zer0luck.eth

Collection of papers on LLM for vulnerability detection github.com/huhusmang/Awes… #llm #infosec hermestool.duckdns.org

Today we're introducing TRIBE v2 (Trimodal Brain Encoder), a foundation model trained to predict how the human brain responds to almost any sight or sound. Building on our Algonauts 2025 award-winning architecture, TRIBE v2 draws on 500+ hours of fMRI recordings from 700+ people to create a digital twin of neural activity and enable zero-shot predictions for new subjects, languages, and tasks. Try the demo and learn more here: go.meta.me/tribe2

Hello! As a follow up to Monday’s (March 23 2026) Arbitrum Sepolia network outage and out of an abundance of caution, we are asking all node operators to upgrade all of their nodes to Nitro 3.9.8 at their earliest convenience. This release is available as a Docker Image on Docker Hub at offchainlabs/nitro-node:v3.9.8-4624977. This recommendation applies to node operators for all Arbitrum chains. Thank you for your understanding!

Google Quantum AI is expanding to include neutral atom research with our superconducting program. Complementary paths will accelerate our mission to build a large-scale, error-corrected quantum computer—an engineering challenge we are excited to tackle. goo.gle/40TycSV

LiteLLM GitHub thread github.com/BerriAI/litell…

Introducing TurboQuant: Our new compression algorithm that reduces LLM key-value cache memory by at least 6x and delivers up to 8x speedup, all with zero accuracy loss, redefining AI efficiency. Read the blog to learn how it achieves these results: goo.gle/4bsq2qI

Today we're releasing MolmoWeb, an open source agent that can navigate + complete tasks in a browser on your behalf. Built on Molmo 2 in 4B & 8B sizes, it sets a new open-weight SOTA across four major web-agent benchmarks & even surpasses agents built on proprietary models. 🧵

If you want to see what Xint Code reports actually look like, check out the app at code.xint.io We've pre-loaded public scan results for two popular web vulnerability benchmarks (NodeGoat and AltoroJ) and two critical OSS projects (PostgreSQL and QEMU).

We just launched Neo, a security testing platform that proves vulnerabilities, not just reports them. • 66 exploitable vulns found in benchmark testing (most of any tool) • 24 findings no other tool caught • 22 CVEs in open source projects, zero direction given prnewswire.com/news-releases/… At #RSAC? Try it live at Booth #3131 → projectdiscovery.io/events/rsac-20…

VScode + llama.cpp + QEMU (container2wasm), all running locally inside a browser Work by @TokunagaKohei #KubeCon github.com/ktock/vscode-l…

We’re introducing Dynamic Workers, which allow you to execute AI-generated code in secure, lightweight isolates. This approach is 100 times faster than traditional containers. cfl.re/4c2NvPl

Update on burp-ai-agent extension drama #issuecomment-4110917796" target="_blank" rel="nofollow noopener">github.com/PortSwigger/ex…

Autoresearch 这种 修改 -> 评估 -> 保留/舍弃 -> 修改... 无限循环迭代的思路太好用了，终于可以实现让 Codex 无人值守 24 小时不停工作了，就是 token 消耗太恐怖了，试了下只是跑 2 个循环任务，4 个 Pro 账号是不够用的

vibe coded a fuzzing ai agent last month and let it run for a week using my $200 claude max. it then found 21 high/critical vulnerabilities in Chrome.

Today, at the @DARPA expMath kickoff, we launched 𝗢𝗽𝗲𝗻𝗚𝗮𝘂𝘀𝘀, an open source and state of the art autoformalization agent harness for developers and practitioners to accelerate progress at the frontier. It is stronger, faster, and more cost-efficient than off-the-shelf alternatives. On FormalQualBench, running with a 4-hour timeout, it beats @HarmonicMath's Aristotle agent with no time limit. Users of OpenGauss can interact with it as much or as little as they want, can easily manage many subagents working in parallel, and can extend / modify / introspect OpenGauss because it is permissively open-source. OpenGauss was developed in close collaboration with maintainers of leading open-source AI tooling for Lean. Read the report and try it out:

We've reached an agreement to acquire Astral. After we close, OpenAI plans for @astral_sh to join our Codex team, with a continued focus on building great tools and advancing the shared mission of making developers more productive. openai.com/index/openai-t…

Comprehensive Formal Verification of @CetusProtocol on @SuiNetwork: core CLMM protocol, standard library (STL), and integer-mate. Our formal verification confirms correctness of critical protocol properties, including deductive proofs that the tick-to-price conversion functions correctly implement the mathematical formulas (√1.0001^tick and its inverse). This goes well beyond all other formal verification work on Uniswap v3-style CLMM protocols, including Uniswap itself. As part of our security partnership with Cetus, we'll continue hardening their protocol. Details and report below.

This is how you run a zero human AI Agent Company in 2026. OpenClaw, Cursor, and Codex agents organized under one org structure, pointed at one goal. Get started in just one command. 100% Opensource.

Sui’s new VM just went public. Bug bounty is open & we want the community's eyes on it before Mainnet. The bella-ciao branch is a rewrite of the execution layer. Faster. Per-package caching. Next gen Move features. Internal review & multiple independent audits from @osec_io & @zellic_io done. Report on @HackenProof. Paid at Mainnet rates, before it hits Testnet.