What should my recent college grad do???
Hi everyone,
By far, the most common question I get after I give a talk isn't about neural networks or GPU clusters. It's much more personal: "My son (or daughter) just graduated from college—what the heck should they do?!?"
New college graduates are facing a job market that has effectively paused entry-level hiring as companies try to figure out how to move at "AI speed.” It’s brutal out there right now.
Years ago, when I was demo'ing GPT-3 for audiences, I used to tell people to steer their kids toward creative professions or STEM. Well, that advice didn't age well. Today, anyone with a laptop can generate stunning artwork, cinematic movies, and professional-grade music in seconds. Even the world of software is shifting - tools like Cursor are already writing 30% of the code at companies like Microsoft.
My thinking has evolved. If I were 22 again (if only!) and looking for my first job (much less enticing), here are the two things I'd do:
1. Learn the crap out of AI
Recent grads have a massive advantage: digital plasticity. I deliver interactive workshops for everyone from high schoolers to Executive MBA students. The younger participants pick up the technical pieces in seconds, while the mid-career professionals often struggle to hold the "toothbrush" in their other hand.
A college buddy of mine, Nikesh Parekh, believes so strongly that AI skills are the "new zero" that he built a platform called Provn (which I wrote about in Future Proof, excerpt below). It's a hiring assessment platform—much like the one I built at Evolv a decade ago—but it’s focused solely on testing AI skills and surfacing the best talent on a leaderboard.
I hopped on a Zoom with him recently to catch up and he joked that all the expensive degrees hanging on the wall behind me (see below) are going to be worthless in the not-too-distant future. Honestly? I completely agree.
If I were a job candidate, I’d be on Provn working my way up that leaderboard. And if I were a hiring manager drowning in AI-juiced resumes, I’d go straight to Provn to find the builders who can actually multiply themselves through agents.
2. Double-down on "Human" skills
No profession is safe from AI, so you have to think in terms of skills. We need to look at what machines do best—pattern recognition, repetitive precision, high-speed data processing, and cold logic—versus what humans do best.
Machines are tireless teammates, but they lack a soul. They can't navigate office politics, they can't hold a client's hand through a crisis, and they certainly can't feel the weight of a high-stakes decision.
If I were 22, I would take every opportunity to develop those human traits: volunteer for projects that require cross-functional collaboration, master storytelling to move people to action, and seek out messy problems that don't have a rulebook. They’re becoming increasingly valuable because they aren't going to be subsumed by a machine anytime soon.
That’s my advice. Pass it along to that 20-something in your life who’s feeling the heat. And if you want to dig deeper, feel free to pick up a copy of Future Proof where I expand on these ideas in much more depth.
Best,
Dr. Michael “House” Housman
The shift is already here. AI skills have quickly become the baseline for landing many types of jobs and keeping them. Employers feel the mandate, and they’re pushing it down to employees in the form of hiring requirements, performance evaluations, and even day-to-day expectations.
Microsoft ran a big survey recently, and the results paint a pretty clear picture. Notably, 77% of leaders said early-in-career talent is going to get bigger responsibilities because of AI. Seventy-one percent said they’d rather hire a less experienced candidate with AI skills than a more experienced one without them.
And 66% of leaders flat-out admitted they wouldn’t hire someone who didn’t know how to use AI.
A friend of mine, Jepson Taylor (another data-slash-mad scientist), is leaning into this reality hard. He’s building an AI startup, and he wants the very best talent on his team. He decided he doesn’t care about resumes, interviews, or pedigree. Those things are all gamed at this point, anyway. He doesn’t care where you went to school, your gender, race, or even if he’s ever met you in person.
His one requirement is to show him you can actually use AI to get things done. In fact, he posted this on LinkedIn: “If you, some random person on the internet, can demonstrate that you can text from your phone and have a Cursor agent do work and send back a text-friendly summary… congratulations, you have an internship at VEOX.”
That’s it. Do the thing, and you’re in. If you can’t, no hard feelings, but you’re not joining the team.
To me, it sounds insane and brilliant all at once. But more importantly, it feels like a preview of where the job market is heading.
Another friend, Nikesh Parekh, is taking this one step further. He’s building a company called Provn, and it reminds me a lot of my days building HR technology. Back then, we tested knowledge, skills, and abilities relevant to a role and gave companies a score to help them make hiring decisions. Nikesh is doing the same thing, but his twist is assessing AI fluency at scale. He’s building assessments that measure how good you are at prompting, iterating, and collaborating with AI tools at the start of the recruiting process, before and eventually in place of resumes and interviews.
The idea is that employers will soon use these scores just like they use SATs, GMATs, or hiring assessments. Instead of asking if you can code or if you’re a good fit for company culture, they’ll ask, ”How good are you at using AI to multiply your output?”