Dan Wallach

Big news! I’ve started a new position today at RAND’s Center for AI, Security, and Technology (CAST). I'll be working on projects that allow me to use my technical expertise to inform policy and shape consequential decisions, such as mitigating the impacts of AI on computer security. It’s going to be an exciting new adventure.

Taking on this new role at RAND marks my departure from DARPA, where I’ve managed a portfolio of programs in cryptography and computer security. One highlight was launching TRACTOR (Translating All C to Rust), which is making remarkable progress at eliminating memory safety vulnerabilities, once and forever more.

I’m also retiring from Rice University, where I’ve been a faculty member since 1998. I’m deeply grateful to the colleagues and collaborators I’ve worked with, and to all of the students I’ve taught and advised over the years. In my new role at RAND, I look forward to staying connected with everyone as we work together on some of the most consequential technological and policy challenges of our generation.

https://www.rand.org/global-and-emerging-risks/centers/ai-security-and-technology.html

Peter G. Neumann has passed away. 93 years old.

Update: John Markoff, at the New York Times, has published an obituary. https://www.nytimes.com/2026/05/17/obituaries/peter-g-neumann-dead.html

Eric Rescorla has taken a 40-page swing at "what policy makers need to know about AI safety and security". Really interesting. https://educatedguesswork.org/assets/ai-security-policy.pdf

I've recently been playing around with vibe coding some basic tree-like data structures (treaps, red-black trees, AVL trees, and hash-array mapped tries) in Rust, and then twisting the arm of the LLM to do an optimization from Sarnak and Tarjan (1986) that lets you keep a version history without paying O(log n) path copying costs. This is the sort of thing that, in the old days, might have made for a useful undergraduate senior thesis that they'd crank on for a semester.

I'm at a point where I now have modest confidence in the correctness of my vibe code (e.g., it's got property-based tests that check the invariants, and also doesn't crash under load, despite lots of internal calls to Option::expect()), but I'm not confident enough to share it. It's not bad but not great.

At some point, I'll write up something useful about what I've learned about how to vibe code (in short, write vicious unit tests or you're doomed), but meanwhile I thought I'd skip straight to the data.

For comparison, I also included Rust's "im_rc" crate, which includes a human-written HAMT by
@bodil@social.treehouse.systems; I'm using the faster "Rc" version since that's how I vibe-coded all those others.

Punchline 1: @bodil@social.treehouse.systems wins. Their work is the "IM HashMap" in the graph. Higher is better. X-axis is problem size and Y-axis is throughput. The benchmark was 80% reads and 20% a mix of inserts, deletions, and updates. Every "version" is saved, creating (hopefully) real memory pressure.

Punchline 2:
The Sarnak-Tarjan optimization definitely helps for the various binary trees, but my attempt to do it for HAMT ended up costing a factor of two in perf. Yikes.

Graph below generated by Criterion.rs. Yes the colors are horrible. Measured on my M1 MacBook Air, because why not.

A useful explainer about why static type systems, in general, and Rust, in particular, are super useful for debugging code.
https://blog.daniel-beskin.com/2025-12-22-the-compiler-is-your-best-friend-stop-lying-to-it

This is great: https://blog.trailofbits.com/2025/11/25/constant-time-support-lands-in-llvm-protecting-cryptographic-code-at-the-compiler-level/

LLVM 22 (and presumably all the subsequent versions) now have a constant time select intrinsic to enable cryptography algorithms to tell the compiler exactly what they need (i.e., evaluate both sides of a conditional expression then select the one you want), replacing gross bit hacking expressions that newer optimizers would unravel.