That's not a great solution, since many people are barely aware what at url is, but I think it should still be one that you and I (as people who are) use.
Similarly the article is fixed by guessing that google chrome is probably at chrome.google.com (also chrome.com), firefox is probably at firefox.com, cnn is probably at cnn.com, gmail is probably at mail.google.com, gmail.google.com, or gmail.com (actually all 3), hacker news is probably at hackernews.com (oops), etc.
Which also risks ending up on the wrong site (e.g. Steam is not at Steam.com). I'd trust Google to know the correct URL more than my guess.
Basically I'm trading Google/Bing and ads known to be malicious, for ICANN/registrars, self interested companies, and a reasonably functional "legal" system. Edit: And less tracking, and faster access to websites.
Guessing domain names is even worse than Googling.
.com is pronounced 网 wǎng, "net", which doesn't leave much conceptual space for other TLDs.