In this podcast, Lockheed Martin's CTO talks about their cool planes and space projects. He thinks the space industry is at its most exciting time, but military systems operate in a complex world with no easy answers. Three key things: F-35 fighter jet (stealthy, connected, like a flying supercomputer); Orion spacecraft (the most complex human-rated vehicle ever built, for Moon and Mars); and Auto GCAS (an automatic collision avoidance system that has saved 8 pilots).
At a Glance Lockheed Martin CTO Keoki Jackson discussed the company's engineering achievements in aerospace and defense on the Lex Fridman podcast, covering iconic projects such as the SR-71 reconnaissance aircraft, the F-104 Starfighter, and the origins of Skunk Works (e.g., the P-80 jet aircraft).
Keoki Jackson (Chief Technology Officer of Lockheed Martin) discussed the company's engineering achievements in the aerospace and defense sectors on the Lex Fridman podcast, covering iconic projects such as the SR-71 reconnaissance aircraft, the F-104 Starfighter, and the origins of Skunk Works (e.g., the P-80 jet aircraft). The core argument is that Lockheed Martin demonstrates the pinnacle of human engineering through hypersonic vehicles, missile defense systems, and deep-space probes (e.g., lunar and Mars missions), but military systems operate in a complex world where good and evil are intertwined, with no simple solutions. Key takeaways include that AI research can serve the side of justice, and that humanity's passion for space exploration is a critical driving force for the future.
Jackson argues that Skunk Works’ success stems from Kelly Johnson’s 14 principles—small teams, high autonomy, and close client relationships—which enabled Lockheed Martin to accomplish seemingly impossible tasks.
In terms of historical lineage, Skunk Works began with the P-80 jet fighter (the first U.S. jet fighter), followed by the F-104 Starfighter (the first Mach 2 fighter, dubbed "a missile with a man in it") and the SR-71 reconnaissance aircraft (the pinnacle of speed and altitude, designed to outpace any air defense system). Jackson quotes Apollo astronaut Charlie Duke, who called the F-104 his favorite aircraft.
Regarding stealth technology, Jackson recounts Ben Rich’s anecdote of using a ball bearing to demonstrate the stealth concept to a general: "Would you like to have an aircraft that has the radar cross section of this ball bearing?" The core of stealth technology lies in geometric design (how to reflect radar waves) and material selection (absorbing or reflecting radar waves)—a continuous "cat-and-mouse game."
Extrapolation: Jackson emphasizes that the Skunk Works model remains relevant today, as seen in the development of the F-35 fighter—which is not just a stealth fighter but a "flying information system," boasting unprecedented data fusion and sensor fusion capabilities.
Jackson argues that the ultimate goal of space exploration is to establish a permanent human presence beyond cislunar space, but robotic exploration is expanding our understanding of the solar system in ways that humans cannot replicate.
In human exploration, Lockheed Martin is building the Orion spacecraft — "the most sophisticated human-rated spacecraft ever built" — designed for deep-space missions, from the Moon to Mars. Jackson proposes the "Mars base camp" concept, aiming to enable astronauts to safely return after spending several weeks on the Martian surface. He hopes to see this goal achieved within his tenure.
In robotic exploration, Jackson cites the OSIRIS-REx mission as an example — the spacecraft is conducting a year-long autonomous mapping of the asteroid Bennu, then will use a "pogo stick"-style sampling device to briefly touch the asteroid's surface for just a few seconds, collect samples, and return to Earth. He emphasizes that robotic exploration capabilities have undergone a qualitative transformation, from initial reconnaissance in the 1970s to today's ability to send probes to Pluto, intercept comets, and bring back stardust.
Key Mechanism: Jackson points out that the irreplaceable value of humans lies in the "ability to adapt on the fly to new information," as demonstrated by the Apollo 13 mission, where ground control and astronauts improvised a carbon dioxide scrubber. AI is not yet capable of this.
Projection: Jackson believes that returning to the Moon within the next decade (2024-2025) is a near-term goal, but the real challenge is not "getting there" but "staying" — establishing permanent habitats, enabling vehicle reuse, and ensuring a continuous supply of fuel and consumables. This is both an engineering challenge and a budgetary challenge.
Jackson argues that AI and autonomy will be embedded across all Lockheed Martin systems, but the verification and validation of non-deterministic learning systems represent the greatest current technical challenge.
He defines four levels of autonomy:
1. No Autonomy: Humans perform all tasks entirely
2. Partial Autonomy: Human in the loop, requiring human approval for every decision
3. Supervised Autonomy: Human on the loop, able to intervene or change direction
4. Full Autonomy: Human off the loop
Jackson cites the F-35's carrier landing as an example — the aircraft's control system has "dialed out the variability" (i.e., eliminated the variables introduced by environmental conditions) to the point where it consistently catches the second arresting cable, causing excessive wear on that cable. He also introduces Auto GCAS (Automatic Ground Collision Avoidance System), which has saved seven F-16s and eight pilots, and is currently being integrated into the F-35.
In the space domain, Lockheed Martin is developing an AI assistant named Maya ("the Alexa in space"), which will possess a digital twin library of the entire spacecraft, including design data, test data, and historical flight data, enabling anomaly monitoring and assisting astronauts.
