- Improving Fable 5's biology safeguards
anthropic-news· 07-ago
Aug 7, 2026Product
- From Continuous Predictors to Clinical Thresholds: Early Evidence on Performance Trade-offs of Guideline-Based Categorisation for Ischaemic Stroke Outcome Prediction
arxiv-ai· 07-ago
arXiv:2608.05203v1 Announce Type: new Abstract: Machine learning models achieve strong predictive accuracy for 90-day outcome prediction in acute ischaemic stroke, yet clinical adoption is limited by the misalignment of model explanations with clinicians' reasoning. Motivated by a clinician user study calling for clinical guideline-aligned cut-offs, we ask whether continuous predictors can be replaced by clinically informed categorical encodings without sacrificing performance. On a multi-centre European registry stratified into three treatment cohorts, we compare standard and fully categorised gradient-boosted models, the latter using stroke guideline-aligned, treatment-specific thresholds. The fully categorised models are statistically indistinguishable from their continuous counterpart…
- Agentic Nesting: A New Methodology for Existing Enterprise Application Integration and Services
arxiv-ai· 07-ago
arXiv:2608.05159v1 Announce Type: new Abstract: Enterprise operations extensively rely on multiple heterogeneous business systems and information applications, which also result in severe data silos and process fragmentation. Enterprises have invested considerable financial and material resources in building these applications, however, effectively leveraging and orchestrating them remains a formidable challenge. Conventional approaches to enterprise application integration, encompassing middleware architectures such as Enterprise Service Bus (ESB), API gateway infrastructures, and Robotic Process Automation (RPA), suffer from inherent limitations like high architectural coupling, escalating operation and maintenance costs, and limited intelligence capabilities. This paper proposes Agenti…
- SkillTrace: Multi-Trace Provenance Auditing for LLM-Agent Skill Reuse
arxiv-ai· 07-ago
arXiv:2608.05204v1 Announce Type: new Abstract: LLM-agent ecosystems are rapidly growing around reusable skills: mixed-modality packages of metadata, natural-language instructions, code, tools, references, and operational workflows. As skills become marketplace artifacts, auditing their reuse is no longer the same problem as ordinary code clone detection. Existing detectors target single-modality source code or whole-package similarity, yet skill reuse evidence is distributed across authored text, implementation fragments, and operational structure. As a result, they can miss reuse that preserves only one part of a skill. We present SKILLTRACE, a multi-trace provenance auditing framework for LLM-agent skill reuse. SKILLTRACE extracts three provenance traces: Expression, Implementation, an…
- Abstract Event Causal Rules: Induction and Application
arxiv-ai· 07-ago
arXiv:2608.05205v1 Announce Type: new Abstract: Event-centric intelligent analytical systems heavily depend on explicit causal event knowledge for risk early warning, decision-making support and narrative comprehension. Nevertheless, existing instance-level causal pairs suffer severe generalization deficits on low-frequency long-tail and unseen event combinations. To address this limitation, this work proposes Abstract Event Causal Rule (AECR), a novel relation-level causal abstraction paradigm that transforms concrete cause-effect pairs into generalized abstract causal logic while retaining their intrinsic causal relationships. We design a multi-agent Concrete-to-Abstract Causal Induction (CACI) system coupled with similarity-constrained clustering to distill trustworthy AECRs from noisy…
- Otter: A Time-Aware, History-Conditioned Human Chess AI
arxiv-ai· 07-ago
arXiv:2608.05206v1 Announce Type: new Abstract: Otter is a 15.3M-parameter human chess AI that predicts human move selection by modeling play as a time-aware, sequential process rather than treating each position in isolation. It combines two conditioning signals: (1) a move history encoder that conditions predictions on the last 20 moves, capturing opening preferences, positional drift, and intra-game behavioral tendencies; and (2) a time control module that modulates predictions based on clock pressure. Otter is trained on 6.1 billion positions from 117 million Lichess rapid games over 30 days on a single T4 GPU. Otter achieves 55.23% top-1 and 90.95% top-5 move-prediction accuracy, surpassing the prior state-of-the-art human chess model, Maia 2, with far fewer parameters and less train…
