SSL tries to solve several problems at once (identity verification + encryption of connection) over a protocol (TCP/IP) that isn’t cryptographically tied together. Some very dedicated, smart people have sketched out some ideas on this. Basically, this compromise only happened because we’re relying on a chain of third parties and intermediaries to do this verification of certs.

Web of trust, blockchain-based system (where a name like ebay.com can be registered and publish their own cert, so you can prove whoever owns ebay.com also owns that cert), a TCP/IP replacement that you can derive a key from (there are several interesting proposals here, some of which run on top of IP), are all ways of solving this problem with different tradeoffs.

This biggest problem like with any protocol change is inertia behind SSL. Same reason a number of excellent anti-spam protocol enhancements never made it to be widely accepted in SMTP and now everybody is fleeing to other methods of communicating without SMTPs issues.

The crazy thing is that everybody I know who works in tech including myself is throwing hundreds of dollars a month at AI tools and they’re still posting massive losses.

The great hardening continues. Good to see.

Allow them. Anti-AI hate is kneejerk sentiment. AI is the future whether your sentiments agree with it or not, and it’s a major enablement tool for OSS developers.

Spend your money wisely. The market can’t sustain what there is no demand for.

SSL has had a broken trust model for decades. We need something truly decentralized and auditable.