Does it matter if the market is manipulated by propaganda? You, the laborer, can either make a native app or a web app, and you know one option is more likely to net a return than the other. Why would you risk making a native app in this market? What farmer plants seeds in the winter?
> and if you've ever used one and an equivalent native app
Of course, I've built an Electron app and a native app as well. And, of course, the native app ran faster. But the electron app took 3 weeks to build and the native app took nearly a year. Both apps are now defunct, but if I had to pick one to keep supporting it would most definitely be the Electron one because that had the preferable simplicity in codebase and toolchain.
I will add that Notion is so incredibly slow to load that it convinced me to pursue native options only, but my non-technical girlfriend and her friends use it every day for planning and have no complaints about it. Would a native version be faster? Duh, but would the Notion team have been able to get their work out to the world this quick if they had pursed native over web? Highly unlikely.
Now you'd say, well Notion's lead is just temporary until a native app matches the feature set and loads instantly. And indeed, Notion would lose out until the next wave of features come and suddenly the native app is struggling to support 3 different OSs while the Notion app continues to innovate at the cost of speed. Developer velocity is a critical metric that we, as developers, seem to discount heavily for some reason, probably because we all suck at estimating the time things take us.
I went through this same cat and mouse game with Sublime Text and VSCode. Sublime was great, but VSC kept getting better and better, with performance improvements along the way, to the point where it didn't make sense to use Sublime Text anymore. I should add that, as a web developer, making VSCode plugins (which is just JS + HTML + CSS) seems like a far more approachable task than doing anything substantial with Sublime's plugin framework. That isn't just important to me, that also means other plugin developers can jump in head first to improve the ecosystem.
> efficiency
Efficiency in what? RAM usage? CPU cycles? If I'm using an application for work I'm going to be using it on a machine with the specs to match the workload. Nobody can seriously claim to be doing professional work on a computer with 2GB of RAM in 2021. So who is this inefficient for? Maybe it sucks for personal computing, if you can't afford to buy a better computer, but that's not the market these applications are targeting; obviously you want to target professionals if you want them to buy into your monthly subscriptions.
One of the biggest Electron apps, Discord, is actually in a very clever position due to its target market of gamers, as those are the people who are most likely to have top-end PCs with high end graphics cards. Rendering Chrome and its three billion helper processes is a probably a breeze on those monsters, but even on my middle ground MBP, it doesn't have any noticeable impact on my workflow. Whatever inefficiency it has is more than justified by the value I gain from using it. And that's really the trade: it may be inefficient, but it has features, which is much better than an efficient native program that doesn't have the features my friends are using.