- SearchAuditor: Auditing and Attributing Failures in Long-Horizon Search Agents
arxiv-ai· 07-ago
arXiv:2608.05212v1 Announce Type: new Abstract: Deep search agents tackle challenging questions through long-horizon web interactions, a process that is both complex and fragile: small reasoning errors may propagate through long, noisy trajectories into fluent but incorrect answers. Diagnosing such failures is difficult, requiring the manual inspection of extremely long execution traces, which could be beyond human capacity. We therefore introduce SearchAuditBench, a benchmark that evaluates whether LLM auditors can localize, attribute, and repair these failures, thereby reducing the human burden. SearchAuditBench comprises 1,243 failed trajectories, averaging 73.1 messages and 65.1K tokens, collected from eight open-weight models on five deep-search benchmarks, each expert-annotated with…
- PD-GS: Phoneme-Driven 3DGS for Audio-Driven Talking Heads
arxiv-ai· 07-ago
arXiv:2608.05218v1 Announce Type: new Abstract: 3D Gaussian Splatting (3DGS) enables fast, photorealistic talking-head rendering, yet accurate lip articulation remains elusive: mouth motion is often over-smoothed and may violate hard articulatory constraints such as bilabial closures, producing the notorious ``leaky mouth'' artifact. A key difficulty is that brief, discrete articulatory events are inferred from a continuous acoustic embedding under a regression objective, which biases predictions toward averaged mouth configurations. While modern self-supervised speech encoders provide rich prosodic and phonetic cues, they do not provide an explicit, frame-aligned linguistic target that reliably disambiguates closure-level events. We propose \textbf{Phoneme-Driven Gaussian Splatting (PD-G…
- When Privileged Guidance Misaligns: State-Matched Routing and Contextualized Self-Distillation for Multi-Turn Agents
arxiv-ai· 07-ago
arXiv:2608.05219v1 Announce Type: new Abstract: Privileged on-policy distillation provides dense supervision for multi-turn agents by allowing a synchronized teacher to re-score the student's response at every turn with access to training-only references, such as successful trajectories. In interactive environments, however, the student's preceding actions continually change the execution state. As the student takes different actions or completes subgoals in a different order, its rollout may reach states not covered by the reference, making the reference an unreliable source of guidance for the state actually reached. Applying privileged distillation indiscriminately therefore creates state--reference mismatch. This mismatch motivates a central objective: providing privileged reference g…
- The Ignition Index: Measuring Global Workspace Dynamics in Language Models
arxiv-ai· 07-ago
arXiv:2608.05160v1 Announce Type: new Abstract: We introduce the Ignition Index (I), a validated scalar metric that operationalizes Global Workspace Theory's (GWT) all-or-none ignition prediction in transformer language models. The metric fits a four-parameter sigmoid to per-layer linear probe accuracy as a function of input signal strength, extracting steepness parameter beta-hat: high values indicate abrupt, ignition-like transitions; low values indicate graded build-up. Across 11 models spanning five architecture families, shuffled-label controls demonstrate 9.6-fold selectivity for genuine linguistic structure over spurious probe capacity (p < 0.001, Mann-Whitney U-test). We find: (1) Feedforward transformers exceed SSMs by 89% in aggregate beta-hat (p < 1e-13, Cohen's d = 0.52), with…
- Woodpecker Distillation: Weak Models Diagnose Reasoning Bugs in Strong Models
arxiv-ai· 07-ago
arXiv:2608.05168v1 Announce Type: new Abstract: Large language models often fail on reasoning tasks despite possessing the capability to solve them. We argue that many such failures arise from localized reasoning bugs in intermediate steps rather than from global incompetence. We show that these bugs are frequently repairable: inserting a short patch generated by a weak probe model after the same strong-model reasoning prefix can redirect the trajectory toward a correct solution. However, this corrective effect is not reliably internalized by directly fine-tuning on weak patches or repaired trajectories, suggesting that the useful signal lies not in the intervention text itself, but in how it reshapes the model's future reasoning distribution. We therefore propose Woodpecker Distillation,…
