Few other "browsers" beat links (or links2 nogui), and they are not even "browsers": e.g., the original netcat, tcpclient and BSD's ftp with http support. No, curl is not on the list. It's dog slow.
No one needs to cater to my attention. Apparently I'm not today's end user. I'm for all intents and purposes a blind user. The web is not for me. I don't even start X11 if it's not necessary. I work with text. Graphics and multimedia are for recreation.
One of my favorite recent HN comments/stories was from Diego Basch. He described what happened at Inktomi, an early search engine that eventually was made all but obsolescent by Google.
In his story, he stated what he saw as one of the sure signs that Inktomi was being overtaken by Google. He said he saw that Inktomi engineers did not use Inktomi's search. They used Google.
Are complex browsers the way of the future? I find it easier to work on _operating systems_ than I do to work on today's "modern browsers". That is how complex (and therby insecure) the code has gotten. I would rather try to understand the code for ffmpeg or mplayer than I would for Chrome or Mozilla. But as I said, I'm not the "end user" to focus on.
Developers/engineers gotta eat. Do what works today. Focus on what you think "end users" are doing. Try to anticpate what they "want".
When I'm pondering "the next big thing" and what may work tomorrow to pay the bills, I will always remember Diego's story of the Inktomi engineers.