formal verification & compilers @plankevm | https://xitter.cf/t.co/tPm7ElS6Lv | ex @SpearbitDAO
Joined January 2019
- Tweets9.3K
- Following531
- Followers6.8K
- Likes13.4K
PSA: For those just hopping on the formal verification train, current frontier models (e.g. Opus 5.5) will introduce nonsensical spec bugs if you don't carefully supervise, so at least for now learning to read & write good specs still requires your input!
A prediction as a result from this post: efforts like Fil-C and "rewrite it all in Rust" will lose steam.
Applying FM to legacy software to eliminate and preclude new bugs is much more attractive at this point than rewriting (Rust) or dragging it down with a GC (Fil-C)
New on our Frontier Red Team blog: using Claude to formally verify software at large scale.
Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time).
With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks.
Read our research in full here: anthropic.com/research/forma…
Anyone still believing in 2026 that "FV doesn't scale to real world code" needs to start updating their priors and pick up a theorem prover asap
New on our Frontier Red Team blog: using Claude to formally verify software at large scale.
Formal verification can prove that code is free of certain bugs, but historically it has only been feasible for small codebases (or required a lot of expert time).
With AI models, however, this is beginning to change. In collaboration with @theoremlabs, we used Claude to prove security properties in over one hundred real-world open-source libraries over the span of a few weeks.
Read our research in full here: anthropic.com/research/forma…
Very exciting and daunting taking my first step into the formal methods (FM) community.
Basically starting from scratch career wise but also so many cool unknown unknowns to learn about.
philogy retweeted
Lean v4.35.0 is out, and it ships with verified kernels.
Run `lake check --paranoid` to recheck your project with every bundled kernel.
Blog post: leodemoura.github.io/blog/20…
Trying claude desktop rn with the old tab paradigm and it's crazy how inferior it already feels after just ~2 weeks of using `optchat`
Been using my own implementation (in pi) of @VictorTaelin's OptChat for 2 days now and so far it's been pretty nice!
Having one agent that "knows" everything, can dispatch and manage many work streams async without tab/window switching is neat.
xitter.cf/VictorTaelin/status/21…
I can't overstate how excited I am about OptChat. It has been the best upgrade to my workflow since GPT 3. Second only to learning VIM.
I exported all my past chats to it, and life has been amazing. The convenience of having a single unified log of my entire life, in a way the AI (and I) can navigate to find any information about me is indescriptible.
Before, I had all these campaigns spread out across different Claude Code sessions, and even different harnesses. Each part of my life was on a different place, and Opus sucks at navigating these past memories because it is a huge mess. Now it is all just there, in one place. Any fact from my life is 3-4 zoom hops away. Any prompt I wrote, any algorithm I designed, is a few hops away. It feels like my life is now burned in the models weights, even though it is just a silly compression trick.
"what is the state of that PR about λ-contraction again?"
(3 zoom hops later...)
"it was 2 weeks ago. you went to eat yogurt and forgot to reply"
---
"remind me of that purely affine sort algorithm?"
(5 zoom hops later...)
"you wrote this 7 months ago: <code>. the trick is..."
---
"find the top 3 pending topics in my life that need my attention"
(several zoom hops later...)
"here they are: ..."
---
It just feels fantastic how well this just works.
Also, the way the context is reseted on every turn clearly makes the AI perform better, it completely eliminates context rot, and I never have to /compact manually again. There is no need to.
And yes, it is surprisingly cache efficient, because in-turn messages (where most of the cost is) are cached, and even cross-turn, most of the view is intact due to the way the binary compression works. I had to fix some silly bugs on OptMem to make it so, but now it just works. My costs DROPPED.
Also the fact I can manually navigate the memory tree myself, starting from the root, and finding anything about my life. It is just so natural. Everything is there.
Also, I realize I can take 90% of my AGENTS.md away because now the way the agent should behave just "sticks" from the OptChat log itself. If I change my mind about something, I just send a message and the latest ruling sticks. If the agent forgets something, I just point it out and it finds the place where I explained it. No need to re-explain. No need to manage AGENTS.md manually. It just... works.
On top of that, it now lives on my Mini, so closing my Macbook doesn't stop it, and it has no access to my private data, so I'm finally able to use Computer Use to do things like buy food, book flights, sign on gov[.br] (THANK GOD!!!!!!!!!!). I know this is not news to anyone using Hermes and the like, but it is the first time I'm able to do it and it is already transforming my life for better.
Sorry for the over-excited tweet about something so mundane. It is just insane how incredibly transformative a simple idea can be. Context rot was the root of all evil and now it is gone. I feel free.
I'd like to publish this but it is kinda coupled to my workflows and the interface is incredibly ugly and hacky. Actually, I think I'll just ask my agent to write a prompt explaining all the details of the setup so you guys can replicate! I'll post it below.
It is essentially infinite context that just works.
I really suggest that you guys try it!!!
It may change your life too
Struggling to get good cache hit rates from claude tho as the Claude SDK that I'm using to bridge doesn't seem to give nice cache break point controls (according to the clanker at least) and I haven't had time to look more deeply into it myself but otherwise no huge issues.
Sometimes it does feel like the agents still "forget" things but if you remind them/bring it up it does seem to eventually learn as the summary solidifies the importance of any given fact.
I also experimented with using Jev to classify hallucinated summaries and retry to improve accuracy which helped as well.
Rust is finally a system's language, LFG
The allocator_api feature has been stabilized!
> Rust’s Allocator API standardizes custom memory allocation, enabling collections to use specialized strategies for better control, performance, and memory efficiency.
ALT Image: The allocator_api feature has been stabilized! > Rust’s Allocator API standardizes custom memory allocation, enabling collections to use specialized strategies for better control, performance, and memory efficiency.
