A Cube Inside a Sphere, Right Between Two Super Hornets

Lex Fridman starts with *Top Gun: Maverick*. Graves, a former Navy F/A-18 Super Hornet pilot with combat-lead, landing-signals-officer, and rescue-mission-commander time, says the flying was accurate. What films rarely catch is the chess game underneath. The canyon strike is a high-threat ingress: stay low, hide in clutter, and stay out of the radar window of an integrated air defense system whose sensors and shooters are deliberately separated. Take out the battery that is shooting at you and another radar can still complete the intercept.

Ingress, he notes, is a choice. Egress after weapons come off is not, so tactics and aggression change. In a less scripted fight, one jet down low acts as a forward air controller, joining other players onto the right target. The back-seater, the weapons systems officer, or Wizzo, works the radar, the radios, and the weapons while the pilot stays alive at about a hundred feet. Graves is blunt about psychology in that moment: there is no time for fear or self-reflection in the cockpit.

The more interesting advantage, he argues, is randomness. A tactic in a manual has to be repeatable, trainable, and lethal enough, which means it is not the most lethal version. Put machine learning in the loop and the war space can look stochastic from the enemy's side: a new solution for each scenario. At that point, Graves says, even friendly fighters might not be able to join the fight, because it would be unsafe for anything else. Knowledge would stop being the variable that decides the engagement.

Dogfighting on screen is mostly fiction. Real merges last milliseconds; a correctly flown fight looks like blue sky and a dot. Close-in guns may still return if electronic warfare dumps missiles out of the sky. Asked for the best fighter ever made, Graves declines the romantic F-14 answer. The F-35 is not the best knife fighter, and it does not have to be. Information, not the last turning fight, is what he thinks wins.

Mach 10 ejections in the movie, Graves wants on the record, are nonsense. Crossing Mach 1 in an F/A-18 feels like crossing a highway speed limit. He has gone about Mach 1.28, usually in a descent in full afterburner, because the Super Hornet is heavy. Pull the throttles from afterburner back to military power and aerodynamic drag throws you into the straps, almost as hard as trapping on the carrier. You feel acceleration and deceleration, not the number on the Mach meter.

Hypersonic human flight is not, as far as he knows, a tactical fighter tool. Hypersonic missiles matter because they shorten the kill chain: locate, identify, authorize, employ, assess. The old fighter loop of observe, orient, decide, and act was a way of getting inside that chain. Graves treats hypersonics as a detection-and-destroy timing problem. You are simply leaving the defender less time to finish the job.

Bird strikes are not rare color for a flying sequence. Graves has hit birds. He knows of a turkey vulture that came through the canopy, knocked the front-seater unconscious, and forced the WSO to eject both of them. "It's like a bowling ball going 250 miles an hour." Navy fields check bird conditions; some have bird radars; some cancel flights; some keep EPA staff to move migratory birds off the runway. If a bird is in front of you, always try to go over it. A startled bird folds its wings and drops, and a dive puts you into it.

Graves joined the Navy in 2009, chose the pilot path so he would not get stuck in another job, and lucked into the one jet slot in his class. That meant tailhook: T-45 Goshawk, then Super Hornet. The jump from a 600-horsepower prop to a jet of roughly 15,000 pounds of thrust is the biggest in Navy training. Radios, gear, and decisions all happen faster. His first T-45 hop was a formation flight in the back seat. On the way home the landing gear would not come down.

They headed toward a controlled-ejection area, cycling the stick to unload and reload the jet. The gear came down halfway on its own; they trapped as a precaution. Graves says he was too young to feel the seriousness. He switched from observer to crewmember and ran checklists. Five or six years later, as aviation safety officer in the same squadron, a student on the same syllabus ejected after runaway trim inverted the jet at about 600 feet. They punched at 300 to 400 feet, slightly outside the envelope, and walked away. Two more ejections followed within months.

