It looks like LetsEncrypt is really the only service that issues ACME certs anymore. So basically it's an SPOF for like 90% of the web. Impressive how well they managed to fuck that one up, nice job EFF.
Follow

@cjd ICANN can (and has) canceled TLDs and domains (on threat of canceling TLD) for political reasons. The shadowy "TLS cabal" selects CAs for inclusion in the "trusted" list for mainstream browsers. Any CA on the list can forge any cert whenever they want.

ICANN is ok for public websites - but if you want actual security, run your own TLD and CA. Like we did in the old days before ICANN.

We've been CLI (manual) signing certs, but I'm about to try this open source ACME server for private CAs/TLDs:
github.com/smallstep/certifica

Regarding the domains, there's sort of a solution to this with PKT domains (my project).

People register pkt.<whatever> and point it to our nameservers, then people who register domains on PKT blockchain get <their domain>.pkt.<every TLD>

So going after these is ... complicated.

But if LetsEncrypt just stops issuing certificates for political reasons then we lose SSL...

@cjd Run your own ACME server for .pkt. Like the one I linked.

The BIG issue is that Normies only comprehend how to add fully trusted CAs to their browser (allowed to validate any domain). We need a normie friendly way to add a PKCS#11 policy that e.g. does *not* trust cabal CAs for .PKT and *only* trusts your CA for .PKT.

If you can convince people to install a VPN app, then you can convince them to install a *local* CA and approve it's CA cert, then it can just sign off on websites on the client side...

@cjd But then, the cabal CAs can still forge certs for the local stuff. Because the cabal CAs are trusted for *everything* - not just ICANN TLDs.

If you trust the VPN and it's DNS system, then you don't need to worry about certs.

@customdesigned
Almost everything supports this. For example, I distribute a CA to all my equipment that's only valid for r000t.com and subdomains
@cjd

@r000t @cjd Right, but the trick is to convince the browser that the cabal CAs are NOT valid for r000t.com and subdomains. That requires PKCS#11

As long as MiTM attacks are not a concern, you don't need to worry about this. And solving MiTM *without* using x509 is better than trying to use it.

@cjd @r000t That's why I use raw IPv6 (cjdns or yggdrasil) for p2p email and VOIP.

If you can't trust DNS or cabal CAs, then not using either leads to fewer mistakes.

@cjd Not the biggest fan of golang - but at least it compiles fast, unlike Rust.

I've used both a fair bit and I prefer Rust because while Go does have a lot of nice features:

x := SomeStruct {
// hehe missing field here, lets just make it nil
}

Lead to so many blowups that it soured my opinion of the entire language. Not to mention most of my go experience was working on terrible code...
Sign in to participate in the conversation
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.