Position A — Hire generalists with one deep dimension
- Operator: Anton Osika (Lovable)
- Card: In an AI-native team, hire generalists with one deep dimension, not specialists
- Claim: In an AI-native team, hire generalists with one deep dimension. Specialists are out of phase with what AI absorbs.
Position B — Hire spiky candidates with real strengths and real weaknesses
- Operator: Brian Halligan (HubSpot)
- Card: Hire spiky candidates with real strengths and real weaknesses, not the un-weakness average
- Claim: Hire spiky — real strengths and real weaknesses, not the un-weakness average. Implies depth/specialty over balance.
Position C — Hire across three archetypes, not one
- Operator: Jenny Wen (Anthropic)
- Card: Hire across three archetypes: block-shape generalist, deep specialist, craft new-grad
- Claim: Hire block-shape generalists and deep specialists and craft new-grads — the team needs all three.
Conditions distinguishing them
- Team size: Osika's "generalist" works for an early-stage AI-native team where each hire covers wide surface. Wen's three-archetype mix works at scale where coverage is collective. Halligan's spiky frame works for senior leadership hires regardless of stage.
- Function: Generalists fit AI-native eng/product. Spiky fits exec/leadership. Three-archetype fits design (where craft heritage is real and not absorbable by AI yet).
- Implicit weight on AI substitution: Osika treats AI as absorbing specialist work; Wen treats craft/specialist work as still differentiated; Halligan is agnostic.
Resolution / synthesis
The contradiction surfaces on whether deep specialty is high-leverage:
- Osika says no (AI absorbs specialist work).
- Halligan says yes (real strength = real depth).
- Wen says it depends on archetype mix.
Reconciliation:
- For a small team (<10): Osika's generalists.
- For an exec hire: Halligan's spiky.
- For a large discipline (design, eng, PMM at scale): Wen's three-archetype mix — generalist + specialist + craft.
The genuine disagreement is on how much AI absorbs specialist craft. Wen and Halligan implicitly bet "less than Osika thinks." Osika is in a fully AI-native context where the bet pays. The other two operate in environments where craft compounds.