Or go the other way and make web pages not programs at all, like they originally were. Unfortunately, just communicating information isn't cool, doesn't monetize, and certainly doesn't get clicks.
It should be the users choice how their browser renders, it's their damn browser. Instead, browsers are built to follow standards that mostly favor the developers preferences above those of the person who actually own the device rendering it.
The modern web is user hostile by design.
A lot of websites back in the 90s were just made up of images that were clickable using image maps. It was not a fun time.
I get what the parent comment is saying, though, and it's why I use Gemini (the protocol[0], not the AI). It doesn't align with corporate interests, but for now I consider that a good thing.
Think "an interactive visualisation for choosing a seat when buying a concert ticket." That's not a "traditional" webpage, should work on whatever device the user is currently on (so native is unsuitable), and shouldn't require the user to install anything.
People clearly want web pages to be exactly this; we're on iteration #3 of the idea now. Iterations 1 and 2 were Java Applets and Flash, not necessarily in that order, and they were massively popular despite their shortcomings. Clearly, if we somehow deprecated the web as it is now (which is an utterly ridiculous idea, which makes me really curious what comments like yours are exactly trying to achieve here), as long as we keep it an open standard, users will choose browsers which offer this capability in some form. We should appreciate that the current iteration is an open standard which anybody is free to implement, instead of some proprietary blob that you need to license from Oracle or Adobe.
So do hackers. A big chunk of the open standards revolve around sandboxing malicious code, and given the amount of tracking happening anyway, fail at it.