- How Big Models Teach Small Models to Be Smart
bytebytego· 05-ago
In this article, we will walk through the idea from the ground up.Aug 5 • ByteByteGo31025
- LLM Security Basics: The Full Threat Model
bytebytego· 03-ago
In this article, we try to build a map of the full attack surface that threatens an LLM’s security.Aug 3 • ByteByteGo367713
- The Read Path versus the Write Path: Strategies and Techniques
bytebytego· 06-ago
In this article, we will look at read path and write path operations and techniques in detail.13 hrs ago • ByteByteGo1212
- Why An LLM’s Memory Gets Expensive and How to Fix It
bytebytego· 04-ago
In this article, we will learn how LLMs use memory, how it gets expensive, and how to fix it.Aug 4 • ByteByteGo351514
- Hiring: Part Time Instructor, Write Production Grade Code with AI
bytebytego· 31-jul
We’re hiring a part-time instructor for “𝐖𝐫𝐢𝐭𝐞 𝐏𝐫𝐨𝐝𝐮𝐜𝐭𝐢𝐨𝐧 𝐆𝐫𝐚𝐝𝐞 𝐂𝐨𝐝𝐞 𝐰𝐢𝐭𝐡 𝐀𝐈”.Jul 31 • ByteByteGo29211
- Cloudflare is the only vendor named a Visionary in 2026 SASE and SSE reports
cloudflare· 05-ago
We're honored to announce that Cloudflare is the only vendor that has been recognized as a Visionary in both the 2026 Gartner® Magic Quadrant™ for SASE Platforms and the 2026 Gartner® Magic Quadrant™ for Security Service Edge reports.
- From ranking to recommended: get your site ready to thrive in the age of AI agents
cloudflare· 06-ago
More than half of requests now come from machines, not people. Agent Readiness shows how well agents can discover and read your site, while Answer Engine Optimization tracks how often AI assistants recommend you.
- Cloudflare AI Search: give your agents a search engine for your data
cloudflare· 06-ago
AI Search makes search easier than ever, with no Cloudflare primitives to stitch together. Point it at your data to create a search for your own files and websites. We're also sharing a preview of our new pricing model.
- Cloudflare OS: an open platform for agents, apps, and work
cloudflare· 05-ago
Cloudflare OS is an open-source platform that lets everyone in your company build apps, automate work, and safely access internal systems, shaped around what your organization knows and how it operates
- How we’re rethinking work at Cloudflare with Cloudflare OS
cloudflare· 05-ago
We built Cloudflare OS to equip our teams to safely rethink how they get work done with AI. The platform brings together the best of our technologies, from our Compute primitives to our Zero Trust suite. This post walks through our journey to give our users the best AI tools available.
- Introducing Kitesurf: The agent-first browser that runs in V8 isolates on Cloudflare Workers
cloudflare· 06-ago
We should be giving all agents tools that excel at what’s important for an AI model. Kitesurf is Cloudflare’s new stateless, highly scalable, and cost-effective web browser that runs entirely on top of Workers and was designed specifically for the Agentic Cloud.
- The next generation of MCP
cloudflare· 06-ago
The next version of MCP has a rewritten, stateless core that just works on Workers. We cover upgrades to the protocol, the new feature lifecycle and SDK migration path, and hear from early adopters already running it in production.
- The Agent Access Model
cloudflare· 05-ago
The Agent Access Model proposes a new architecture to secure task-scoped agents using strict identity brokering, continuous mediation, and stateful trust.
- Building an open Agentic Internet: readable, discoverable, callable, and payable
cloudflare· 06-ago
Agents are a new kind of visitor. They don't render CSS or click ads, but they have a paying human on the other end. Block them and you block your customer. We're building the open tools and protocols so publishers and agents can cooperate and not collide.
- Give any website a WebMCP interface
cloudflare· 06-ago
Today we're launching a developer preview of WebMCP on Cloudflare. With one switch, any site becomes usable by browser AI agents — no new APIs, no origin changes — while the human stays in control and creators keep their traffic.
