- OriginBlame: Record- and Token-Level Data Provenance for AI Training Datasets
arxiv-ai· 17-jul
arXiv:2607.13037v1 Announce Type: new Abstract: When a data contributor requests removal, model trainers face a practical gap: unlearning algorithms require a forget set, yet no tool can locate which training records belong to a given author. Existing provenance systems operate at file or dataset level, forcing catastrophic over-deletion. We present ob, a record- and token-level data provenance system that propagates author identity through data processing pipelines and resolves revocation requests into precise forget sets via deterministic queries. Evaluation on 219,555 Wikipedia pages demonstrates that record-level provenance eliminates dataset-level over-deletion (from 101x to 1.3x), while integration adds 1.3-4.0% throughput overhead (HuggingFace) and 2.1-19.0% (Datatrove) on wiki dat…
- SPINE: Bridging the Cyber-Physical Gap with Agentic AI
arxiv-ai· 17-jul
arXiv:2607.13049v1 Announce Type: new Abstract: Foundation models have given robots a sophisticated brain for complex decision-making, yet deploying that intelligence into a physical platform still demands tedious, expert-driven calibration. This deployment gap, the robot's spinal cord, remains a primary bottleneck to scalable Embodied AI. Hence, we propose SPINE (Scalable Physical Integration with ageNtic Expertise): an agentic framework for systematically debugging and deploying bimanual robots with minimal robotics expertise. SPINE's harness comprises two orchestrated multi-agent workflows: a profile builder that creates robot-specific context, and a debugger that cycles through diagnosis, repair, and validation until teleoperation works. Across seven DOBOT X-Trainer debugging scenario…
- Probabilistic Extension of Neuro-Symbolic AGI Robots based on Belnap's Typed Intensional FOL
arxiv-ai· 17-jul
arXiv:2607.13073v1 Announce Type: new Abstract: Neuro-symbolic AI based on $IFOL_B$ is a way to combine neural learning and symbolic reasoning to overcome limitations of purely neural systems (like lack of interpretability and logical structure) with formal logical machinery for self-reference. In this paper we expand the cognitive power of $IFOL_B$ by using the probability computation for the currently unknown sentences, based on Nilsson's probability structure for the $IFOL_B$. We introduce the global symmetry transformation that preserves the current knowledge database and logical deduction, and the local one used for real-time decisions about concrete (sub)problems that involve only a very strict subset of $IFOL_B$ predicates. The computation of probability density function $KI$ in bo…
- Interventional Grounding Audits: Black-Box Premise-Dependency Tests for LLM Chain-of-Thought via Predicate Substitution
arxiv-ai· 17-jul
arXiv:2607.13069v1 Announce Type: new Abstract: Large language models produce chain-of-thought (CoT) reasoning that appears logically sound yet may not genuinely depend on its stated premises. We introduce interventional grounding audits, a black-box, step-level test of premise dependency: we intervene on a single premise by substituting its target predicate with a fresh symbol, re-run the model, and check whether each reasoning step's normalized conclusion (canonical predicate form) changes. We evaluate on ProntoQA, a synthetic multi-hop deductive reasoning benchmark with gold proof trees, where step-level premise dependencies are known. Applied to 50 ProntoQA problems with GPT-4o, our method achieves F1 = 0.806 on detecting proof-tree dependencies (F1 = 0.885 on predicate-determining de…
- Self-Improvements in Modern Agentic Systems: A Survey
arxiv-ai· 17-jul
arXiv:2607.13104v1 Announce Type: new Abstract: Self-improving autonomous agents are moving from research prototypes to deployed systems. The primary goal is controllable evolution, or adaptation, from experience with minimal or even no human input. This survey frames modern self-improving agents as adaptive systems that convert experience into accumulated capability gains. We offer a system-level framework that represents a modern agent as a configuration coupling a foundation model with an operational scaffold of prompts, memory, tools, and control logic. Within this framework, self-improvement is formalized as a self-induced update operator that obtains and commits updates to model parameters or scaffold components. We organize prior work by update target and by the signals that drive …
