Hydra beats humans but Centaurs beat Hydra
How will we configure our orgs to partner with the technology we use/build/buy?
I love the full quote, ‘Jack of all trades, master of none, often time better than the master of one.’ So many people stop after that first clause and don’t get the full meaning of the quote. Yes, I know, I have a thing about quotes. I really feel “seen” by this quote is probably why I love it so much. Along the same lines, I also really liked David Epstein’s book, Range: Why Generalists Triumph In A Specialized World, from waaay back in 2019.
I was looking back through it and one section stood out to me as a pointer to what we might be able to engineer with AI moving into so many environments. Right at the start, Epstein lays out the history of computers beating people like Gary Kasparov at chess. What’s interesting isn’t that that happened, chess is a highly codified game with tons of available training data (sound like anything else, I can only think of about a dozen work areas…), what’s interesting is that Kasparov was open-minded enough and thoughtful enough to see the evolutionary potentials here. He grokked Moravec’s Paradox, that is that what we think of as hard for us humans, is often quite easy for computers and vice versa. So memorizing and instantly recalling thousands and thousands of opening gambits, middle games and end games, requires a lifetime of work for a human but a computer can do it with zero effort. Kasparov got this and as Epstein recounts, started an “advanced chess” tournament that paired a human player and a computer as a team. This leveled the tactical memory field and put strategy at the fore. Now players that Kasparov had trounced on their own, were playing him to a draw. Sound familiar? Like the results we’re already seeing from pairing humans and AI? Evolution wasn’t done though.
A few years past the advent of advanced chess, “freestyle chess” landed and it pitted teams of humans and computers. Now amateurs with laptops were crushing chess mainframes with human grandmasters. What happened? Turns out that just like with all-star sports teams, its often the best teams that win, not the ones with the most stars. The skills needed to “coach” and lead teams made up of humans and computers were different than knowing the most about chess. The all-star teams were known as Hydra and amateur human/computer teams were Centaurs. One of the biggest lessons here is that the Centaurs learned that what mattered most is this cyborg environment was not domain-specific knowledge but the ability to lead and direct the combined team with a thoughtful strategy that made the best use of the combined talents.
This may all sound blindingly obvious but I think there are deep implications here for multiple groups across the enterprise.
Recruiting and hiring: What strengths are you going to look for? What examples from past work will be the most relevant? How will you fashion questions that get at someone’s ability to lead a cyborg team? What will that job description look like?
Onboarding: We’ve all had onboarding plans that have say multiple lunches or coffees with new teammates (or we did in the Before Times) but how will you onboard a new hire to the tech systems they’ll not just be using but working with, partnering with? What will be the expectation gap between what they’ve been using in their personal lives versus what they’ll have access to in the enterprise?
Teams/Orgs: Hiring managers - what will you look for? How will you understand the gaps in your team’s capabilities? Once those new hires are in place, how will you rate and assess them? What will their contributions to the team look like?
Learning and Development: How will you come to know both the teams’ requirements and the skill gaps that the current and future members represent? What will your output look like beyond courses? What systems/technology do you need…will you need in the next 24 months? How do you build roadmaps and curricula that include both humans and digital partners?
All the others: Finance, IT, Facilities, Ops, - there will be impacts across all these teams. Now is the time to stop worrying about the tech and working on what the org will look like.
Look, I know we’re all wild-eyed and stunned by the daily march of AI and GenAI into places in the enterprise. That’s missing the point though. The tech will be there, the hard part is going to be figuring out how to best employ that tech in concert with our employees. This is where the hard math is, the hard decisions are - not whose job gets replaced by AI but what is the optimum configuration of ANY of our jobs. AI will impact every corner of the enterprise, even before ChatGPT, AI was almost invisibly impacting a lot of what we do anyway. Our choice, the work we can choose to do, is thinking about what that impact will look like and how to shape it to best benefit our customers and our colleagues.