Uncertainty: Jackson acknowledges that the verification and validation of non-deterministic learning systems is an "extremely difficult challenge." Lockheed Martin is funding university research at institutions such as MIT and collaborating with startups. He emphasizes that trust, explainability, and understandability are core requirements for safety-critical systems.
Extrapolation: Jackson believes that future fighter jets will achieve "human-machine teaming," as demonstrated by the Have Raider program — where the F-16 serves as a loyal wingman, and the human pilot, acting as a mission manager, can control multiple aircraft. However, he does not foresee the complete removal of humans from the cockpit in the near term.
Jackson argues that nuclear deterrence prevented large-scale war during the Cold War, but the world today has become far more complex—ballistic missile capabilities have spread from a handful of nations during the Cold War to over 70 countries, and cyberattacks have emerged as a new strategic threat.
Lockheed Martin participates in multiple pillars of U.S. strategic deterrence, including the Trident submarine-launched ballistic missile. Jackson emphasizes that the core of deterrence is to "prevent the use of any of these weapons," with the underlying logic being the ability to deliver unacceptable retaliation even after an attack.
Key Change: Jackson notes that strategic deterrence today has expanded from nuclear, chemical, and biological weapons to the cyber domain—attacks on critical infrastructure (power grids, water supply systems) could have strategic effects comparable to nuclear weapons. He hopes that AI and autonomous technologies can help address these new threats.
On the Military-Industrial Complex: In response to Eisenhower's warning, Jackson points out that Lockheed Martin also engages in commercial businesses (energy, space, commercial aircraft), and that U.S. defense spending as a share of GDP is far lower than during the peak of the Cold War. He believes that changes in global threats "would warrant relevant investments in defense."
Jackson welcomes competition from new entrants like SpaceX, calling it the "most exciting time to be in the space business in my entire life," but emphasizes the need for a level playing field.
Lockheed Martin is also innovating on its own—partnering with Stanford University to build the first U.S.-launched CubeSat (QuakeSat), and is developing "smart sats" and software-defined satellites that can reconfigure mission objectives in orbit.
Extrapolation: Looking ahead to the next century, Jackson believes the world will "move faster, be more connected, more autonomous, and more complex." Lockheed Martin is investing in technologies such as quantum computing, AI, and autonomy to address military threats not yet imagined, and to achieve full-domain awareness from the depths of the ocean to the edge of the solar system.
| Position | Analyst View | Key Data |
|---|---|---|
| F-35 Fighter Jet | Bullish | Hundreds delivered, annual production near 100 units; capabilities include supersonic cruise, stealth, and sensor fusion |
| Orion Spacecraft | Bullish | "Most complex human-rated spacecraft," designed for lunar and Mars missions |
| THAAD System | Bullish | Deployed with the U.S. military, with international customers purchasing |
| Aegis Combat System | Bullish | Protects carrier strike groups from ballistic missiles, aircraft, and cruise missile threats; land-based version Aegis Ashore developed |
| Trident Submarine-Launched Ballistic Missile | Bullish | Component of U.S. strategic deterrence, currently undergoing modernization |
| OSIRIS-REx Mission | Bullish | Currently mapping asteroid Bennu, will collect and return samples to Earth |
| Maya AI Assistant | Bullish | Space AI assistant under development, equipped with a library of spacecraft digital twins |
| Auto GCAS System | Bullish | Has saved 7 F-16s and 8 pilots |
| Matrix System (Sikorsky) | Bullish | FAA-certified "optimal piloting" system, capable of landing on oil platforms in zero-visibility conditions |
| SpaceX | Neutral (Competitive Relationship) | Viewed as a competitive force, but Jackson welcomes the innovation brought by competition |
1. Jackson believes the core of Skunk Works is "small teams + high autonomy" — Kelly Johnson's 14 principles enable engineers to shield themselves from bureaucratic influence and build close relationships with clients, thereby accomplishing seemingly impossible tasks such as stealth technology.
2. "The big challenge is not so much to go, but to stay" — Jackson points out that returning to the Moon is technically less difficult than establishing a permanent habitat, which requires solving the continuous supply of fuel, air, consumables, and spare parts.
3. Jackson proposes a four-tier framework for autonomy — ranging from no autonomy to full autonomy, with different tasks requiring different levels; the F-35's carrier landing is already close to supervised autonomy, while Auto GCAS is a "last-resort system."
4. "A nanosecond to a second is like a second to 32 years" — Jackson uses this analogy to illustrate the value of AI on the battlefield: compressing the OODA loop (Observe-Orient-Decide-Act) to make decisions before the opponent can react.
5. Jackson argues that the irreplaceable value of humans is "real-time adaptation to new information" — the Apollo 13 case proves that AI is currently unable to replicate the improvisational creativity humans demonstrate in emergencies.
6. Jackson notes that strategic deterrence has expanded into the cyber domain — attacks on power grids and water supply systems could have strategic effects equivalent to nuclear weapons, and AI and autonomous technologies may help address these new threats.
7. Jackson believes that verification of non-deterministic learning systems is an "extremely difficult challenge" — Lockheed Martin is funding research at institutions such as MIT, but trust, explainability, and understandability remain core requirements for safety-critical systems.
8. "This is the most exciting time to be in the space business in my entire life" — Jackson welcomes the competition and innovation brought by new entrants like SpaceX, but emphasizes the need for a level playing field.