Ejection, as described to him, is about 250 g for a few milliseconds, possible only because of inertia in the blood, then 40 or 50 g to clear the jet. You may wake up in a tree, still falling, or in the water. He has never done it, and he is glad. Formation, by contrast, is a long stress he knows from both seats. It is all manual: stare at a visual triangle on the lead's wing, sometimes for hours, sometimes in weather a few feet away. In a four-ship, Dash-4 sits at the end of the whip. Each jet flies the errors of everyone inboard, so the last jet moves the most.

As an instructor, Graves could fix a student's formation with three words: "Wiggle your toes." The near leg had locked on the rudder. The body was trying to get away from the other airplane. In cloud, the lead becomes your gyroscope; you cannot break out to the instruments without hitting him. He has thought about automating formation and then asked why the formation exists at all. Tight parade is for weather and for seeing the lead. In a tactical fight, the useful formation might be no formation, or a random one, with a human still adjusting the dance.

Graves is an outlier among fighter pilots. The romance of inverted earth and clouds did not pull him in. He came from a small town and a small school and wanted to stand next to the edge of human engineering. Mastery came later, when the jet felt like an extension of his body. He had wanted test-pilot school and an astronaut path; a broken foot closed that window. Going to the instructor tour instead, he says, is why he could talk publicly about UAP. Staying in would have made the conversation impossible.

Catapult shots are a pocket watch. You are a gear in the machine; press the right buttons. Landing is another animal. A perfect pass means flying your head through a one-foot box, plus or minus about five knots, judged by angle of attack. At touchdown you go to full power as if boltering. Bring the throttles back early, as Graves once did as a new guy, and you have committed a serious error called ease-guns-land. The crash of a good trap, when it is good, is still one of the best feelings in the job.

The "ball" is a mirror with lights, a lagging indicator, and the engines lag too. The landing area is moving and angled, so every lineup correction is a bank that costs lift and must be bought back with power. Descent rate rises from about 200 feet per minute to 650. Then the burble, the downwash off the island, pulls the jet down just as the wings cross the deck. Graves has seen a hook strike the ramp. He puts the blame in the cockpit, while noting that new pilots and complacent ones make the same mistake: trying to recenter a high ball in close.

The worst call is "99 taxi lights on." Weather is so bad the landing signals officer cannot see you either. You turn the taxi light on so he can talk you down with a radio that looks like a 1980s telephone, into a zero-zero landing you never see. You hit at about 1,600 feet per minute, into darkness or into white light. Night recoveries ten or fifteen miles aft, lights dimmed, cloud deck blocking the stars, feel like flying with your eyes closed. The jet could be inverted and you would not know. Graves calls it almost meditative until four miles, when the light grows and he has to snap back into the body that still has to land.

After a 2012 Afghanistan cruise, VFA-11, the Red Rippers, upgraded Super Hornet radars from the mechanically scanned APG-73 to the AESA APG-79. Graves compares the leap to analog versus digital. The next day, in the same working areas off Virginia Beach, the new radar showed objects the old one had not. A crew might fly the old set in the morning and the new set that evening. They first assumed software, calibration, false tracks. New gear always has bugs, and they were used to that.

There is no picture of the target, only symbols. One display is a forward view of what the radar sees; the situational-awareness page is a God's-eye view that fuses own-ship radar and data link, with the jet in the center. Graves tells the story from that SA page. Each track has a target-aspect line. On a normal airplane the line is steady and turns smoothly. On these objects it jumped around the full 360 degrees, as if the radar could not decide which way they were going, or as if they were stationary.

They sat still at 15,000 or 20,000 feet, and at 500 feet, with wind. In that volume of air, Graves says, everything goes with the current, like boats in a sea. Anything that does not go with the wind is immediately strange. A consumer drone in 30 or 40 knots might look still from miles away. It cannot do that and then go 0.6 to 0.8 Mach for four hours. He would see circular or racetrack patterns, or a push east, typically subsonic for as long as a sortie lasted, occasionally 1.1 or 1.2 Mach. The paths were more erratic than a human hand-flying, and nothing like an autopilot's locked line.

