The pilot owns the mission.
Purpose, values, risk tolerance, accountability, and the final decision remain human responsibilities.
Pilot judgment • RIO intelligence • Bayesian updating
Cockpit Coherence
Goose Never Dies.
A decision framework for keeping the human in command while a second intelligence helps build, challenge, and continuously update the operating picture. The framework draws not only from real pilot/RIO architecture, but also from the cultural imprint of the Top Gun films and what they revealed about trust, ego, teamwork, and mission-first thinking.
The core idea
AI should not fly your airplane.
Most conversations about generative AI swing between two bad extremes: surrender judgment to the machine, or fear the machine and refuse the leverage. Cockpit Coherence takes a different position. The human remains the pilot. AI becomes the second seat.
The RIO improves the picture.
AI can research, compare, remember, pattern-match, challenge, synthesize, and maintain situational awareness at a scale the pilot cannot match alone.
The vector is never sacred.
New evidence changes confidence. Contradictions are not annoyances to suppress. They are signals that the operating picture needs to be updated.
The rule
The mission outranks the original plan. Evidence outranks ego. The pilot keeps the stick.
Why the F-14?
Because the second seat is visible in the architecture, and Top Gun made people feel it.
Cockpit Coherence does not use the F-14 because old technology is automatically better than new technology. It uses the Tomcat because the aircraft makes the human relationship impossible to ignore: pilot and Radar Intercept Officer, distinct jobs, different information, one mission. The Top Gun films also helped crystallize that relationship in the public imagination, which is part of why they belong in the origin story.
The F-14 was a U.S. Navy carrier fighter built around a two-person crew. The pilot flew and fought the aircraft. The Radar Intercept Officer managed a sophisticated tactical picture, radar information, weapons-system tasks, and threat awareness from the second seat.
That division of labor is the metaphor. A strong pilot does not become weaker by using a strong second seat. The crew becomes more capable when the roles are clear, trust is earned, disagreement is useful, and both seats stay oriented to the mission.
The original Top Gun helped fix the F-14 in Eric Steele's imagination after his own Air Force pilot path closed. Top Gun: Maverick later reinforced the same lesson from the other direction: old hardware in the hands of a disciplined crew can still outthink more advanced systems when judgment, teamwork, and nerve stay coherent.
The F-14 is not the point. The crew relationship is.
Cockpit Coherence asks what happens when the second seat changes form, but the need for a second intelligence does not.
The operating stack
Five layers inside one coherent cockpit.
Cognition
How each seat perceives, remembers, reasons, recognizes patterns, and generates possible explanations.
Convergence
How separate streams of experience, data, intuition, research, and machine analysis come together around the mission.
Calibration
The Bayesian layer. Confidence changes as evidence changes. Beliefs are updated instead of defended for pride.
Judgment
Information does not make the decision by itself. The pilot weighs risk, context, uncertainty, consequences, and mission.
Coherence
The shared operating picture becomes stable enough to act, while remaining open enough to be corrected by reality.
The hidden danger
False coherence.
Agreement can feel like clarity and still be completely wrong. A pilot and RIO can reinforce the same bad assumption. A leadership team can share the same flawed data. A human and AI can build a beautifully consistent answer on a false premise.
That is why Cockpit Coherence is not a philosophy of agreement. It is a philosophy of reality-tested coherence. The cockpit must remain exposed to disconfirming evidence.
Coherence without calibration becomes certainty. Calibration without coherence becomes chaos.
The origin
A cockpit that started long before AI.
At 19, Eric Steele was a high-school dropout trying to decide what came next. An Air Force recruiting piece arrived in the mail. He responded, walked into a recruiting office in Wichita, and told the recruiter that the possibility of becoming a pilot was what had caught his attention.
The first answer was simple: take the ASVAB and see what the scores say. Steele's recollection is that his scores changed the conversation quickly. The recruiter told him his aptitude placed him in pilot-eligible territory and that finishing his high-school education could be addressed after enlistment.
Then came MEPS. Steele measured roughly 6'6". He recalls being told that the pilot path was closed because the applicable standing-height limit was 6'5".
The flying career disappeared. The fascination with the mentality did not.
Then the movies arrived. The first Top Gun came out the next year and helped turn a personal disappointment into a deeper, enduring study of pilot mentality, crew trust, and cockpit psychology. Decades later, Top Gun: Maverick only sharpened the framework by putting an older F-14 back at the center of a mission solved through human skill and second-seat coherence.
The present mission
How do you actually use generative AI?
Cockpit Coherence is Eric Steele's answer to that question. Do not worship the tool. Do not panic about the tool. Build a working relationship with it.
Train the second seat on the mission context. Use it to widen the radar picture. Let it challenge your assumptions. Force it to show contradictions. Keep updating. Keep judgment human.
Then fly the airplane.