I think the most frustrating thing here is it doesn't need to be this bad.
Rendering HTML and running a JS VM might seem like a long way around to get a desktop interface, but you're reading this in just that context, alongside a couple of dozen tabs, many running webapps with complexities to rival Microsoft Word '97. You're not using 4GB per tab; you probably aren't even getting up to 4GB with all your tabs.
Think of electron as a single tab open in a browser instance. If you open 12 tabs in chrome, it doesn't use 12x the RAM, because most of the initial RAM is used maintaining the browser instance, not the rendered tab. most electron apps don't have "tabs" but if they did they wouldn't multiply the memory usage either
Three Web apps and 12 tabs might not be the equivalent of that one app in Electron, if it's a heavy one. If a Chrome browser instance running Slack uses less RAM than the Slack Electron app does, __then__ we can say Electron uses more RAM than it should.
I'm not saying it does or doesn't, and I'm not defending it, I just think we need to be a little more fair with our comparisons
I did a comparison a while back (though I'd hardly call it scientific). I built a basic "hello world" Electron app, which consumed 69MB of RAM on Windows 10. When I opened the html file of that app in Chrome it took 82MB.
Rendering HTML and running a JS VM might seem like a long way around to get a desktop interface, but you're reading this in just that context, alongside a couple of dozen tabs, many running webapps with complexities to rival Microsoft Word '97. You're not using 4GB per tab; you probably aren't even getting up to 4GB with all your tabs.
Electron can do better.