Hacker News new | ask | show | jobs
This is a valid domain name (remove www) (ai)
37 points by theseobosscom 1478 days ago
16 comments

So they registered the domain "www" just like they could have registered "google" or "lol".

You can access Google by typing:

- google.com (domain name)

- www.google.com (subdomain)

The same for the domain in question:

- www.ai (domain name)

- www.www.ai (subdomain)

~~The reason why typing ".ai" or "ai", etc. In your browser works is in IMO due to the internals of the browser that automatically adds a www in front of the domain if it can't resolve it because historically so many websites run on http://www.example.com and neglect to setup http://example.com~~

Edit: as pointed by another comment (https://news.ycombinator.com/item?id=31690379) there are actually TLDs with A records which is the case here.

So to take my example from above: The registry responsible for the AI TLD set up a website on the domain name "www.ai" but neglected to set it up on the subdomain "www.www.ai" Instead, they set it up on the TLD "ai" which is indeed very unusual.

This is not correct. The TLD alone is a valid domain name, although in most cases it must be followed by a trailing dot. "ai." is valid, "ai" may not be.
It is a valid domain in the technical DNS sense. But practically speaking, each TLD has "ns" entries in the root zone database, not "A", "CNAME", etc. entries. This is just a guy who registered www under a TLD that's it.
This has nothing to do with the "www" subdomain. I assure you that the unix "host" tool is not trying to guess subdomains if the lookup fails:

  $ host "ai."
  ai has address 209.59.119.34
  ai mail is handled by 10 mail.offshore.ai.
I'm not saying "ai" is not a valid domain. I'm just guessing why people get to a website when they just type "ai".

I'm on my phone now, if you search for the ai entry in the root zone file (https://www.iana.org/domains/root/files) you should find that IP returned by your host command.

Can you try curling "ai" and see if the content returned (if any) is the same as the one from "www.ai"?

They are the same, but that's not really the point. www has been registered in addition to the entries for the root TLD. You said "the internals of the browser that automatically adds a www in front" - that's irrelevant. host isn't doing that, curl isn't doing that.
There's a bunch of TLDs with A records, even Google used to point .dev to 127.0.0.53 for a while.
You are right I didn't know it was allowed
I thought 'ai' was valid according to spec, but depending on your browser it may or may not actually go to 'http://ai./'
If you are now just learning that .ai domains are available, I'll save you some time, samur.ai is taken.

Unfortunately it appears a link farmer has it rather than a cool hacker.

I was about to purchase it, lol
To all who can't replicate it: Clicking a link, copy/pasting the URL, and typing it from the keyboard all behave differently. Make sure that you manually type

    http://ai/
into the location bar (do not use https). Also try with and without the trailing slash.
Safari on macOS will not load this address. It does resolve, weirdly, "host" in the Terminal works and I can see Safari's requests being answered by my home DNS server in the logs.

But Safari just throws an error: Can't connect to the server "ai".

Update: It works if you include a trailing dot: http://ai.

Update 2: HN's URL parser does not include the dot in the hyperlink. Frustrating.

http://ai. Works in safari for iOS, shockingly.
Huh. It does not work on my iPhone. Same error. I wonder what the difference is.
The link didn't include it, but try again with the trailing '.' period
Aha, yes, this works.
Weird. `dig` on macOS does not resolve it but `host` does.
That's interesting. dig works for me.

I suspect all these difference depend on what DNS server you are using.

The proper URI should be "http://ai." (with the ending dot)
n@ai is also a valid email address. Owned by a guy named Ian.
A long time back I entertained the idea of buying hc.im and using the email address le@hc.im. (my name is michael) Perhaps making business cards with a laser cutout so I could turn the card over if it wasn't apparent. But two-letter im domains are a very expensive ongoing cost so it didn't seem worth it for the gimmick.
I have a very short email address (in the form a@aa.aa) and I haven't found anywhere that doesn't accept it - except for humans, who often refuse to believe it is real and will insist on a longer one.

(I have a backup longer one.)

I have a similar one and it has been rejected many times. I still write to the company but nobody gives a shit. I also tried on twitter.

Surprisingly people are fine with it, I think that this is thanks to social media handlers such as @wazoo - people group all these together as "email and such"

I suppose it should be possible to get an MX record to work with a root domain, but that person that has it would have to probably work for or be the owner.
MX records fallback to A records, so ~10ish TLDs work this way already.
sun@. would likely blow up even the best email checking regexes.
creative!
I've been maintaining a list here for a few years: https://captnemo.in/tld-a-record/

(Automatically updated, but Travis broke this a while ago). There's some older diffs on my blog[1] if you search for "A Record")

[1]: https://captnemo.in/archive.html

The page itself has a (hidden) link to its mentions in popular news and elsewhere

www.ai/publicity

"Server Not found" for any of:

- https://ai/

- https://www.ai/

- http://ai/

- http://ai.

