bit

7-node ESP32-S3 cluster running a 0.4B LLM via 1.58-bit (BitNet) ternary quantization over SPI daisy-chain - Low-Zi-Hong/ESP32s3-LLM-Cluster

Let's chill just a little bit on the "TLA+ will save AI from itself" narrative

It’s finally time for me to wrap up this ZX Spectrum series, with a look at how to get the simple 1-bit beeper on every Spectrum model to emit high-quality digital audio. The results are quit…

<p>NEW FEATURES AND NOTEWORTHY NEW PACKAGES</p> <ul> <li>Support for 64-bit little-endian POWER machines (ppc64le)</li> <li>Support for 64-bit IBM z Systems (s390x)</li> <li>Rust 1.17.0 and Cargo 0.18.0</li> <li>GHC 8.0.2</li> <li>Julia 0.5.2</li> </ul> <p>SIGNIFICANT UPDATES</p> <ul> <li>GCC 6.3</li> <li>LLVM 4.0</li> <li>PHP 7.1</li> <li>Go 1.8</li> <li>Python 3.6</li> <li>LuaJIT 2.1beta3</li> <li>nginx 1.12</li> <li>Ruby 2.4</li> </ul> <p>NOTEWORTHY CHANGES</p> <ul> <li>The llvm package has been changed to be provided by a versioned llvm&lt;X&gt; package, which is presently llvm4;</li> <li>The -grsec kernel-related packages have been renamed to -hardened;</li> </ul> <p>DEVELOPMENT-RELATED CHANGES</p> <ul> <li>MD5 and SHA-1 hashes have been removed from APKBUILDs, being obsoleted by SHA-512; set -e is now implied in APKBUILDs, automatically failing the build upon unhandled failing commands;</li> <li>A check() function has been added to APKBUILDs that allows packages to run test suites after build(), ensuring no regressions have occurred. This has been implemented for a number of packages, and policy onward will be to have them either be present or explicitly opted-out of with good reasoning;</li> </ul> <p>CREDITS</p> <ul> <li>Thanks to everyone sending in patches, bug reports, new and updated aports</li> <li>Thanks to GIGABYTE, Scaleway, Fastly, IBM, Packet, and vpsFree for providing us with hardware and hosting.</li> </ul>

Let's chill just a little bit on the "TLA+ will save AI from itself" narrative

With Rust, you can do low-level number-crunching and bit-by-bit processing, while enjoying memory safety and concurrency features. With Helix, you can use your Rust code inside of a Rails project.

What if renewing a passport was a JRPG quest? Not metaphorically. I mean six signatures, absurd...

You're probably familiar with modern processors made by Advanced Micro Devices. But AMD's processors go back to 1975, when AMD introduced t...

VUPEN Security provides vulnerability management and intelligence solutions (advisories, analysis, exploits).

It&#039;s a drop in replacement, it&#039;s faster, and it&#039;s cheap. What&#039;s not to like?

<p>Alternate title: Nearly SNFS-Speed Signature Forgery Sans Factoring N (NSNFSSSFSFN)</p> <p><strong>Abstract.</strong> The security of RSA is generally understood to be based on the complexity of factoring, and key size parameters are extrapolated from the general number field sieve (GNFS). However, this may not accurately represent RSA security in practical scenarios.</p> <p>An under-appreciated 2007 algorithm of Joux, Naccache, and Thom´e allows an attacker to forge RSA signatures after temporary access to a raw RSA signing/decryption oracle in time close to the <em>special</em> number field sieve (SNFS) without factoring the key. We implement and run this algorithm for 1024-bit RSA. In total, the attack took 1380 CPU core-years over five calendar months, and made 232 oracle queries. Most of this time is precomputation; after the precomputation the attacker can forge any signature of choice, offline, in 180 core-years.</p> <p>We carried out our attack using a hardware security module (HSM) as the signing oracle, thus demonstrating the ability to impersonate the HSM through black-box API interactions, without exfiltrating the key. Blind RSA schemes also provide such a signing oracle.</p> <p>Extrapolating our empirical running times to larger key sizes, we conclude that the concrete security of RSA with a signing oracle should be 15 to 30 bits lower than the factoring-based security estimates for the 1024-bit to 4096-bit RSA parameters that are common in practice. Even 4096-bit RSA does not appear to meet a 128-bit security level in this attack model. This highlights a gap in current RSA-type security assumptions, and gives classical cryptanalytic evidence in favor of moving away from RSA entirely during the current post-quantum transition.</p>

How the Internet ended up with 4.3 billion addresses. Vint Cerf picked a number in 1977. It had to be enough for an experiment.

Let’s cut through the noise, one bit at a time

Just to step back and get a look at the big picture (or the big pictures), when we talk about AI, which these days almost always comes down ...

Glyd stores open models' bf16 weights in about 11 bits instead of 16 and decodes them inside the GPU's matrix multiply: bit for bit, often faster.

A 128-bit universal hash function designed for fast file/data stream integrity checks with cryptographic strength, written in Rust. - orlp/polyxor

A new hex puzzle every day. Trap the cat before it slips off the edge — can you solve today's?

Nearly SNFS-Speed Signature Forgery Sans Factoring N - ucsd-hacc/NSNFSSSFSFN

Taking 'execute logging' a bit too literally - CVE-2026-88771

https://genesis.mzpx.ai — genesisplanet streams live on Twitch! Check out their videos, sign up to chat, and join their community.

Nearly SNFS-Speed Signature Forgery Sans Factoring N - ucsd-hacc/NSNFSSSFSFN

My article from two weeks ago on fancy 1-bit sound on the ZX Spectrum was a bit incomplete. I used a shortcut to test out the &#8220;interleaved oscillator&#8221; technique, and I couldn&#8217;t ge…

Jim walks through the classic 2038 problem — Unix's 32-bit signed timestamp (seconds since Jan 1, 1970) overflows on January 19, 2038, flipping negative and breaking anything that depends on file/time comparisons…

<p>My fedora co-maintainer wrote this about the release of terminator-1.92.</p>

Repacking Gemma 4's QAT weights five ways and serving each on the same SageMaker NVIDIA L4 endpoint: int4 linears, int4 embeddings and lm_head, FP8 and int8, across E2B, E4B, 12B, 26B A4B and 31B.

A modern IDE and 64-bit runtime for Visual FoxPro 9 applications: the same language, rebuilt on Electron, React and a Rust virtual machine in WebAssembly.