Radar could still have been an electromagnetic ghost. Crews slewed the ATFLIR targeting pod, an infrared camera meant for air-to-ground work, onto the lock. There was IR energy coming from that piece of sky. Graves estimates 80 to 90 percent of the squadron had seen the tracks by then. Neighboring squadrons with the same radar saw them too, and they warned each other. People started flying by to look. Most of the time they saw nothing. Jets go by fast; the object may have descended, or may not have been there.

Then two Super Hornets entered at the published entry point, one of the busiest pieces of sky on the East Coast. An object sat at that altitude and that geographic position and went between the two jets. Graves does not think it charged them; he thinks they flew through its location. He suspects their radar was down, which would explain why they did not avoid something the squadron already treated as physical. They described "a dark gray or black cube inside a clear translucent sphere," the apexes of the cube touching the inside of the sphere. It was a perfectly clear day.

They canceled the rest of the event, still in their gear, and told the skipper they had almost hit one of those damn things. That was the watershed. Balloons, drones, and radar bugs were no longer enough. If you can hit it, someone has to own it. They filed hazard reports into naval aviation's safety network, a collection system, not an investigative one. Graves hoped a classified program they did not know about would get the message: you could kill somebody. There was no response. He did not think of the objects as a shooter. An F/A-18 going down to one of them would make little tactical sense except as fear. What did scare him was the intelligence problem: years of crypto, tactics, and electronic warfare emitted in the same patch of sky, collected cheaply, forever.

Graves sorts the possibilities into buckets. One is U.S. technology tested in the wrong sky, which he finds institutionally plausible given how stove-piped the military is. After the hearings and public statements, he thinks that explanation is now worse than the objects: if this is still a communications failure, the country is in bigger trouble than a UAP mystery. Another is a foreign power, not only China or Russia, and not only adversaries. A third is a non-government entity, which he finds unlikely but will not drop. A fourth is poorly understood natural phenomena. He allows that for some cases. He does not think it covers the East Coast objects, the Roosevelt Gimbal events, or the Nimitz Tic Tac.

If a competitor had this, he would expect remnants in the commercial economy. Technology that advanced should leak into other military uses and into products. A leader might lock it in the vault. A second-place power, he argues, would take bigger risks and show its hand. He has tried to steel-man a drone-swap operation that replaces airframes when nobody is looking, to explain the endurance. The infrastructure required, he says, collapses the idea. Night-illusion training, which aviators get in hours, does not cover objects that radar, infrared, and eyeballs all find in daylight.

He dislikes the word "alien." It smuggles in a distant planet, interstellar transit, a particular body. He prefers non-human intelligence, or simply: is this demonstrating intelligence? Start with one object: kinematics, energy source, g-forces inside. Step out: interaction with fighters, with weather, with each other. Off the East Coast the objects often appear in multiples, so swarm behavior might reveal sensors and areas of interest. A single video frame cannot do that. The irony is that the best sensors sit under a military umbrella that will not easily release the data, even internally. That is why he has been working with the American Institute of Aeronautics and Astronautics to stand up cross-domain teams for when the data comes.

Graves sat in a House Intelligence subcommittee hearing chaired by Rep. Andre Carson. From his seat it felt a bit disingenuous. Members advertised a love of science fiction and talked around the subject. They did not treat it as a possible foreign tactical threat. He grants that their job is not exploratory science, and he wants them on the intelligence problem. His worry is that they played a role they are not suited for: doing science.

The Pentagon's All-domain Anomaly Resolution Office, AARO, is a step he calls necessary because the previous one did not happen. Its predecessor, the Airborne Object Identification and Synchronization Management Group, was "a confusing mess of words" designed, in his view, to make the topic unpalatable. AARO at least opened a Twitter account and said it wanted to talk. Classification will hit that wall. He likes that director Sean Kirkpatrick comes from a research background, and he gives the same nod to David Spergel's NASA study. Small doors, opened after a lot of negotiation.

