AI learns what an organization is designed to teach it.
Teaching an intelligent system what works is good. Teaching an intelligent system what works and what doesn’t / hasn’t worked is great.
At last week’s Primary Venture Partners Summit, during his outstanding conversation with Patrick O'Shaughnessy, Etched co-founder + CEO Gavin Uberti described the job he hires 40-year veterans aka “legends” to do.
Say don’t do that. We tried it. Here’s exactly how it failed.
For example, instead of 40 years, take an industry veteran with 30+ years of experience and their entry level AI native colleague who just joined their team.
AI can’t make one of them younger or the other older, but it can make both of them younger and older, at the same time.
E-shaped is the new T-shaped.
*link to AIWA TOT #09
AI can make anyone broader by compressing learning cycles and giving them capabilities they previously would have had to hit the 10,000 hour rule to acquire.
We’ve seen the results of pairing an AI native engineer with an operator in a non-technical functional role.
*Link to AIWA TOT #27
Pairing sets the floor but when you think of AI as the medium between two people, 30+ years apart, you take the ceiling off the top of your organization.
The second order effect of the pairing is that it becomes the capture mechanism. The Intelligence Layer is the recorder.
*link to AIWA POV “Every Company Will Have An Intelligence Layer”
Instead of focusing on change management, you optimize for outcomes generated and iterate accordingly.
A veteran becomes E-shaped in the directions that matter the most. Without retraining.
The junior person’s intuition compresses, interrogated and accelerated by the system. Shaped by the veteran aka legend who helps them see around corners via hard-won, lived experience.
The human in the loop. The human on the loop.
Translating the stuff you can learn but, historically, can’t teach.
Now, there’s an Intelligence Layer for that.
Watching a veteran get broader is easy. Measuring whether the junior’s intuition actually compressed is the harder problem.
The enterprise’s unsolved problem isn’t how well the next model scored on an abstract exam. It’s the deployment overhang.
AI agents need to know how things relate, not just what they are called.
Not just what happened but how they happened and why.
A concept core to your value chain viewed by ten different stakeholder groups has subtle nuances specific to each group. Those nuances need to be codified.
Enabling the intelligent system to meet people where they are. Viewing the world through their lens.
At the Primary Summit, Rogo founder + CEO Gabriel Stengel put a number on it.
Roughly 99% of the data in a private markets deal never gets written down. And that unwritten 99% is what builds intuition.
Think knowledge graph. Think ontology.
The solve? A junior, a veteran and an Intelligence Layer.
The system’s enterprise intuition compounds.
*link to AIWA TOT #34
Raising the floor while taking the ceiling off.
Humans + Machines. Never Humans vs. Machines.