- Improving Molecular Property Prediction in Small Language Models Using Graph-based Tools
arxiv-ai· 17-jul
arXiv:2607.13115v1 Announce Type: new Abstract: Small language models (SLMs) have shown promise for zero-shot molecular property prediction from SMILES strings, yet they often suffer from structural blindness because sequence representations under-specify key graph-topological cues. We propose a modular Context-Augmented Prompting framework that enables agentic tool use at inference time: a trained GNN expert model provides a predictive hint with confidence, and a GNN extracts an instance-specific explanatory subgraph (e.g., a subgraph SMILES and an accompanying explanatory paragraph). We evaluate three commonly used SLMs on MUTAG and Tox21 under five prompting configurations ranging from SMILES-only to using all available tools at hand. Across two datasets, enriching prompts with graph-d…
- Oracle Agent Memory as an Enterprise Memory Substrate for Long-Horizon AI Agents
arxiv-ai· 17-jul
arXiv:2607.13157v1 Announce Type: new Abstract: Agent memory is a systems problem for long-horizon agents. Practical deployments require retention of task state across extended conversations, recovery of user-specific facts and preferences across sessions, and accumulation of procedural knowledge from prior outcomes. These requirements extend beyond document retrieval: a memory layer must determine which interactions become durable state, how that state is scoped, how it is retrieved under latency constraints, and how it is revised or removed over time. This report studies Oracle Agent Memory as a database-native memory substrate built on Oracle Database. Three themes organize the discussion: memory as a lifecycle spanning ingestion, extraction, consolidation, retrieval, summarization, an…
- Learning Safe Agent Behaviour from Human Preferences and Justifications via World Models
arxiv-ai· 17-jul
arXiv:2607.13172v1 Announce Type: new Abstract: We address the problem of safely training an agent policy and deploying a good and safe policy, in settings where the environment dynamics are unknown and no suitable reward function is available. In the context of safety-critical environments, we consider traditional reinforcement learning impractical and resort to the resource of human input. We introduce DROPJ, a human-centred method for both safe training and deployment. We first learn a world model (a learned simulator) from a dataset of prior real-world trajectories. A human then plays the game in this learned simulator to extract several informative simulated trajectories. From these, we sample pairs of simulated trajectory segments and elicit from a human their preference over these …
- CayleyR: Solving the TopSpin puzzle via cycle intersection
arxiv-ai· 17-jul
arXiv:2607.13219v1 Announce Type: new Abstract: We present cayleyR, an R package for solving permutation puzzles by detecting cycle intersections in Cayley graphs. The core algorithm performs an iterative bidirectional search: from both the initial and target permutation states, random operation sequences generate cycles in the Cayley graph of the symmetric group Sn; their intersection yields a connecting path. When no direct intersection is found, a distance-guided bridge selection narrows the gap, and the process repeats. The package targets the TopSpin(n,k) puzzle, whose state space is a Cayley graph of Sn generated by a cyclic shift and a prefix reversal. We describe the mathematical framework, the algorithm, and its implementation, which combines a C++ hash-indexed state store with o…
- Networked Intelligence: Active Shared Context Graphs for Human-AI Team Science
arxiv-ai· 17-jul