- addyosmani/agent-skills
github-trending
Production-grade engineering skills for AI coding agents. Agent Skills Production-grade engineering skills for AI coding agents. Skills encode the workflows, quality gates, and best practices that senior engineers use when building software. These ones are packaged so AI agents follow them consistently across every phase of development. DEFINE PLAN BUILD VERIFY REVIEW SHIP ┌──────┐ ┌──────┐ ┌──────┐ ┌──────┐ ┌──────┐ ┌──────┐ │ Idea │ ───▶ │ Spec │ ───▶ │ Code │ ───▶ │ Test │ ───▶ │ QA │ ───▶ │ Go │ │Refine│ │ PRD │ │ Impl │ │Debug │ │ Gate │ │ Live │ └──────┘ └──────┘ └──────┘ └──────┘ └──────┘ └──────┘ /spec /plan /build /test /review /ship Commands 8 slash commands that map to the development lifecycle. Each one activates the right skills automatically. What you're doing Command Key pr…
- goauthentik/authentik
github-trending
The authentication glue you need. What is authentik? authentik is an open-source Identity Provider (IdP) for modern SSO. It supports SAML, OAuth2/OIDC, LDAP, RADIUS, and more, designed for self-hosting from small labs to large production clusters. Our enterprise offering is available for organizations to securely replace existing IdPs such as Okta, Auth0, Entra ID, and Ping Identity for robust, large-scale identity management. Installation Docker Compose: recommended for small/test setups. See the documentation. Kubernetes (Helm Chart): recommended for larger setups. See the documentation and the Helm chart repository. AWS CloudFormation: deploy on AWS using our official templates. See the documentation. DigitalOcean Marketplace: one-click deployment via the official Marketplace app. See …
- google/guava
github-trending
Google core libraries for Java Guava: Google Core Libraries for Java Guava is a set of core Java libraries from Google that includes new collection types (such as multimap and multiset), immutable collections, a graph library, and utilities for concurrency, I/O, hashing, primitives, strings, and more! It is widely used on most Java projects within Google, and widely used by many other companies as well. Guava comes in two flavors: The JRE flavor requires JDK 1.8 or higher. If you need support for Android, use the Android flavor. You can find the Android Guava source in the android directory. Adding Guava to your build Guava's Maven group ID is com.google.guava, and its artifact ID is guava. Guava provides two different "flavors": one for use on a (Java 8+) JRE and one for use on Android o…
- huangruiteng/loopx
github-trending
Lightweight loop engineering state kernel for long-running AI agent teams. Agent-loop agnostic across Codex, Claude Code, and other coding agents, with durable goals, quota-aware auto-wake, executable todos, evidence logs, and verifiable handoffs. LoopX The local control plane for long-running AI agent work. Keep objectives, gates, todos, evidence, quota, and handoffs stable while Codex, Claude Code, Cursor, or your own runtime executes bounded turns. Public website · Docs · Try LoopX · See real loops · How it works · User manual · 简体中文 把会干活的 Agent，接成可管理、可复盘、可持续改进的数字员工。 A lightweight state kernel and agent-agnostic local control plane for loop engineering, LoopX keeps long-running work reviewable, restartable, and easier to hand off across turns, tools, and agents. It does not replace you…
- mattpocock/skills
github-trending
Skills for Real Engineers. Straight from my .agents directory. Skills For Real Engineers My agent skills that I use every day to do real engineering - not vibe coding. Developing real applications is hard. Approaches like GSD, BMAD, and Spec-Kit try to help by owning the process. But while doing so, they take away your control and make bugs in the process hard to resolve. These skills are designed to be small, easy to adapt, and composable. They work with any model. They're based on decades of engineering experience. Hack around with them. Make them your own. Enjoy. If you want to keep up with changes to these skills, and any new ones I create, you can join ~60,000 other devs on my newsletter: Sign Up To The Newsletter Installation (30-second setup) Two ways in, two philosophies. The Clau…
