Yesterday, I attended a defense startup event. A bunch of people building defense startups were talking about what they are working on and what is happening in the industry.

I always thought defense was an attractive space. It sounds important, high stakes, manly, and there is a lot of money in the U.S. defense market.

But! While in normal startups, you can build some shit, show a nice demo to investors, exaggerate what it can do, and maybe fix it later. In defense, if your shit does not work, people die. That is the customer.

A person who may literally die because your drone failed, your software made the wrong decision, or your system stopped working at the worst possible moment.

There seems to be this weird situation where some defense startups are trying to sell the U.S. government a $200,000 drone that barely flies. Nobody really checks it under real conditions. The demo looks good, the presentation looks good, investors are impressed, and everyone acts like it is some kind of magic.

Then you bring it into a real war and it does not fucking work. Now people are dead, and you are like, well, ok.

The price difference is also insane. In a real war, people may need a drone that costs $600, can be built quickly, and is disposable. They do not need a beautiful $200,000 prototype that works once during a controlled demo and then breaks in the field.

It made me think: what the fuck are you actually building? Are you building something for the people who will use it? Or are you building shiny shit for investors, procurement officers, and conference stages?

And if you are building a defense startup but are not willing to go near the war zone, watch people use your product, understand how it fails, and feel responsible for what happens, are you really serving your customer?

Defense should not be pursued for glory. It should not be pursued because it feels cool, important, masculine, or well funded. It should be pursued because you genuinely want to help the people whose lives depend on what you build.

And then I started thinking that maybe every startup should be viewed in a similar way. Imagine that every customer would die if your product failed. Would you still build some shitty AI agent that does not work? Would you still release unreliable software, hide the problems behind a nice interface, and tell everyone that it is revolutionary? Maybe not.

Obviously, most software does not literally kill people. But thinking this way changes your standard. It makes you ask whether you are actually helping the customer, whether your product really works, and whether you are willing to take responsibility for the outcome.

Maybe we would build much less shit if we thought about every customer that way.

Thanks, David Venegas.