All articles

Entry-level jobs are getting killed. Here's the skill that still gets you hired.

Reece Callahan·Jul 12, 2026·9 min read

Look, I'm going to be straight with you because nobody else will.

Entry-level jobs are getting killed. Not "shrinking a little." Not "changing." Getting killed. And the reason nobody's telling you plainly is that saying it out loud sounds mean, so instead you get a career coach on TikTok telling you to fix your resume and manifest harder.

Fix your resume. Manifest harder. Both are lies.

Here's the actual math and then I'll tell you what I think you should do about it, big-brother style, no sales pitch, no career-coach voice.

The stats you're not supposed to see this bluntly

According to the Bureau of Labor Statistics, new Gen Z workers are the highest share of the unemployed in nearly 40 years. Only 30% of recent grads land a full-time job in their field of study. Meanwhile 98% of HR leaders say they can't find talent, and 89% of them refuse to hire Gen Z. Read that sentence twice. You are the "talent shortage" they're complaining about, and they still won't hire you.

By 2027, a resume.org survey found that 47% of employers expect to eliminate entry-level roles entirely. Stanford's Digital Economy Lab already tracked a 16% decline in Gen Z employment (ages 22 to 25) in occupations most exposed to AI. Older workers in those same occupations kept their jobs. That's not a rumor. That's data.

Two-thirds of "entry-level" roles today are filled by workers with 7+ years of experience. 61% of full-time entry-level jobs require 3 to 5 years of experience. That is not a typo. Entry-level positions require several years of prior work. If that sounds insane it's because it is.

And 73% of underemployed grads will still be underemployed a decade later, per the Burning Glass Institute. Ten years. Not a phase.

Why "learn to code" isn't the fix this time

Every previous generation got told to grind, get a degree, and things would work out. Then millennials got told "learn to code" and it worked for about seven years until it didn't. Now the same voices are telling Gen Z to "go into the trades" as if 60 million people can get absorbed by drywall and HVAC.

The uncomfortable truth is the funnel that used to turn a 22-year-old into a competent professional is broken. Companies used to hire for potential. You'd sit in meetings, take notes, learn the culture, get promoted to something real. That funnel is gone. MIT found that 95% of generative AI pilots at companies are failing, which means employers aren't even sure AI can replace you, but they're firing you anyway because Wall Street likes hearing them say "AI efficiency." That's the actual story.

So nobody's going to hire you for potential. You have to bring skills. Actual, usable, "I can do this on Monday" skills.

What "skills" actually means in 2026

I'm going to say this plainly. Every job description in the last 18 months has quietly added "AI proficiency" or "familiar with LLMs" or "comfortable using AI tools" or some version of it. Not because HR loves buzzwords. Because managers watched half their team get replaced by three people using Claude or ChatGPT and they don't want to hire someone who has to be taught the basics.

You don't need to become a machine learning engineer. Nobody's asking you to build GPT. What they're asking is whether you can walk into their office and cut a 4-hour task down to 30 minutes because you know how to prompt properly and stitch a workflow together. That is the entire bar. Nobody's telling you this because it doesn't sound impressive enough to sell as a $2,000 course.

Here are the AI skills that actually get you hired.

  • You can take a real, messy business problem and write a prompt that produces a usable draft
  • You can chain two or three AI tools together (writing tool, image tool, spreadsheet, whatever the job needs) without breaking
  • You can explain to a non-technical manager why your workflow works, in normal English
  • You know when to trust the AI's output and when to catch its mistakes
  • You've built at least one small thing (a report generator, a research bot, an email draft assistant, an ad generator) that you can demo in an interview

That last one matters more than every certificate. When a hiring manager sees a live demo, everything else on your resume becomes noise. They stop caring about your GPA. They stop caring about the internship you didn't get. They see the demo and they think "this person can do the job."

I put mine together from a mix of YouTube tutorials that finally clicked once I had the vocabulary, a Skool community I lurked in for two months, and a short course on Mindwand that broke the whole thing into 15-minute daily lessons I actually finished. Coursera has similar material if that fits better. Pick one. Finish it.