arXiv:2607.13220v2 Announce Type: new Abstract: Most AI-for-science systems focus on scaling a single reasoning process by using better models, larger context windows, long-horizon agentic execution, or digital co-scientists working with one principal user. However, challenging scientific problems are rarely solved by one reasoner alone. They are solved by teams whose members carry different priors, experimental background, tacit knowledge, and domain-trained intuitions. The open problem is therefore not only how to scale models, but how to develop "networked intelligence", scaling the connections between humans and AI systems so that a result or hypothesis produced in one context reaches another person, agent, instrument or robot that can act on it. We introduce Mycelium, an active share…
- How Microsoft Ships AI Agents at Enterprise Scale
bytebytego· 13-jul
To understand what it actually takes to ship agents at that scale, we spoke with Marco Casalaina, VP of Products for Microsoft Core AI.Jul 13 • ByteByteGo2938
- A Guide to Multi-Tenancy: Benefits and Challenges
bytebytego· 16-jul
In this article, we will understand multi-tenant architecture from the basics, along with its various benefits and challenges.13 hrs ago • ByteByteGo1474
- AI Customer Support at Scale: The Travel Industry’s $Billion Bet
bytebytego· 15-jul
In this article, we will look more closely at the different solutions by following the support pipeline from first principles, show why a tail of cases…Jul 15 • ByteByteGo2664
- EP221: How Docker Works Under the Hood
bytebytego· 11-jul
A Docker container starts with a single command, but that command has to be turned into a running Linux process. Here is what actually happens.Jul 11 • ByteByteGo311616
- How LLMs Learn to Be Helpful (RLHF vs DPO)
bytebytego· 14-jul
In this article, we will look at how that learning actually happens, starting with why instruction-following alone falls short, then walking through the…Jul 14 • ByteByteGo31426
- Loading...
claude-changelog
Loading...Loading...Loading...Loading...Loading...Loading...Loading...Loading...Loading...Loading...Loading...Loading...Loading...Loading...Loading...
- Claude Platform release notes
claude-changelog
Release notesCopy pageUpdates to the Claude Platform, including the Claude API, client SDKs, and the Claude Console.Copy pageFor release notes on Claude Apps, see the Release notes for Claude Apps in the Claude Help Center.For updates to Claude Code, see the complete CHANGELOG.md in the claude-code repository. July 15, 2026 Mid-conversation system messages are available on Claude Fable 5, Claude Mythos 5, and Claude Opus 4.8, on the Claude API, Claude in Amazon Bedrock, and Google Cloud. No beta header is required. This corrects earlier availability notes. July 14, 2026 You can now manage the people in your Claude Enterprise (claude.ai) organization with the Admin API, in beta for all Claude Enterprise organizations: list members and look them up by email address, change a member's r…
- A broken DNSSEC rollover took down .al. Now 1.1.1.1 tells you when validation is bypassed
cloudflare· 14-jul
When a failed DNSSEC key rollover took down the .al TLD, we deployed a Negative Trust Anchor to restore resolution. This time, though, clients didn't have to take our word for it: 1.1.1.1 returned EDE 33, a new DNS error code that signals directly in the response that DNSSEC validation was bypassed.
- Introducing Precursor: detecting agentic behavior with continuous client-side signals
cloudflare· 13-jul
Precursor, our new continuous behavioral validation engine for bot management, offers visibility into how humans and bots actually interact across the full user journey. By turning session-level behavior into bot detection signals, it identifies advanced automation with higher precision — while reducing friction for legitimate users.
- Improving Smart Tiered Cache for Public Cloud Regions
cloudflare· 10-jul
Smart Tiered Cache allows for precise upper tier selection for origins hosted on AWS, GCP, Azure, and Oracle Cloud with customer-provided cloud region hints.
