursuing strategic efficiencies to optimize our operational structure in alignment with our long-term growth objectives." They say restructuring. And here's a mechanism that makes a quiet replacement [music] almost impossible to measure. Someone leaves a role that AI can partially perform. The manager is handed a software license.

[00:05:15]>> [music] >> The job listing is opened, reviewed, and then quietly, without announcements, without severance, [music] with a single press release, closed. No one was fired. The job simply ceased to exist. A researcher [music] named David Shapiro ran two independent analyses of AI's real employment impact in 2025. Both [music] methods converge on the same number, between 200,000 and 300,000 jobs, not eliminated in dramatic layoffs, but quietly and permanently removed from the economy four to six times higher than what companies officially admitted. [music] The gap between what corporations reported and what actually happened is not a [music] rounding error. It is a choice. Now, I want to introduce you to a concept that I think is the most important and least discussed consequences of all of this.

[00:06:06]There's a reason that the most experienced attorneys started as junior associates doing research and document review. There is a reason that the most seasoned software architects began by fixing other people's bugs. There's a reason that the CFOs of Fortune 500 companies once spent [music] years reconciling spreadsheets at 11:00 p.m.

[00:06:28]till midnight. Entry-level work is not just work. It is training. You don't pay a young paralegal 48,000 a year to review contracts. You pay them 48,000 a year to learn how contracts work so that [music] in 15 years, when they are sitting across the table from a client during a crisis, >> [music] >> they know things that cannot be taught in a classroom. These jobs are disappearing the fastest right now.

[00:06:54]Entry-level jobs are not senior-level jobs. Aura, a legal AI startup, raised $550 million in 2025 to automate contract analysis, legal research, and litigation support.

[00:07:09]So, everything that a junior attorney would do, a paralegal would do. And Oracle announced around 20,000 job cuts while simultaneously announcing record AI investment. [music] The jobs that disappeared first were not the executive roles. They were the roles where you learn to become an executive.

[00:07:27]So, a world with no beginner roles becomes a world with no senior [music] roles. That's just the point that every economist reports that about. [music] We're not just removing jobs from the economy, we are removing the job ladder.

[00:07:42]In 1990s, there was a debate in manufacturing circles around what robots could do to a factory job. The conventional wisdom was simple. Robots do the repetitive work, >> [music] >> humans do the complex work. Robots tighten the bolts, humans solve the problem. It was a reasonable framework.

[00:08:00]It was also eventually completely wrong because the robots got better and the definition of complex work kept [music] shrinking. So, let's just think about this history constantly and when we talk about this AI agents, everything we have been discussing, such [music] as current reviewing AI outputs, the paralegal managing the contract analysis platform, the software engineer prompting the code generator, all of it shares one common characteristic. A human is still driving. Every tool, every co-pilot, every assistant requires a human to initiate the task, [music] shift the prompt, and verify the result. The AI is fast and capable and increasingly accurate, but it waits. AI agents do [music] not wait. An AI agent executes multiple step task without being asked.

[00:08:47]It coordinates with other agents. It books the meeting, drafts the proposal, [music] analyzes the contract, files the report, routes the decision. Not because a human told it to start, because the conditions were met. Data conversation about AI and jobs, [music] every government sex report, every E the V E F statistic, every panel discussion at every conference for the last 3 years was a conversation about AI as a tool. A hammer does not replace the carpenter.

[00:09:17]It makes the carpenter [music] faster.

[00:09:19]But an AI agent is not a hammer. It is closer to a contractor who works 24 hours a day, never sleeps, >> [music] >> never asks for a raise, never has a bad week, and gets measurable, more capable every 6 months. If AI is taking the entry-level jobs, what do we do? And I've been thinking about the question ever since. And the easy answer is learn to use AI, be adoptable, upskill constantly. Always felt to me someone It's like telling someone to swim faster without acknowledging that current has changed. [music] Here's what I think.

[00:09:51]The capabilities AI replacing first are the ones that are primarily about execution, processing, [music] pattern recognition, first drafts. Those are fast, cheap, and increasingly automated.

[00:10:03][music] What AI cannot do is build trust with a client who just got the bad news.

[00:10:09]You cannot read the room when the numbers are technically correct, but something feels wrong. Can't be >> [music] >> the person who says quietly, with the confidence of experience, "I've seen this before. This is not what it looks like." That is still a human skill.

[00:10:23]Every CEO I have spoken to on this channel have said a version of the same thing. [music] The future belongs to people who know when the AI is wrong.

[00:10:32]So, let's go back to Karen. Last year, her firm promoted her. She saw senior client manager, a title that did not exist at the firm three years ago. She oversees a portfolio of 200 accounts.

[00:10:43]She manages two junior staff members.

[00:10:46]She also works, by her own estimate, around 30% more than she did before the AI platform was deployed. Karen's original job is gone. The three junior positions her firm would have hired for are gone. The entry-level pipeline for the accounting position is quietly, [music] systematically narrowing. The economists who said no would change were wrong. The tech CEOs who said everything would change overnight were also wrong. The truth is something more unsettling that's happening than either prediction.

[00:11:16][music] It's the future of work is is on how Karen kept her job. She just does not recognize it anymore. So, if this episode made you think differently about what's happening, share it with someone who's navigating this right now. [music] I host AI Revolution I guess where I sit down with CEOs and innovators who are making these decisions at the highest level. The conversation doesn't [music] end there. It starts here.

[00:11:42]>> [music]