- Significant-Gravitas/AutoGPT
github-trending
AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters. AutoGPT — AI agents that finish the work Get 10 hours back every week. Describe what you want done. AutoGPT builds the agent, runs it, and reports back. Get started · Tour · Pricing · Docs · Discord · Self-host The open-source platform for AI agents AutoGPT lets you build, deploy, and run AI agents that carry out complete workflows. Describe an outcome in plain English or shape every step in the visual builder, then run the agent on demand, on a schedule, or from a trigger. 185,000+ GitHub stars. Cited by: “Next frontier of prompt engineering imo: ‘AutoGPTs’.” Andrej Karpathy, founding member of OpenAI “If you have a phone you can run Aut…
- TapXWorld/ChinaTextbook
github-trending
所有小初高、大学PDF教材。 项目的由来 虽然国内教育网站已提供免费资源，但大多数普通人获取信息的途径依然受限。有些人利用这一点，在某站上销售这些带有私人水印的资源。为了应对这种情况，我计划将这些资源集中并开源，以促进义务教育的普及和消除地区间的教育贫困。 还有一个最重要的原因是，希望海外华人能够让自己的孩子继续了解国内教育。 学习数学 希望未来出现更多不是为了考学而读书的人。 小学数学 一年级上册 一年级下册 二年级上册 二年级下册 三年级上册 三年级下册 四年级上册 四年级下册 五年级上册 五年级下册 六年级上册 六年级下册 初中数学 初一上册 初一下册 初二上册 初二下册 初三上册 初三下册 高中数学 目录 大学数学 高等数学 线性代数 离散数学 概率论 更多数学资料-(大学数学网) 问题：如何合并被拆分的文件？ 由于 GitHub 对单个文件的上传有最大限制，超过 100MB 的文件会被拒绝上传，超过 50MB 的文件上传时会收到警告。因此，文件大小超过 50MB 的文件会被拆分成每个 35MB 的多个文件。 示例 文件被拆分的示例： 义务教育教科书 · 数学一年级上册.pdf.1 义务教育教科书 · 数学一年级上册.pdf.2 解决办法 要合并这些被拆分的文件，您只需执行以下步骤(其他操作系统同理)： 将合并程序 mergePDFs-windows-amd64.exe 下载到包含 PDF 文件的文件夹中。 确保 mergePDFs-windows-amd64.exe 和被拆分的 PDF 文件在同一目录下。 双击 mergePDFs-windows-amd64.exe 程序即可自动完成文件合并。 下载方式 您可以通过以下链接，下载文件合并程序： 下载文件合并程序 文件和程序示例 mergePDFs-windows-amd64.exe 义务教育教科书 · 数学一年级…
- tirth8205/code-review-graph
github-trending
Local-first code intelligence graph for MCP and CLI. Builds a persistent map of your codebase so AI coding tools read only what matters, with benchmarked context reductions on reviews and large-repo workflows. code-review-graph Stop burning tokens. Start reviewing smarter. English | 简体中文 | 日本語 | 한국어 | हिन्दी Usage · Commands · FAQ · Troubleshooting · GitHub Action · Reproducing the benchmarks · Roadmap AI coding tools can end up re-reading large parts of your codebase on review tasks. code-review-graph fixes that. It builds a structural map of your code with Tree-sitter, tracks changes incrementally, and gives your AI assistant precise context via MCP so it reads only what matters. Quick Start pip install code-review-graph # or: pipx install code-review-graph code-review-graph install # a…
- cloudflare/computer
github-trending
Give your agent a computer 👾 Cloudflare Computer Cloudflare Computer is a virtual filesystem that lives inside a Durable Object. The Durable Object holds the authoritative state in SQLite and exposes one pluggable execution surface through workspace.runtime. Three backends ship today: Container projects the SQLite state into a sandbox container as a real FUSE mount. A sandbox-side daemon (computerd) mounts the state as a filesystem and syncs changes back over a capnweb RPC channel. Full Linux userland, real binaries, real network. Isolate shell runs just-bash in a Dynamic Worker. It reaches the authoritative Workspace over Workers RPC, so there is no second store or sync round trip. Isolate JavaScript runs an ECMAScript module in a fresh Dynamic Worker with structured input/results, dura…
- [AINews] Jeff, Sanjay, Oriol, and Quoc depart DeepMind; Demis to Chair; Koray to SVP — what is going on at GDM???
latentspace· 06-ago
The end of an era.
