Fast Company ran a piece this week by Marcus Collins arguing that the skills we have historically called soft, empathy, communication, social intelligence, are in fact the hardest to develop and the most economically valuable. The timing is pointed. The same week, The New Yorker's farewell to Hacks makes a structurally identical argument through five seasons of television: the creative partnership between Jean Smart's Deborah Vance and Hannah Einbinder's Ava works not because they are similar, but because their friction generates something neither could produce alone. Meanwhile, a new arXiv paper, "When Should AI Read the Room?" by Mathur et al., finds that public perceptions of socially intelligent AI are deeply context-dependent. People want AI to have social awareness until it does, at which point it feels invasive. The skill that is hardest for humans is also the skill AI cannot safely perform without triggering discomfort.
The Creative Partnership Economy
Hacks is, among other things, a case study in mentorship as a productive collision. The show's argument is that creative excellence requires someone who will tell you the truth you cannot tell yourself. Collins's Fast Company piece comes at it from the organizational behavior angle: the Army's 1968 distinction between "hard" and "soft" skills embedded a hierarchy that undervalued exactly the capacities that make teams function under pressure. A 2022 meta-analysis in the Journal of Applied Psychology by Heckman and Kautz confirmed that social and emotional skills predict career outcomes more robustly than technical competencies across most sectors. The market has been slow to price this in.
Why This Is a Funding Problem Too
Accelerator culture has a measurable bias toward technical founders. Pattern matching on STEM credentials is faster than evaluating interpersonal intelligence. But the companies that scale are typically the ones where someone in the room can read it. TurboFund's accelerator matching quiz is built to surface fit beyond the pitch deck, because the soft signal is often the hard one that matters. Hacks knew this. The Army figured it out in 1968. The venture ecosystem is still catching up.