Earlier quoted context omitted.
What if the website told the browser to load malware? Browsers in the modern web need to defend the user, not execute arbitrary instructions from random websites that nobody cares about.
A website cannot tell a browser to "load malware", unless we're talking about a exploit, which should be patched. (Please don't say "if I send you a malformed png file you have to execute the exploit, otherwise your argument breaks down".)
When I buy and read a newspaper, I don't expect the publisher to start following me everywhere and keeping a log of my life. When I read an article online, I shouldn't have to think about that either. But sites have so flagrantly abused the ability to deliver more than just the content I've deliberately requested, in order to track (and monetize) user behavior everywhere, that it's entirely appropriate for my User Agent to take steps to defend me.
I don't mind a site delivering some ads alongside the content I've asked for, just like I accept some ads in a printed magazine. But I don't expect my magazine to come with an embedded tracking device that will stick to me like a burr, even long after I've read the content and recycled the pages.