Public pressure matters more than it used to, not because Pentagon leaders find the topic toxic in themselves, but because they spent careers in a culture that did. Taking a stance is against their nature. Graves says his own speaking-out never felt like a bet on aliens. He was a Navy-trained aviation safety officer asking about a problem that was still happening. About eight months before the conversation, a former student, now in an East Coast F/A-18 squadron, called. Weather was too poor for dogfighting. The senior aviator suggested a UAP hunt. They went, and the student physically saw one. "Hey sir, I saw a cube and sphere. They're still out here," years later. Young pilots can talk now. Reporting still does not become investigation.

Of the Navy videos, the Nimitz Tic Tac is the one Graves cannot break. He has tried the thought experiment of thousands of drones in clever relays. He cannot get there unless the pilots are lying, and he will not take that exit. He knows David Fravor, considers him a friend, and has "zero reason to disbelieve anything he says." What interests Graves more than the airshow kinematics is the water. If he had advanced technology, he would operate partly underwater, because you can hide a lot there. From a fighter, the ocean is a black box. There is no cross-domain tracking: a radar contact at 40,000 feet cannot be handed to a system that follows it to 200,000 feet, let alone under the surface.

He was airborne the day Gimbal was recorded. The WSO on the tape is a close friend who later transitioned to pilot under Graves's instruction. It was a workup, a high-stress month of almost wartime ops, people peeling off ones and twos back to the boat. The crew thought the contacts were range aggressors trying to penetrate. Then they got FLIR. The object looks like two plates with a small funnel of IR energy top and bottom, no exhaust, no skin, no outgassing. On the SA page, offset north, four to six more objects flew in a loose wedge, not a stiff autopilot formation. They turned sharply, still with a radius, scattered in the turn, then tightened again on the reciprocal. While that happened the Gimbal object reversed to follow.

Later analysis suggested the Gimbal object climbed, so the "instant" turn was vertical. Fighters do not change direction in the vertical if they can help it. It burns gas. A surveillance platform that wants to loiter does not climb 500 feet every turn. They were not recording these things all day. They put them in a safety bucket: objects over there, do not go near them. Lex observes that a busy day of tactics and fuel math trains you to say there is weird stuff over there and keep working. Graves says that is exactly right, and it is why he did not give the problem its due until he left the fleet for an instructor tour with time to think. On the day Gimbal was taped he watched an admiral come into the intel space, look at the film for five or six seconds, make a small sound, and walk out. Graves's read: he had seen this before. He later heard the admiral reported it toward the Pentagon in real time, which is as close as Graves, as an aviator, ever got to confirming that AATIP, or something like it, had a place to call.

Asked about Mick West, Graves splits the method from the stance. Building models is needed. "Trust but verify should certainly be the mantra." What he rejects is starting as a debunker, which tells him the analyst is driving toward a conclusion. For years that was a safe business: telling people they had not seen aliens. The tactic he cannot accept is discarding what the aircrew said. Taking the human out of the analysis, he argues, is a fallacy when the first data are human observations.

On night-vision goggles over the ocean, the sky is crowded with stars. Graves thinks it would be statistically silly to assume we are alone, while conceding we might be early, rare, or unrecognizable. He worries more about anthropomorphizing. Popular culture has trained people what an encounter is supposed to look like, and that closes off what they might actually see. A mature intelligence, he suspects, would first show mastery of matter: treating material as information, designing what nature does not make, getting past the materials wall that still blocks compact fusion. Then fields: propellantless movement, even speculative coupling to a planet's gravity. "We can only throw so many rocks at the back."

That is the work he now does at Quantum Generative Materials, GenMat: bridging classical simulation and future quantum computers so you can design materials from the atoms up, with machine learning in the loop. The company was preparing to launch a hyperspectral satellite to look back at Earth for deposits, using that chemistry. Asked whether the U.S. government is hiding spacecraft, he says he has no information, and if he did he probably would not share it. His heart tells him something is going on. He has no evidence. He also wonders whether secrecy is less about hoarding technology than about a shock to the socioeconomic system, and whether a more anti-fragile public, with faster information, could hear it now. The 2017 videos, as he understands it, were already out. The Pentagon could lie or admit. Culture made lying uglier.