- The Conductor Developer
martin-fowler· 31-jul
TL;DR Why I think software development is starting to feel a little more like conducting an orchestra. There’s a shift happening in software development that I don’t think we’re talking about clearly enough. For the last couple of years we’ve framed AI as a productivity tool. How much faster can it write code? How many more features can we ship? How much cheaper can we build software? I think that’s the wrong question, but I understand why. The first thing AI became good at was writing code, so naturally that’s where we focused. As AI got better at coding, I expected the bottlenecks to move through the software delivery lifecycle: from coding to design and specification, architecture, then verification. And they have. We spent a lot of time at the most recent FOSE event discussing how we …
- Fragments: August 4
martin-fowler· 04-ago
There’s been a fair bit of publicity of the Open AI “rogue agent” that hacked into Hugging Face. This prompted Anthropic to check what their models were up to and, to my complete lack of surprise, discovered three incidents where models had gained unauthorized access to data in other organizations. Simon Wilison concluded: It’s abundantly clear now that running evals of cyberattack potential in models is a spectacularly risky business. Every AI lab needs to pay attention to this. Keeping a close eye on what’s happening in those sandboxes is crucial It strikes me that this is akin to a virus escaping from a laboratory. It makes clear that the model builders are not putting sufficient controls in place to prevent these lab escapes. They are morally responsible for any consequences of this, …
- Modeling Device Capabilities for Analytics
netflix-tech· 31-jul
- One-shotting a Raccoon Heist game using Claude Fable 5
simonw· 05-ago
<p>Back in 2022 <a href="https://twitter.com/simonw/status/1555626060384911360">I tweeted</a> screenshots of a game concept generated by GPT-3 and some concept "art" created using DALL-E. Today, on the fourth anniversary of that tweet, I decided to see if Claude Fable 5 (running in <a href="https://code.claude.com/docs/en/claude-code-on-the-web">Claude Code for web</a>) could build the entire game from the content of that tweet. It did a pretty good job of it!</p> <p>You can <a href="https://simonw.github.io/raccoon-heist/">play the game here</a>. Here's <a href="https://github.com/simonw/raccoon-heist/">the GitHub repo</a>, and a short video demo:</p> <p><video controls="controls" preload="none" poster="https://static.simonwillison.net/static/2026/raccoon-heist-poster.jpg" width="1280" h…
- Third-party cyber evaluations involving OpenAI models
simonw· 05-ago
<p><strong><a href="https://openai.com/index/third-party-cyber-evaluations-involving-openai-models/">Third-party cyber evaluations involving OpenAI models</a></strong></p> And <em>another one</em>. I had to create a <a href="https://simonwillison.net/tags/accidental-cyberattacks/">accidental-cyberattacks tag</a> to keep track of them all!</p> <p>This post from OpenAI covers both the UK AI Safety Institute attack (see <a href="https://simonwillison.net/2026/Aug/5/incident-report/">my previous post</a>) and another attack enabled by <a href="https://www.irregular.com">Irregular</a>:</p> <blockquote> <p>Irregular, one of our external cybersecurity testing partners, was running Capture-the-Flag-style evaluations intended to be isolated from the internet, but a testing-environment misconfigura…
- An AI model from Meta also hacked another company during testing
simonw· 06-ago
<p><strong><a href="https://www.cnn.com/2026/08/05/tech/meta-ai-hacking">An AI model from Meta also hacked another company during testing</a></strong></p> Stop me if you've <a href="https://simonwillison.net/tags/accidental-cyberattacks/">heard this one before</a>:</p> <blockquote> <p>An AI model from the parent company of Facebook and Instagram hacked into another company’s systems during cybersecurity testing, a spokesperson confirmed on Wednesday.</p> <p>Meta says the breach occurred because of an inadvertent error during testing of the model, similar to previously disclosed incidents with OpenAI and Anthropic.</p> <p>“A misconfiguration by Irregular, an independent testing company Meta uses, inadvertently allowed one of our models access to the internet during evaluation,” the Meta sp…
