If I were to recommend one cryptography book for implementors in 2026, would it be:

(Edit, would love your comments as to why.)

We @binarly_io just open-sourced our VulHunt framework at @REverseConf

GitHub: https://github.com/vulhunt-re/vulhunt
Documentation: https://vulhunt.re/docs
Slack: https://join.slack.com/t/vulhunt/shared_invite/zt-3pyjsrqef-xMVFB5C4J6718nPNJAk0Fw

If a Klein bottle could wear pants, would it be like this or like this?

#mathstodon #math #maths #shitpost

One of the biggest statistical biases one encounters when trying to assess the true success rate of AI tools is the strong reporting bias against disclosing negative results. If an individual or AI company research group applies their AI tool to an open problem, but makes no substantial progress, there is little incentive for the user of that tool to report the negative statement; furthermore, even if such results are reported, they are less likely to go "viral" on social media than positive results. As a consequence, the results one actually hears about on such media is inevitably highly skewed towards the positive results.

With that in mind, I commend this recent initiative of Paata Ivanisvili and Mehmet Mars Seven to systematically document the outcomes (both positive and negative) of applying frontier LLMs to open problems, such as the Erdos problems: https://mehmetmars7.github.io/Erdosproblems-llm-hunter/index.html

As one can see, the true success rate of these tools for, say, the Erdos problems is actually only on the level of a percentage point or two; but with over 600 outstanding open problems, this still leads to an impressively large (and non-trivial) set of actual AI contributions to these problems, though overwhelmingly concentrated near the easy end of the difficulty spectrum, and not yet a harbinger that the median Erdos problem is anywhere within reach of these tools.

I reverse engineered DexProtector, the security solution protecting applications like Revolut and other banking apps.

From custom ELF loaders to vtable hooking, here is an insight into how these protections work and their limitations.

https://www.romainthomas.fr/post/26-01-dexprotector/

RE: https://mathstodon.xyz/@robinhouston/115821134807127928

Wow, someone discovered a more efficient way of multiplying two 3×3 matrices! You might think this would have already been solved.

This new method, due to A. Perminov, uses 58 addition/subtractions and 23 multiplications. The previous best used 60 addition/subtractions and 23 multiplications - and that was discovered only in August last year!

This is not at all the sort of math I'd ever want to work on; it reminds me of pole-vaulting and other specialized athletic competitions. But it's one of those problems that's easy to explain, hard to solve. And those are great for luring people into the subject.

(As @robinhouston emphasizes, this new scheme is the only the best known if you want to allow for the possibility that multiplication could be noncommutative for the 'numbers' in your matrices - like if they're other matrices, which can be useful in recursive algorithms for multiplying large matrices . For commutative rings, like the reals and complex numbers, people already know how to do multiply 3×3 matrices with just 21 multiplications.)

Ran into a problem in prod?
Just generate a fake cloudflare error page and blame it on them - gives you time to fix.

https://github.com/donlon/cloudflare-error-page

#foss #devops #cloudflare #infosec