If we met a civilization thousands or millions of years ahead, he thinks they would have learned that cooperation outproduces conflict. A planet working as one unit might approach us planet-to-planet, not person-to-person, assuming any intelligence would already know that working together is better. His heart tells him that past a certain point, love, care, and the desire to collaborate are more powerful than the technological progress war brings. Lex says he hopes so too.

Graves's favorite idea in collaborative autonomy is still randomness: a tactical envelope the enemy cannot predict, plus simulation that designs boutique hardware for a single battle instead of one big system for every war. On manned-unmanned teaming he wants the human at the highest possible level. He wants to say that building should go away, not to place vehicles, count fuel, or know the loadout. He wants a cloud of capabilities that handles defenses and electronic warfare and prosecutes to the commander's required probability.

Lex pushes the next order: fully autonomous drones told to clear terrorists from a city. Graves does not start from the fear of a wider swath of destruction. Assuming just and moral leadership, he is an advocate for more autonomy because precision should mean fewer effects you do not want. He knows the answer is unsatisfying. He has seen a truck bomb hit a school bus in Afghanistan. Lex, having been in Ukraine, says the murder of civilians, of schools and hospitals, has a distinct unethical character, and that precision matters if the leadership is ethical. The question underneath is still who is making the decisions.

What stands out to Graves in Ukraine is not a particular airframe so much as commercial drones with weapons strapped on, and war on Twitter. Ten years from now, he says, autonomous package UAVs will be everywhere, and any one of them can be a weapon without security designed in. Lex notes civilians buying DJI airframes on their own salaries to fight. Graves thinks that kind of access can turn a land war into a quagmire and might even demotivate aggression. Will there always be war? Until we are past resource limits, and past the lizard brain, he thinks yes on this world. More than one world might be a valve.

He did not hate the people on the other side of his missions. He was also never in a true kill-or-be-killed merge. Tactical aviation, from high up, even carries a kinship: you know the training, the personality, the hundred-million-dollar jet that did not take off because you made someone angry. He calls that a luxury, a castle in the sky. Everything looks small from above. As autonomy abstracts killing from sword to gun to remote drone to a strategic order, Lex's fear is that people without a voice, without a Twitter account, the ones who only need to put food on the table, disappear from the calculation. Graves does not argue.

Advice for the young, whether they love jets or stars: do not fear looking foolish. Graves spent years treating other people's laughter as a signal to stop. It does not matter if you are good yet. Find something you can apply love and care to; you will get better, and the world will too. Lex adds that this is not only for the young. The older you get, the more you are expected to have it figured out, and the more a beginner's mind is worth.

The hardest stretch of his life was not combat. It was getting out. Fighter culture plans the next paycheck and the next city, not the mind. Studies, he notes, show reduced adrenaline production in fighter pilots for years after they leave. He had anxiety because he did not know how to fill the space that the job had occupied. He would tell anyone leaving to respect the change. He and his wife, after eight or nine Navy moves, bought land near Boston for privacy, not for a house. Fiber in the walls, wilderness out the door. Nature, he says, is vastly superior to almost all of our technology. It is being surrounded by perfection.

He accepted too much risk when he was young and felt immortal. War did not change that; he was up on a horse. Emergencies could have killed him. He did not sit with mortality afterward. He had signed a check, and the machine does not give you time to philosophize, which is the same logic that put the objects off Virginia Beach in a bucket. Staring at what might kill you, all day, would make you worse at the job. Meaning, as he takes it, is curiosity: understand as much of the environment as possible. The human mind can output new information in a universe that is otherwise stale. We can imagine what could never occur naturally, and with enough belief we can make it happen. His job is to encourage that interaction. Lex closes with Buzz Aldrin: bravery comes along as a gradual accumulation of discipline.