LLenny's Podcast
← All frameworks
LeadershipAnton Osika (co-founder and CEO)

Generalist-With-A-Superpower Hiring

Hire people who care obsessively, have a generalist brain, and one absolute superpower — then trial them for a week.

Difficulty
Advanced
Time to result
~weeks to results
Steps
4
Confidence
85%

A hiring model for small, high-leverage teams in the AI era: prioritise obsessive care, a generalist brain that learns any skill fast, and one dimension of absolute superpower — then verify it through a multi-day work simulation rather than interviews alone. Use it when a tiny team must cover many skill sets per person.

Origin

Anton runs Lovable with 18 people (12+ writing code) growing >$1M ARR/week. He argues being a generalist matters far more than it used to, and that he'd 'obsess about getting as many skill sets as possible for each person' while still demanding one superpower.

Core principles

  • 01Care, preferably obsession, is the strongest predictor of fit
  • 02Raw cognitive capability is the strongest correlate of performance
  • 03Want a generalist brain plus one absolute superpower
  • 04Value a startup mindset: move fast over process; own the business over guarding your craft
  • 05Verify with real work, not just interviews

How to run it

  1. 1

    Screen for care and ownership

    Ask what people have done before and look for genuine obsession over past work and details, plus caring about the team and users — not just showing up for a job.

    Pro tip People who cared deeply about a past project tend to keep caring; probe the technical details they went into.

  2. 2

    Test raw thinking on an unorthodox problem

    Present a very hard, slightly unusual problem they haven't seen before and watch how they reason through it.

  3. 3

    Look for generalist range plus one superpower

    Seek a brain that quickly learns any skill (architecture, design, product taste, talking to users) while being exceptional in one dimension.

    Pro tip At Lovable the prized superpower is cramming maximum output from LLMs and knowing the full parameter space to tune the product.

    Watch out This profile is genuinely rare — 'easier said than done' to find people strong across all of it.

  4. 4

    Run a work simulation

    Have candidates join the team and do real work for at least a day, often a full week, before deciding.

    Pro tip The trial doubles as a culture and pace filter — the intense environment self-selects.

In the wild

The Shackleton-style job post

Lovable's job description warns of long hours, high pace, 'high urgency under AGI timelines', honor in success, and 'those seeking comfortable work need not apply' — a deliberate filter so only people who want that intensity apply.

A strong self-selection filter that attracts the obsessed and repels the comfort-seeking.

Common mistakes

Hiring passengers

People 'here for a job' as a passenger on the journey lack the care and ownership that carries a tiny team a long way.

Interview-only decisions

Skipping the work trial means you never see how someone actually contributes to the team's pace and dynamics.

Is it for you?

Best for

Founders building small, high-intensity, high-leverage teams where each person spans many roles

Not ideal for

Large organisations hiring for narrow, specialised roles with heavy process

From the transcript

the most important thing is that people care a lot

41:30

have this generalist brain that quickly learns any skills but be super super good in in one dimension

42:30

I pretty much always have people join the work simulation for at least a day often a full week

44:00

From the episode

Building Lovable: $10M ARR in 60 days with 15 people

Anton Osika (co-founder and CEO)