- datasette 0.65.3
simonw· 06-ago
<p><strong>Release:</strong> <a href="https://github.com/simonw/datasette/releases/tag/0.65.3">datasette 0.65.3</a></p> <p>Back-ported the SQL Injection security fix from <a href="https://simonwillison.net/2026/Aug/6/datasette/">1.0a38</a>.</p> <p>Tags: <a href="https://simonwillison.net/tags/datasette">datasette</a></p>
- datasette 1.0a38
simonw· 06-ago
<p><strong>Release:</strong> <a href="https://github.com/simonw/datasette/releases/tag/1.0a38">datasette 1.0a38</a></p> <blockquote> <p>This release fixes a <strong>SQL injection</strong> security issue that affects Datasette instances that serve a <strong>mixture of public and private tables</strong> in the same database, with access configured using the <a href="https://docs.datasette.io/en/latest/authentication.html">Datasette permissions system</a>.</p> <p>Site administrators who serve private tables in this way are advised to disable the <a href="https://docs.datasette.io/en/latest/authentication.html#execute-sql">execute-sql permission</a> <actions_execute_sql>` on that database to prevent users from accessing private tables using raw SQL queries. The bug that has been fixed would h…
- Simon Willison on Technical Blogging
simonw· 06-ago
<p><strong><a href="https://writethatblog.substack.com/p/simon-willison-on-technical-blogging">Simon Willison on Technical Blogging</a></strong></p> I was interviewed by Cynthia Dunlop for her "Write that blog!" series back in January, but I just realized I never linked to the interview from my own blog!</p> <p>It includes my answers to the following questions:</p> <ul> <li>Why did you start blogging – and why do you continue?</li> <li>What has been the most surprising impact of blogging for you?</li> <li>What blog post are you most proud of and why?</li> <li>What post was the most difficult to write and how did you tackle it?</li> <li>Any lessons learned that you want to share with the community?</li> <li>Your advice for people just getting started with blogging?</li> <li>A few blogs tha…
- Introducing Muse Code and Muse Spark 1.2
simonw· 05-ago
<p><strong><a href="https://research.meta.ai/blog/introducing-muse-code-and-muse-spark-1-2">Introducing Muse Code and Muse Spark 1.2</a></strong></p> Yet more evidence that the most important characteristic of any model these days is long-sequence agentic tool calling. Meta shipped their own coding agent as part of getting that to work!</p> <blockquote> <p>Muse Spark 1.2 is a coding-focused update to Muse Spark 1.1, with improvements in code generation, complex debugging, codebase understanding, and end-to-end developer workflows. In Muse Spark 1.2, we significantly scaled up training compute on coding tasks while expanding training environment diversity. The model also maintains its strength in other key areas like general agents. [...]</p> <p>We co-trained Muse Spark 1.2 with Muse Code …
- Incident Report: unsanctioned agent behaviour during cyber testing
simonw· 05-ago
<p><strong><a href="https://www.aisi.gov.uk/blog/incident-report-unsanctioned-agent-behaviour-during-cyber-testing">Incident Report: unsanctioned agent behaviour during cyber testing</a></strong></p> It happened <em>again</em>. This time it was the UK government's AI Security Institute who accidentally attacked other companies while running an evaluation with models with the safety filters turned off. From <a href="https://cdn.prod.website-files.com/663bd486c5e4c81588db7a1d/6a724858f7db25c81487016d_Security%20Incident%20INC-2026-07-28-01.pdf">their technical paper</a> (PDF):</p> <blockquote> <p>During a cyber evaluation, from 25 to 28 July 2026, AI agents engaged in sustained, unsanctioned activity directed at what were, in practice, real people and organisations. These attempts were unsu…