Why Is Trump Going to Steam Catapults?
"What aircraft-carrier catapults taught me about confirmation bias—and how to use AI to challenge what you already believe."
I thought it might be interesting to demonstrate how to use AI to test our views. I saw the following quote from Senator Mark Kelly:
“I’ve launched off the front of aircraft carriers hundreds of times, have a Master’s degree in aeronautical engineering, and am a test pilot and even I wouldn’t suggest to the Navy how to engineer specific systems on its ships. Donald Trump should stick to what he knows best — ballrooms, bankruptcy, and bullshit.”
The quote is in reference to using steam catapults rather than electromagnetic on aircraft carriers. “Even I” suggests that Kelly believes he is better equipped than President Trump to make decisions about aircraft carriers but he wouldn’t do it. Of course the politics is that Presidents ultimately must make decisions. No president of anything should make decisions just based on personal skill set or experience. Decisions should be made based on data, when available.
So, I decided to go to Grok. First, I presented the question, “Why is Trump going to steam catapults?”
Grok returned the following:
“President Trump ordered the U.S. Navy to switch future Gerald R. Ford-class aircraft carriers back to traditional steam-powered catapults (and hydraulic elevators) instead of the Electromagnetic Aircraft Launch System (EMALS), primarily because he has long viewed the newer system as overly complex, expensive, and less reliable.”
Grok also gave more supporting statements by Trump, going back to “at least 2017” but all seemed to be his opinion.
“In short, the move stems from Trump’s years-long conviction—rooted in anecdotes, cost concerns, and a preference for proven older technology—that steam is better and more practical for carriers.”
The casual reader would have stopped there. I admit that I prefer the idea of electromagnetic to steam. That certainly seems to support the mantra - Trump is an idiot. I could have stopped there, but I wanted to know whether there was data behind his opinion.
Then I asked, “What are the reliability statistics?”
Grok returned:
“EMALS (Electromagnetic Aircraft Launch System) on Ford-class carriers has consistently fallen short of its formal reliability targets according to Pentagon testing data, while traditional steam catapults on Nimitz-class and earlier carriers are mature systems with very high availability.”
Grok gave more details and I reviewed with Chatgpt:
The Navy’s formal requirement is 4,166 mean cycles between operational mission failures (MCBOMF), where a “cycle” equals one aircraft launch. Reported figures from the Pentagon’s Director of Operational Test and Evaluation (DOT&E) in Fiscal Year 2022 reported 614 MCBOMF.
Using Chatgpt and Grok, I learned that the Navy was emphasizing 98% operational availability since 2023. DOT&E was discussing MCBOMF while the Navy was emphasizing operational availability.
The Navy has also cited operational availability around 98% for recent sets of launches in some public statements, focusing on uptime rather than strict MCBOMF.
Grok went on to state that the EMALS systems share power which can make it difficult to isolate and repair systems. It can take an hour and a half for an EMALS system to power down. Depending on the failure, all four catapults can be down during the repair time.
As for steam, Grok replied:
Steam catapults (used for decades on Nimitz-class and earlier carriers) are a mature technology refined over roughly 70 years. Carriers equipped with four steam catapults have historically achieved ~99.5% availability of at least one catapult. Failures are relatively rare in operational service, and repairs can typically be handled shipboard with more conventional tools and procedures.
I asked, “How many catapults are on a carrier?” The answer is four.
I asked, “If one fails, do all four fail?” Grok replied:
“In short: one steam catapult failure leaves the others usable; one EMALS failure is more likely to affect overall launch capability until repairs are complete.”
I asked, “Can both systems be repaired during combat?”
“Both systems can be repaired during operations or combat to some degree, but steam catapults are substantially easier and more practical to fix under combat conditions.”
Grok ended with its assessment:
“Steam → more repairable by the crew during combat, with better ability to keep launching while fixing.
EMALS → more constrained; isolation limits and specialized support needs make sustained combat repairs harder.”
I also learned that EMALS may provide quicker launches, provide more precisely controlled acceleration, can accommodate a broader range of aircraft weights, reduces stress on aircraft, and was intended to reduce maintenance and manpower. However, its reliability is documented to be less than the original requirement and certain failures can affect more than one catapult.
DOT&E data showed 614 MCBOMF in 2022. It also states there have been 8,725 launches in a 262-day deployment and claims it doesn’t have enough data to determine MCBOMF. How can 8,725 launches not give enough data to determine MCBOMF?
The Navy has emphasized approximately 98% operational availability. Availability and MCBOMF aren’t the same thing. A system can be available 98% of the time and still fail far more frequently than its original reliability requirement.
When anybody starts changing how data is collected or reported, I suspect shenanigans.
I still don’t know whether Trump is making the right decision. Changing a carrier already under construction may cost more than any benefit from returning to steam. That is a judgment call.
I learned something from this exercise. AI can generate something to support our cognitive bias. While doing that, it can inform us and help us ask the next question, and the next, and so on. IF, we are willing to question.
AI can support our confirmation bias or it can be a big help in getting to the truth. If we are willing to question it.