Ian Miller

I interviewed at Anduril in the summer of 2018. I had just finished my freshman year at MIT. It was my first time in Southern California. I stepped off the plane in Santa Ana and was convinced I was on another planet. The world looked tinted gold with palm trees and beautiful beaches and sleepy beach towns dotting the coast. The 20 or so Anduril employees worked out of a garage that bordered the same John Wayne airport I had flown into that day. Plane activity was a constant source of amusement to some, slight annoyance to most others. I arrived on a Sunday and showed up at the office next to John Wayne airport the following morning. I interviewed with 4 people. 2 of those people were the CEO Brian Schimpf, and the founder Palmer Luckey. Admittedly, I wasn’t aware of Palmer Luckey beore I had first read about Anduril in Wired magazine. I had vaguely heard of Oculus and commercial VR products but no one in my suburban town in Central Florida had any. That Wired magazine article was Anduril’s exit from stealth mode. Ironically, the article that many of those first 20 had viewed as a hit piece (the title in particular being a thinly veiled, politically charged criticism) made me immediately send an email to recruiting@anduril.com. I had reached out at the exact perfect moment to get brought in to do an interview, right before a wave of talented people started lining up to hop aboard the rocket ship that Palmer captained. In July of 2018, I wrote:

“Dear Mr. Luckey,

I first learned about Anduril by way of an article in Wired magazine. There are so many applications for Computer Vision and so few of them it seems, manifested in everyday life. Your company is applying a relatively novel technology to an area in dire need of updating. The application of Computer Vision and, more broadly speaking, AI technology, to the defense industry is something I have thought about for a large part of my young adult life. I am excited and inspired that a man like yourself has that vision and passion to go out and make something happen. I believe I can contribute to what you are doing. I am an Army ROTC cadet at MIT. I have a passion not only for cutting edge technology, but also for my country. I crave the adventure that will come with being a soldier. I look forward to learning about specific areas where AI can make a difference in the performance of the American military and law enforcement. I have research experience working on Visual Recognition enemy-civilian identification, something not far off from the technology detailed in that Wired article. I have work experience in Computer Vision with SmartShot Inc, and am currently busying myself with an Audio Recognition project in the FinTech industry in New York City. I am still in school for a large part of the year, but would be incredibly excited for the opportunity to work for Anduril Industries whether it be in an internship-type role in the short term or a career in the long term.

Sincerely,

Ian Miller “

A combination of desperation, slight writing talent, and association with MIT landed me an invitation to “come check out what we’re building” from Palmer. At that point, Anduril was actively building 2 things: the “Sentry Tower”, which they had deployed a prototype version of down in Texas on a wealthy farmer’s ranch, and an autonomous firefighting tank. The tank is a rare false start in Anduril’s otherwise stellar product development history early on. Jamie Hyneman (of Mythbusters’ acclaim) contracted to build the vehicle. Hyneman brought publicity, and one could argue that was a significant part of the value-add, from the founders perspective. In any case, there were several reasons why the tank was shelved: 1- the tank did not represent the product vision that Anduril would eventually settle on, 2- local governments did not have the coffers to justify such a large R&D investment, and 3- Jamie was difficult to work with. Of course, history is written by the victors and it may also be true that a fledgling startup headed by a 24 year old billionaire was difficult to work with.

From the very beginning, Anduril felt pre-ordained to me. In reality, Luckey and company were attempting something that very few people thought was possible. Anduril aimed its sights on a market that hadn’t been breached more than twice in the past 50 years. The common refrain in the early days was that there had “only been 2 successful defense startups in the past 50 years (SpaceX and Palantir) and both of them had to sue the government to reach that point.” I distinctly remember Matt Steckman, CRO, shrugging his shoulders at lunch one day as he conceded that Anduril might be forced into a similar course of action. The military is a unique customer and very few people know how that customer buys things. Oftentimes, the military wants to build its own things instead of buying an equivalent system on the open market. SpaceX and Palantir found themselves in this exact situation in the mid 2010s. Both companies successfully proved that the government had violated the Federal Acquisition Streamlining Act of 1994 by not procuring technology from them and instead attempting to build copies in-house. Steckman came from Palantir and was intimately familiar with all of these details. As a 20 year old sophomore in undergrad, I was quite oblivious to this fact (and many others). Companies do not necessarily succeed because they build useful things. In fact, many companies succeed without building useful things. No companies succeed without the ability to sell. And in the early days, this was the decisive talent that Anduril had recruited in the form of Palmer Luckey and Matt Steckman. As a 20 year old, I was primarily mesmerized by the prospect of writing software for a startup in California. It was something I’d always dreamed of doing, and to be doing it for Palmer Luckey was pretty much the ideal case.

That wasn’t the first thing I ever heard Matt Steckman say though. “Is it possible spend 1% of what the US is spending right now on defense, and maintain our current military capability?” This was Steck’s grand vision, his daily motivation: the US military should be outfitted with many cheap, attritable, autonomous systems instead of a handful of complex, expensive systems. This statement describes the counterposition that Anduril had to take against “the defense primes.” Defense primes build fighter jets and aircraft carriers. It’s effectively impossible to do that with private money. For such complex systems, you need the customer to shoulder some of the engineering risk. In the early days, both Palmer and Steck loved to beat up on cost-plus contracting. Certainly, much interesting discussion can be had about the warts of a cost-plus contracting model. In the early days, cost-plus was what Anduril promised not to do. Instead, Anduril would build things with their own money and then sell those things to the government. Steck thought this would save the taxpayer billions. I interview lots of SWEs nowadays and probably 75% of them mention something about private R&D as the reason they want to come work at Anduril. Probably half of them explicitly call out “cost plus” as a waste inflater. Steck’s plan for many attritable, autonomous systems would not only save the taxpayer money — it was also likely be necessary to save the homeland in the case of all-out war. In the case of all-out war with a near-peer adversary, aircraft carriers for example would be easy targets, wiped out in minutes by a battery of cruise missiles that are at this moment pointed toward those same carriers.

Before Anduril could “change the way that the military acquired technology”, they needed a product and revenue. The first product ever, the short-range Sentry tower, would effectively form an “electronic border wall”. Hundreds of these towers would line the southern border between the US and Mexico and detect illegal foot traffic. The tower would make the job of Customs and Border Patrol agents much easier. Of course, working with CBP earned Anduril some early enemies and immediately thrust the company into the political arena. Until 2019, the majority of the company worked on the Sentry tower. At this same time, I had started trying to recruit my MIT friends to join as interns. Another startup challenge I did not appreciate was recruiting. Even in 2019, it was not easy to convince computer nerds from San Francisco to work at a defense company. It wasn’t easy to convince computer nerds from relatively apolitical MIT to work at a defense company. I remember having several conversations with classmates who simply couldn’t stomach the thought that what they built would someday be used to kill someone. Some of the early engineers at Anduril felt the same way. I was an intern in 2019 when the head of the Perception team posted a video on youtube explaining his rationale for working at Anduril in spite of this anti-weapon attitude. In the video, he starts with the claim that what he cares most about is the environment - preserving it for future generations, saving the ice caps, reversing climate change. He then performs some incredible mental gymnastics to explain how working at Anduril is actually the best way to add value toward the environmentalism effort. This is one example. Especially in the early days, Anduril was composed of a huge variety of people. Lots of those people would never ever consider working in defense except that this seemed like a uniquely high growth opportunity. This diversity of thought was absolutely crucial to keep focused on product.