@ApacheArrow / Databases / Compilers. (past @SDFLabs, VoDa, @Spotify). Rust/C++/TLA⁺🇧🇷 → 🇸🇪 → 🌎

Joined June 2008
Isolation solves every problem in software engineering, except the problem of too many isolated components.
Python's venv system sucks and the inventor should feel bad, you're telling me every venv has to be updated independently on my machine? I have hundreds.
3
3
34
922
With Macaroons even clients can append a duration caveat without any database interaction. Cryptographic magic.
I've been cooking on this post for a bit, but with the cyber security events of this week I had to ship it: Long-live static credentials must die. We never want to rotate API keys again. Turns out only 8% of cloud and SaaS vendors support secure alternative. This is our call to action to change this AND our commitment to do our part: - We are requiring vendors on our marketplace to support OIDC - We shipped @shadcn components to make it easier to ship OIDC configuration - We are expanding OIDC support on our own platform The last generation got the web from HTTP to HTTPS. It is on us to move past API keys. vercel.com/blog/death-to-lon…
6
373
Felipe O. Carvalho retweeted
Convinced myself that this is hard but is the way to go with agents. You get a request from the user, attenuate the macaroons from the token to the bare minimum necessary for a task, pass that token to an LLM that pass the token around to do anything it decides to do.
Macaroons (2014) hasn’t been widely adopted as JWTs and I can see why. It requires programmers to start thinking about AuthZ in a very non-intuitive way. Each item in a Macaroon shrinks the set of allowed actions. More perms are “added” by not having caveats that forbid actions.
2
5
1
14
2,280
Theoretically possible now equals practically real.
2
1
9
366
You have to exorcise your programs from bundled certificates!
PSA: If you use Node or Bun, please upgrade even if you’ve enabled --use-system-ca! That flag unions the system roots with the the bundled roots, so older versions will continue to accept the eMudhra cert even if your OS removes it.
5
471
If you could store one private key per proton in the universe (you can't), you would only be able to store 2^263 keys (log_2(Eddington Number)~=263). A truly random 384-bit key has never been written down or appeared anywhere in the universe. But sure, pick the number 4.
5
383
Very pretentious from eMudhra’s part to say they *chose* to leave the business. The business is trust and they can’t be trusted. There is no business for them anymore.
I suspect the real reason eMudhra is exiting the SSL business is Google indicated it is most likely going to drop eMudhra from being trusted by Chrome AT ALL, after observing inexplicable incompetence in how they mis-generated their private key for the CA certificate.
11
90
1
1,468
10,073
This CA debacle reminds me of Mozilla simply refusing to take SERPRO (a technical Brazilian agency) seriously in its attempts to add itself as a root CA.
2
11
632
I'm adding a `cargo deny` rule that prevents dependency on the bundled CA certificate store in Rust projects. Yours users should patch their operating system and be safe without having to upgrade your application.
1
1
12
514
You know what’s really messed up? We didn’t really need AI to know that handling web requests with PHP is an endless source of security issues. We should have a mechanism to quickly revoke certs and secrets when they are guessed correctly (e.g. 4) or leaked. And yet, here we are.
3
25
665
PHP can't be stopped
1995: PHP is dead, learn ColdFusion 2002: PHP is dead, learn ASP​.net 2003: PHP is dead, learn Django 2004: PHP is dead, learn Ruby on Rails 2010: PHP is dead, learn Flask 2011: PHP is dead, learn AngularJS 2016: PHP is dead, learn Next.js 2022: PHP is dead, learn Solidity 2026: The NSA is advising that you turn off PHP
6
858
The problem wasn't the 4 in the lowest byte, but the long sequence of padding zeroes.
1
18
730
You wouldn’t see me picking an algorithm that uses a non-power-of-2 key size. 384 bits in a key? Clownish. Get a grip on reality! Use AES-256-GCM or Chacha20-Poly1305.
1
8
554
Digital security is important.
1
5
373
If you’re legit, you should have nothing to hide! 🫣/s
The situation where a certificate's private key only has 8 bits of randomness instead of 384 is unacceptable. We should urge all certificate authorities to immediately publish their private keys so that the community can make sure they are random enough.
9
908
Me generating a private key for a side-project I will never launch vs. a leading provider of digital signature solutions trusted by millions.
1
1
11
471
I’m finding not knowing about how hard a problem is to be a blessing more than a curse these days. Knowing how hard some things are and seeing agents make mistakes all the time is wearing me out.
Replying to @_Felipe
We have some war stories about this. It’s… harder than you might expect.
11
913