Mark Zuckerberg hand-delivered homemade soup to OpenAI researchers’ front doors in an attempt to recruit them. Elon Musk pledged to personally read every SpaceX AI application—no résumé required; pedigree optional. When the world’s richest founders are cold-calling engineers with soup and big-figure job offers, it’s clear the AI talent war has reached a new, absurd level.
But Meta and SpaceX aren’t the only ones taking outlandish measures to attract talent in the tightest hiring market in tech history. The median base salary for startup engineers is $200,000, a 25% increase since 2022 and a far cry from the equity-first model that startups have relied on for decades.
Artificial intelligence roles are taking a long time to fill, and there are plenty of them. AI, machine learning, and data science postings jumped 163% in 2025 from 2024, yet 72% of employers say they’re struggling to fill roles. The challenge isn’t just supply. It’s also a hiring strategy issue. The old model of a charismatic founder with a vision—and often no product yet—is out. AI talent wants to see hard proof, not a pitch deck.
What AI talent responds to
I recently published research on how founders and AI specialists interact on CoffeeSpace, a platform focused on helping startups build early teams from cofounder matching to recruitment. The matching works like a dating app; users send invitations the same way you swipe on Tinder or Hinge. Except instead of singles searching for a match, founders and engineers decide whether to accept or decline an invitation to start a conversation. Of its roughly 25,000 users and 1 million swipes, we narrowed the analysis to 6,868 people involved in AI-relevant matches—357 of whom qualified as AI talent— across about 160,000 decisions.
I found a disconnect between what founders think AI talent wants, and what AI talent is actually looking for.
Founder ambition and equity offers aren’t enough to pique a candidate’s interest anymore. Instead, AI specialists want to see momentum. Founders working full-time on their project were 20% more likely to have their invitations accepted than those still in the exploration phase. But momentum doesn’t require a finished product. A prototype and a handful of early users can signal that an idea has traction, which AI specialists want to see.
A PhD, an MBA, or a stint at Meta or Google used to matter. Today, they are weaker predictors than most founders assume. AI professionals are seeking out founders with nontechnical backgrounds. Founders with legal experience saw 18% higher acceptance rates; those with sales experience saw a 10% lift.
The reasoning tracks. A founder with a legal background can lower perceived risks around compliance, IP, and contracts. A sales background demonstrates that a founder knows how to get products in front of paying customers, not just investors. Any professional experience outside of tech can signal to talent that this founder will let them focus on building, rather than reviewing investor decks.
Of course, AI specialists also make decisions based on compensation, equity, and remote work options, but later in the process than founders might realize. Perks like these don’t create interest on their own, but they can remove friction for candidates who are already in the negotiation stage.
What founders can do about it
Most founders can’t compete with Zuckerberg’s theatrics or Musk’s budget. The good news is that the data suggests they don’t need to.
People looking for top AI talent can make progress visible. Publishing a build log, sharing a demo video, or putting a working prototype in front of even a handful of users before recruiting begins can demonstrate momentum.
If a team is heavy on technical expertise, it may help to hire people with legal or sales experience before stacking the team with another engineer. Having a coworker with experience in these fields can do more to attract AI talent. And save the perks conversation for later, as flexible compensation and remote work options close deals rather than opening them.
AI pros want a viable prototype more than a soup delivery. They’d rather be left to work on the product than get sucked into strategy meetings. They want committed founders, not a Steve Jobs-like visionary running on charisma alone. In short, they want proof that what they’re building is real, and that the people building it are grounded. No amount of soup can make up for that.