- http://ai

and http://.ai/ gets redirected to a search engine searchfor ".ai" by firefox.

But http://www.ai/ works for some reason.

http://ai

works for me in firefox. perhaps you are using a wrong browser?

The http://ai/ works with Edge on Android.
ai. in Safari on iPhone works. But I have to type exactly that.
I was mistaken.

Thanks for the -30 downvotes HN.

It is interesting because http://.ai works (it says remove www, not the dot).

Edit: The HN generated link does not work because it appends a trailing slash to the end.

Works without the dot for me. (Not the https variant though.)
Yes, it worked for me too (Firefox). Interestingly http://ai goes to http://www.ai.com while http://ai/ works. ¯\_(ツ)_/¯
If I remove "www" as per title, ie. if I type ".ai" or "https://.ai" it doesn't work. What am I missing?
Interestingly, most WHOIS services fail to resolve this domain, and show alternative such as ai.com. Its like going off the marauders map :)

The page informs us it is registered in Anguilla.

Even if you remove www it works, i.e. http://ai/
Doesn't work for me (macOS, Firefox)
Works fine for me (macOS, Firefox)
Clicking a link, copy/pasting it, and manually typing it all behave differently. Try typing it.
Works for me on Firefox Android
Works for me on MacOS Chrome
Doesn't work for me on firefox or curl. dig can find it but only with +trace, host reports SERVFAIL.

"Valid domain name" seems a stretch.

Doesn't work for me (Windows, Vivaldi)
try: http://ai. (note the dot on the end)
This one works for me. But I don't get it. Can someone explain? This isn't browser trickery, nslookup succeeds for "ai." but not "ai".

> ai

** UnKnown can't find ai: Non-existent domain

Server: UnKnown

Address: 192.168.86.1

> www.ai

Non-authoritative answer:

Server: UnKnown

Address: 192.168.86.1

Name: www.ai

Address: 209.59.119.34

> ai.

Non-authoritative answer:

Server: UnKnown

Address: 192.168.86.1

Name: ai

Address: 209.59.119.34

>

The period forces it to actually do a lookup. Without it it may be only asking the local dns server or appending the default domain.
http://ai./ as well
Can someone explain to me why this is a valid URL?
>A domain is a non-empty ASCII string that identifies a realm within a network. [RFC1034][1]

>The example.com and example.com. domains are not equivalent and typically treated as distinct.[1]

    When a user needs to type a domain name, the length of each label is
    omitted and the labels are separated by dots (".").  Since a complete
    domain name ends with the root label, this leads to a printed form which
    ends in a dot.  We use this property to distinguish between:
    
       - a character string which represents a complete domain name
         (often called "absolute").  For example, "poneria.ISI.EDU."
    
       - a character string that represents the starting labels of a
         domain name which is incomplete, and should be completed by
         local software using knowledge of the local domain (often
         called "relative").  For example, "poneria" used in the
         ISI.EDU domain.
    
    Relative names are either taken relative to a well known origin, or to a
    list of domains used as a search list.  Relative names appear mostly at
    the user interface, where their interpretation varies from
    implementation to implementation, and in master files, where they are
    relative to a single origin domain name.  The most common interpretation
    uses the root "." as either the single origin or as one of the members
    of the search list, so a multi-label relative name is often one where
    the trailing dot has been omitted to save typing.[2]
[1] https://url.spec.whatwg.org/#concept-domain

[2] https://www.rfc-editor.org/rfc/rfc1034

That works for me in Safari on macOS.
Doesn’t work for me on iOS Safari.
Works for me—make sure you're getting http, not https, though.
Only http://ai (with or without trailing dot) did not work for me using DuckDuckGo browser on Android phone.
Works for Edge on Android.
Okay..WTF? http://ai. Didn’t work on Chrome Mac but it works in Safari for iOS. Is this a bug or how the hell is this possible? It’s flipping everything I thought I knew about domains upside down.
'ai' is a resolvable DNS name, just like you can resolve 'localhost'

The owner of 'ai' decided to add an A record on their zone.

And just like that, all the www.* domain names have been bought.
So did someone buy an empty ".ai" domain? Is that how?
Most likely it's the registry of the .ai who manually created a DNS A record for "ai."
Huh, another dodgy tax haven I did not know about!
How?
I notice it works for pretty much every TLD, including .com too.

http://com. resolves to http://www.com

http://cd. shows a list of all such sites.

It seems likely that the browser has some kind of logic in the resolver to make this work specially for “www” domains.