The pitch he heard
Give us the money. We will do good things.
— DARPA program managers to George Heilmeier, 1975. He stopped funding it.
The pitch he funded
Build an aircraft radar cannot see. Two demonstrators, thirty million dollars, first flight in twenty months. If the model on the pole reflects more than the pole, stop.
— The stealth program, answered through his eight questions. Contract signed April 1976.
George H. Heilmeier had already invented the liquid crystal display at RCA when he took over DARPA in 1975. He inherited an agency under pressure. Vietnam had ended, defense research money was shrinking, and the 1969 Mansfield Amendment required the Pentagon to show a military purpose for every research dollar. Congress wanted to know what the agency was buying.
Heilmeier asked his own program managers first. The people requesting the largest sums could not tell him what they were going to do, what would change if it worked, or how anyone would know. The pitch was the one at the top of this page: give us the money, we will do good things. Most founders have made it. So have we.
So he wrote eight questions, and no program got money without answering all eight. The stealth program answered them. So did every DARPA program after it. Fifty years on, the questions are still published on darpa.mil, and IARPA, ARPA-E, and ARPA-H screen proposals with them.
In October 1973 the Israeli Air Force lost 109 aircraft in 18 days, most of them to Soviet-built missiles and guns the United States expected to face in Europe. The aircraft were American. Two years later DARPA had a director who required eight answers before any program got money, and a Lockheed team claiming it could build an aircraft radar could not track. Here are its eight answers.
Build an aircraft that radar cannot see well enough to shoot at. One sentence. No acronyms. A senator could repeat it.
Fly low and fast, jam the radars, and send Wild Weasel crews in first to kill the missile sites. The limit: jamming is an arms race the Soviets keep winning with new radars, and the Weasels trade aircraft for sites.
Shape the airframe so radar energy reflects away from the receiver, using a Soviet diffraction paper the Air Force had translated and a Lockheed program that predicts radar return before metal is cut. Why it will work: the return falls by orders of magnitude, and the range data already shows it.
Every pilot ordered into a missile belt, and every planner whose air campaign begins by trading aircraft for SAM sites. The difference: the Soviet air-defense investment stops working on the first night.
A shape chosen for radar may not fly; the design is unstable and needs a fly-by-wire computer borrowed from the F-16. The model may not predict full scale. The secret may leak. Both demonstrators later crashed. The program still passed, because the risk register had said the exam was the radar range.
Two demonstrators. About $20 million of government money and $10 million from Lockheed's own board. Everyone in the room knew the split before the contract was signed.
Contract in April 1976. First flight 1 December 1977. Twenty months, and the date held.
Midterm: a full-scale model on a pole at the White Sands radar range, March 1976. The model returned less energy than the pole holding it. Final: a flying demonstrator against Air Force radars. Passing it produced the F-117 contract in November 1978.
Eighteen years separate the pole test from the first night over Baghdad. The answers to questions one and four, written in 1975, described that night before anyone had built the aircraft. That is what the catechism is for.
Sources: DARPA, Stealth · GAO NSIAD-97-134, Operation Desert Storm · Air & Space Forces, How the Skunk Works Fielded Stealth
What are you trying to do? Articulate your objectives using absolutely no jargon.
Write the sentence a squadron commander would repeat to their boss without looking at your slide. If it needs "AI-enabled", "multi-domain", or "next-generation" to survive, the objective is missing. Heilmeier's test was a program manager explaining to a congressional staffer why the money should move. Yours is a mission owner explaining to a program executive officer why the requirement should. Same sentence.
Write your answer ↓How is it done today, and what are the limits of current practice?
Name the incumbent workflow, the people who run it, and the exact point where it stops working. "Manual" and "legacy" are jargon in disguise. The 1975 answer was specific: jammers, low-level ingress, and Wild Weasel crews trading aircraft for missile sites. If you cannot describe today's method with respect, you have not seen it up close, and the operator reading your deck will know.
Write your answer ↓What is new in your approach and why do you think it will be successful?
Two questions in one, and founders usually answer only the first. State the mechanism that changes the outcome, then state the evidence that the mechanism works: a pole test, a field trial, an exercise result, a deployment. Lockheed did not walk in and say "stealth". They said radar return falls by orders of magnitude when you shape the airframe with these equations, and here is the range data.
Write your answer ↓Who cares? If you are successful, what difference will it make?
Name the person, the command, and the budget line that changes if you succeed. Enthusiasm from operators is a signal. Money moves when a program office with a validated requirement owns the outcome. For stealth the answer was every pilot ordered into a missile belt, and that answer held up eighteen years later over Baghdad. If your "who cares" is a market size, you have answered a different question.
Write your answer ↓What are the risks?
List the ways this fails, in order of likelihood, and say what you will do about each. Technical risk is the easy part. Transition risk, color-of-money risk, authority-to-operate drag, and a champion rotating out are what kill funded prototypes. Heilmeier's original wording paired risks with payoffs. Keep that pairing: a risk register with no matching payoff is a reason to stop.
Write your answer ↓How much will it cost?
A number, with what it buys. Have Blue was two aircraft for about thirty million dollars, and everyone in the room knew the split between DARPA, the Air Force, and Lockheed. A range is acceptable. "It depends" is a failed answer. Cost also has a color: state which appropriation pays, because a Phase II SBIR and a procurement line cannot buy the same thing.
Write your answer ↓How long will it take?
A date, and the mission window it has to land inside. Contract to first flight for Have Blue was twenty months. Defense ventures live inside a twelve to eighteen month window before a champion rotates, a budget cycle turns, or a competitor fields. If your schedule spans two PPBE cycles, say so, and say what you deliver inside the first.
Write your answer ↓What are the mid-term and final "exams" to check for success?
Define the test that would make you stop, and put a date on it. The midterm for stealth was a full-scale model on a pole at White Sands in March 1976. The final was a flying demonstrator against Air Force radars. Both were pass-or-cancel gates, agreed before the money moved. Investors call these milestones. Heilmeier called them exams because you can fail them.
Write your answer ↓Fill in what you can in the form on this page. Blank answers are useful: they show which questions you still have to answer. When you send it, one email goes to Merge Combinator and a copy goes to you. We will read it and reply if we think we can help. No mailing list, no follow-up sequence.
September 2026 · About the studio