Why Build Native Instead of Web Applications?

When I was a boy, logging into AOL and downloading game demos was the best. While my mom and sister were playing Slingo through AOL's interface, I was busy installing and playing the Command & Conquer demos. Browsing around, finding games to download was nearly as enjoyable as the games themselves.

Installing the games was fantastic, too. Starting up the Command & Conquer install wizard, watching the progress bar slowly increase over the next 20-40 minutes--different images from the game cycling while I waited...the anticipation was incredible.

I had a similar feeling when installing other software. Remember ICQ? Remember when installing Photoshop was exciting?

At some point, we lost that. Not only did installing applications stop being a happy event, it started becoming a humiliation ritual. Why download and install Discord when it's just a web application I can use in my browser? Why do I have to pay a subscription for absolutely every single piece of useful software (hint: Adobe started it)? Why can't I just purchase and own something?

The rise of web applications is partially a story about programmers making technical decisions, and partially a story about pointy-haired bosses vibe-managing their subordinates to build something useful as fast as possible with no thought given to respecting the user. Programmers know that there were roughly two options:

1. Make an application that is trivially cross-platform from the beginning--a web application. Sure, it'll take more compute, memory, and battery power, and it may not be a buttery-smooth experience for the user, but they'll be able to experience it.
2. Make an application that is cross-platform with considerable effort--a native GUI application. That application would respect the user's time and resources more, but comes at a great cost in time and effort on the part of the programmer.

If the programmer was the only one making the decision, he'd probably pick option 2 most of the time, but programmers usually don't make the decisions--that's the manager's job. The manager sees things differently:

1. Use lower-skilled, homogenized, easily replaceable web developers to build the thing fast.
2. Use highly-skilled, opinionated, difficult to replace native developers to build the thing slow.

Speed and stability of development is the key--get to market fast, and don't blow your budget getting there. Getting the product finished and in customer's hands quickly and cheaply is important because nobody really knows which software is going to be successful, but we know most of it won't be, so it's important to iterate and try lots of things out and see what the customers like.

Men that share both skill stacks and understand both the technical and business challenges that need to be overcome are rare, and they usually have a boss that isn't technical. They must make a cost-benefit analysis taking into account both the technical and business sides of software development, and usually they will weigh the business side more heavily out of both a desire to make cool things that people like and to climb the career ladder.

I don't want to cheapen the costs of native development either. Building software is risky because we don't know the future. We don't know what people really like or want much of the time, especially if we are making something completely new. Native development explodes the size of that risk enormously. If risk is expensive, it's scary. And if it's scary, it probably won't happen.

I actually find the SpaceX Raptor engine development as the model for good software development: initially overly complex, ugly, and inefficient--with the intention of iterating upon and improving the final product into something beautiful--something you can be proud of both for its utility and aesthetic appeal both to normies and your peers.

That's why I always supported businesses and individuals starting with an Electron application. Making a web-native application is ugly and nothing you can be proud of as a programmer. But it allows you to try things out cheaply and quickly, get feedback from users, and iterate on the product after that. Software development that follows the Raptor model is the ideal and should be encouraged.

But the problem is that the decision makers never have the intention of following that model of development. They are happy to vibe-manage their way to seniority and a golden parachute when the consequences of their vibing catch up to them. Usually they aren't in a good position to do otherwise--a series of catastrophic decisions by vibe-politicians (WWI & WWII were optional; Keynesian monetary policy has been pure cultural poison; Communists were not systematically identified and executed; post WWII wealth in America produced a population of weak, soft, and selfish old men who did not care to maintain the prosperous civilization they inherited), vibe-investors (clueless, degenerate, Mammon-worshipping VCs pumping companies with strings-attached money), and vibe-educators (those demonic Communists making young people political revolutionaries devoid of even basic, marketable skills) have so distorted and corrupted the civilization that built the Golden Gate Bridge and Hoover Dam (in MONTHS), sent men to the Moon on kilobytes of memory, and produced energy from glowing hot metal that the vibe-managers can hardly be blamed for not giving a damn about the memory efficiency of their vibe-managed chat client or the poor performance of their vibe-managed video game.

The calculus changes with LLMs, however. People who give a damn are slowly being granted superhuman powers to actually fight back against the people who don't. The risk of native development--a proxy term for any development conducted by people who care about making the future better for everybody via thoughtful software development--is greatly decreased with LLMs. The Good Guys have the ability to circumvent the destructive influence of the wicked Vibers and make great software. You can make a native application quickly and cheaply with a much smaller group of people (maybe even just you, the handsome reader!). You don't have to answer to anyone or anything else except your users (I pray you get many), your inner voice, and to God.

That's why I'm giving native application development a try. I also believe this will be the trend for the next decade and generation of programmers. That explains a strange phenomenon skeptical software engineers have noticed: Tremendous enthusiasm and hype behind LLMs, but no tangible increase in the quality of our current software. For the most part, software developers have witness only a visceral decrease in quality of software because of LLMs. Businesses clumsily adding "AI" to their product and distracting the user; New software on app markets that is obviously vibe coded with no respect to the user--just a quick payday; Vibe-copied software that is visibly inferior to the original.

Furthermore, normies have noticed that they can't pick up their phone or read their email without some Nigerian or Indian scammer trying to steal their life-savings or Bitcoins from them. We, the Situation Monitors, know that these scams are supercharged by LLMs. Scammers use LLMs to correct their English to make their communications appear genuine. They use LLMs to produce voices of family members during phone calls or produce realistic human images during completely LLM conducted video calls. Avoiding these scams is hard for our elder parents and those calls contribute to an ever decreasing trust in society.

So for various reasons, both the Initiated and the Commoner have mostly been exposed to the negative influences of LLMs.

But LLMs are a force multiplier, either for Good or Evil. And We Who Give A Damn are busily working outside the current corporate/investment structure to find our place as part of the forces for Good. That would explain why the initial, visible products of LLM use appear negative, yet there are still so many people optimistic for a future using LLMs. The Sons of Mammon have taken first blood in the battle for the future of software and civilization, but the war has just begun.