The Effort Premium: How to Get Hired in 2026

Please stop using AI in your job interviews.

No, not never touch it. I mean the cover letter it writes for you, the company research you fake with a Claude summary, the take-home you let it design and then hand in as your own.

Everyone does it. Most don't even know the result sounds like ai, the way you can't hear your own accent.

That's why getting hired in 2026 feels impossible. It's also why standing out has never been this easy.

I want to show you the three places every interview falls apart now, and why most people are unhireable without realizing it.

Then the embarrassingly simple thing that can beat every other applicant.

In a frozen market, it's the closest thing to an unfair advantage you've got.

The structure that stopped working

Having spent the past six months interviewing hundreds of candidates for roles across sales, marketing, partnerships, engineering, interns, and product, the slop is constant, and the people sending it can't hear it, the way you can't hear your own accent. They run it through Claude, decide it sounds sharp, and send. It does sound sharp(ish), but so does everyone else's which uses the same phrasing and structure. Which means the thing the applicant thinks makes him stand out is actually making him identical.

So what is a job interview actually for, now that looking competent costs nothing?

It used to measure something. A good cover letter took an afternoon. Researching a company took actual time. A take-home cost you a weekend. Those costs were the point. Nobody cared about the cover letter itself, they cared about what the effort behind it told them about you. Because the number one thing a company hires for is effort.

People have stopped understanding that.

A fifty-year-old idea in economics explains this better than any hiring advice. In 1973, Michael Spence published a paper on what he called “job market signaling”, and it later won him a Nobel. The idea is simple. An employer can't see inside your head. They can't watch how good you actually are, so they read signals instead. Maybe that’s where you went to school, how you write, how you handle a case study, how you speak, or who you know. A signal only works if it's costly, specifically if it costs more to fake than to earn honestly. A degree separates people because coasting through four years is hard. A sharp, specific cover letter separated people because writing one took effort, and effort tracked how much you cared.

AI took the cost of that effort and dropped it to zero.

Now the polished deliverables cost nothing. And anything that costs nothing to produce tells the company nothing about you. When everyone can generate the same semi-competent output in 10 minutes, the output stops separating anyone from anyone. The market for proof collapses, and people have yet to realize it.

Somewhere between 40 and 80 percent of applicants now run their materials through AI. Only about a third of employers still trust credentials as a real measure of ability. Applications are growing roughly four times faster than the jobs they're chasing.

The three tells

There are three places this shows up in an interview, and people fail all three the same way. I’m telling you so you don’t make the same mistake.

The first is the application itself. It's the first impression you have, which is exactly why it's the most flooded. The format they taught you in school, the “I am writing to express my interest in the position." I skip it all. I don't mean that as advice, I mean I physically move to the next one, and I work at a startup company where your letter actually gets read most of the time. At a larger company yours is one of thousands. The people who post "I applied to 1,000 jobs this month and heard nothing" think they're describing bad luck. But it’s not luck, it’s the fact they’re sending the same slop over and over again.

The second is discovery. The first call is usually a fit conversation (at least for us), and half of any interview lives right here, in whether you actually know what we do. Asking Claude to summarize a company or research it for you is not research. It's the appearance of research, which is worse, because it produces a candidate who can say "I love the problem you're solving" about a product they have never once opened. Ours is free. You can sign up in a minute. 75% of applicants don’t, then tell us how much they admire the way we solved a problem they've never watched us solve. Thirty minutes of actually looking (the founders, the product, the competitors, what the team ships) puts you ahead of most of the field, because most of the field outsourced it. This used to be common, now it separates you.

The third is the deliverable. The take-home, the case study, the thing built to look like the work you'd actually do in the role. This is where it matters most and where AI does the most damage. An AI-built deck is beautiful and tells me nothing, because I could have made it myself in the time it took you to send it. What I'm hiring is your judgment, your taste, the one decision only you would have made. Worst of all, you think you can decipher ai vs non ai writing/design. But you can’t. Speaking personally, I spend 50% of my day reading ai outputs, evaluating, and writing and that builds a good understanding. Trust me when I say your professors or teachers or even “ai detectors” cannot tell. You need to do it yourself.

We even put a line in our case study asking people not to use AI, and they use it anyway, in plain sight. Either you don’t care enough to read the instructions or you have become so reliant on ai you can’t function yourself. Both are bad.

The floor nobody steps over

But in the collapse lies the opportunity. When the entire field defaults to the slop, the one person who puts in real effort is impossible to miss. AI took away the effort for mediocrity. Now going above the medium is easier than ever.

So stop optimizing the wrong thing. Applying to a thousand roles isn't a good thing, and it's a pretty embarrassing to announce. Pick the handful of jobs you actually want. Then spend the time nobody else is spending. If it's a design role, design something for the company in your initial application. Do the thirty minutes and actually use the product. Build the deliverable by hand, think the problem through yourself, hand over something unmistakably yours. It might be rougher than the AI version but it will win anyway, because polished is now free and yours is not.

A big part of unemployment in new grads is, in my opinion, because people have become overly reliant on ai, hence making them unhireable.

Use it, but after

Some might say, “but isn’t this hypocritical Gabe?”. I’ve spent two years at an AI company and here I am telling you not to touch it in an interview.

But I didn't say that, not exactly. You can use it, but use it after. Build your own answer first, then hand it to the model and ask where you're wrong, maybe what you missed or what's weak. Work backwards from your own thinking instead of forward from its. The order is the whole thing. Start from the machine and you get the average of everyone who started from the machine (which is most people). Start from yourself and the machine sharpens you instead of replacing you. Use it as an advisor, not an employee.

Interviews are testing whether there's a you worth multiplying, because that's all AI does to good judgment, it multiplies it. Run it on top of nothing and it faithfully returns nothing, just faster. Unless the role is literally about prompting these systems, in which case, fine, show me. But that's a different job than the one most people are interviewing for.

What I don't know is how long it lasts. All of this works right now because almost no one does it. There may be a point where the line between your work and your-work-with-AI dissolves completely, where there's nothing an employer can read to tell one careful person from one careless one. We're not there. We might get there. But for now it works.

Standing out takes time. Use it to your advantage. — Gabriel