How to actually sell yourself once you have the skill

The other half nobody teaches you. Skills alone don't get you hired. You have to be findable, reachable, and impossible to say no to.

Cold email the hiring manager directly. Not the HR filter. Find the person on LinkedIn, get their work email off Hunter or Apollo, send a 4-sentence email with a link to the demo you built. Two of my friends got hired this year doing exactly this after 100+ ghosted applications on Indeed. The application funnel is broken. Cold outreach is the workaround.

Build a public portfolio. GitHub repo, Notion page, YouTube channel, whatever. Something a hiring manager can find in 30 seconds and see "oh, they actually do the work." Most Gen Z candidates don't have this. Having it puts you in the top 5% of applicants for any AI-adjacent role.

Treat every interview like you already work there. Come in with a specific idea for something you'd build for them in your first 30 days. Say the words "here's what I'd ship in my first month." Hiring managers have never been more starved for someone who acts like they want to do work, because most candidates arrive expecting to be taught.

The blunt close

Nobody is coming to fix the entry-level job crisis in time for you. Not Congress. Not universal basic income. Not the trades absorbing 60 million people. Not the next AI regulation.

The system is broken. That is true. You still have to live inside it for the next 40 years. Sitting in the doom-scroll and waiting for someone to say "OK, we've fixed it, come get your career now" is a strategy that ends the way you already know it ends.

What actually works is unglamorous. Learn one AI skill deeply. Build one small thing you can demo. Cold email 50 companies. Keep going when the first 30 ghost you.

That's the entire play. It's not a secret. It's not viral. It's just what the 5% of Gen Z who actually got hired this year all did. Nobody who is currently making it in 2026 is doing it by grinding harder on Indeed. They're not applying more. They're building demos and hitting up the manager directly.

I know you've heard "hustle harder" your whole life and this sounds like more of that. It is not. It's the exact opposite. It's stop doing the thing everyone else is doing (Indeed applications, career fairs, generic cover letters) and start doing the two things almost nobody does (build a demo, cold email a manager).

If you want the structured path instead of piecing it together across a year of trial and error, Mindwand, Coursera, and Skool all teach some version of this. Format matters more than brand. Whichever fits the way you actually learn, that's the one.

Do the two things. Then tell me it didn't work.

Frequently asked questions

Yes, and the difference matters. 'Learn to code' was a decade-long path with a well-paved job funnel behind it. 'Learn AI' is a three-to-six-month path to being useful in almost any office job that exists. You are not competing with software engineers. You are competing with the 27-year-old down the hall who still uses AI like a search engine. The bar is that low right now. It will not stay that low for long, which is why the window is right now.

Depending on how many hours a week you can commit, somewhere between six and twelve weeks to be interview-ready with one small demo project in your portfolio. That includes learning how to prompt properly, chaining two or three tools together, and building one working thing you can screen-record. Anyone selling you 'AI expert in 7 days' is lying. Anyone telling you it takes 18 months is also lying, or trying to sell you a masters program.

Being able to build a small automation that removes 4-6 hours of somebody's week. That's it. Not 'prompt engineering' as a bullet. Not 'certified in Claude.' A one-line description of a workflow you built that saved real time, plus a link to a 90-second demo video. That combination beats every generic AI-skills bullet on LinkedIn.

Yes, more than they have in years. The reason is that Indeed is broken and every hiring manager knows their inbox is full of AI-generated cover letters. A specific, four-sentence, direct-to-them email with a demo link is now one of the most differentiated things you can send. Response rates for people I know sending them look roughly 3 to 8 percent, which is 30x the reply rate on job boards.

The people who can use AI to do more work will keep their jobs. The people who cannot will not. If AI becomes even more capable in two years, the person best positioned is the one who spent this year learning to use it, because they know how to steer the next version too. The worst position to be in is 'I skipped learning AI because it might be replaced by AI.' That is exactly the trap the middle of the workforce is currently walking into.

Keywords

AI CareerGen Z JobsHiringAI SkillsJob Market