Well, as a text-only browser user for decades, all I can say is that there is a great deal I can do with the text-only only browser than I can no longer do with the enormous "modern" graphical browsers. Even something simple like opening a very large HTML file will cause Chromium to choke.

With text-only browser I can fly around the www and consume information at a speed that is just impossible with a "modern" browser. No auto-loading resources, no pre-fetching, minimal headers, no user-agent, no cookies, no Javascript or CSS, and so on.

In the "modern" browser one ends up with an ever-growing number of tabs. In text-only browser I have no tabs. Yet I have a complete navigable browsing history of content URLs.^1 I have enhanced this with a TLS proxy that records every URL. In a "modern" browser any such comprehensive browsing history generally becomes a liability, along with cookies and too many other things. And the history would be cluttered with auto-loading URLs for no-content URLs that must be filtered out Too complex. Too much work.

I'm in textmode most of the time, the "modern" browser generally cannot operate there, so it is of limited utility. Even just the sizes of the binaries for "modern" browsers are absurd. Chrome is like 150MB. To try to compile a "modern" browser is, IME, more difficult than compiling an entire UNIX-like operating system. Many folks simply refuse to even try or just give up.

One can still download W3C's line mode browser from the 1990's, it will still compile and it still works. I use a curses-based browser but I still do many HTTP requests using nc or similar along with localhost-bound TLS forward proxy. HTTP/1.1 pipelining for bulk text retrieval is a favourite feature of the www and that cannot be done inside a browser.

1. "Content URLs" here means URLs pointing only to pages at the domainname I entered, pointing to pages with text content that I can read, not stock images, pixels, ad servers, telemetry and other garbage.