- openinterpreter/openinterpreter
github-trending
A coding agent for open models like Kimi K3 Open Interpreter A coding agent optimized for low-cost models. Blog post ↗ Note Today: Kimi K3 is here. We have reimplemented the provider-recommended Kimi Code harness in Rust, giving you maximum K3 performance with a Codex-like interface. Kimi Docs → Installation macOS and Linux: curl -fsSL https://www.openinterpreter.com/install | sh Windows: irm https://www.openinterpreter.com/install.ps1 | iex Then type i or interpreter in your terminal to start a session. Harness Emulation Open Interpreter is a fork of OpenAI's Codex, with a focus on emulating the agent harness that gets the best performance out of low-cost models. Use /harness to switch the active harness: > /harness native claude-code claude-code-bare zcode kimi-code kimi-cli qwen-code d…
- PrismML-Eng/Bonsai-demo
github-trending
Bonsai Demo Bonsai Demo Website | GitHub | Discord HF Collections: Bonsai 27B · Bonsai (1-bit) · Ternary-Bonsai Whitepapers: Bonsai 27B · 1-bit Bonsai 8B · Ternary-Bonsai 8B Using this demo repository you can run Bonsai (1-bit) and Ternary-Bonsai language models locally on Mac (Metal), Linux/Windows (CUDA, Vulkan, ROCm), or CPU. 🌱 New: Bonsai 27B The family's newest and largest generation, and its first vision-language models (Bonsai 27B collection): Vision: send photos, screenshots, and PDFs; ask about them (see VISION.md). Agentic tool calling: native OpenAI-style tool_calls with full round-trips, plus MCP servers in both demo UIs (see TOOLS.md). Thinking: a reasoning model; pick the reasoning effort per chat in the UI or budget it per request. Long context: 256k+ token conversations. …
- YimMenu/YimMenuV2
github-trending
Experimental menu for GTA 5: Enhanced YimMenuV2 Experimental menu for GTA 5: Enhanced How to use Download the latest version of FSL from here and place version.dll in your GTA V directory. Using FSL is now optional but highly recommended for account safety Download YimMenuV2 from GitHub Releases Download an injector, such as Xenos Open Rockstar Launcher, select Grand Theft Auto V Enhanced, go to settings, and disable BattlEye. If you are using Steam or Epic Games, you may have to pass the -nobattleye command line parameter as well Launch GTA V, then use your injector to inject YimMenuV2.dll at the main menu How to open the menu? Press the INSERT key or Ctrl+\ to open the menu Common issues I keep getting desynced from public sessions every five minutes We currently do not have a BattlEye …
- apache/ossie
github-trending
Apache Ossie, industry wide specification effort to standardize how we exchange semantic metadata across analytics, AI and BI platforms, providing a vendor neutral, single source of truth for semantic data Apache Ossie (incubating) Apache Ossie is a collaborative, open-source effort dedicated to standardizing and streamlining semantic model exchange and utilization across the diverse array of tools and platforms within the data analytics, AI, and BI ecosystem. Our shared vision is to establish a common, vendor-agnostic semantic model specification, promoting unparalleled interoperability, efficiency, and collaboration among all participants. By providing a single, consistent source of truth, this vendor-agnostic standard ensures that your data's definitions and value remain consistent as …
- lobehub/lobehub
github-trending
🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team. LobeHub LobeHub organizes your agents into 7×24 operation. It hires, schedules, reports on your entire AI team. You stay in charge — without staying online. English · 简体中文 · Official Site · Changelog · Documents · Blog · Feedback Share LobeHub Repository Your Chief Agent Operator Table of contents TOC 👋🏻 Getting Started & Join Our Community ✨ Features Operator: Agents as the Unit of Work Create: Agents as the Unit of Work Collaborate: Scale New Forms of Collaboration Networks Evolve: Co-evolution of Humans and Agents 🛳 Self Hosting A Deploying with Vercel, Zeabur , Sealos or Alibaba Cloud B Deploying with Docker Environment Variable 📦 Ecosys…
- PostHog/posthog
github-trending
🦔 PostHog is the leading platform for building self-driving products. Our developer tools – AI observability, analytics, session replay, flags, experiments, error tracking, logs, and more – capture all the context agents need to diagnose problems, uncover opportunities, and ship fixes. Steer it all from Slack, web, desktop, or the MCP. Docs - Community - Roadmap - Why PostHog? - Changelog - Bug reports PostHog is the open source platform for building self-driving products PostHog provides every tool you need to build a successful product, and captures all the context agents need to proactively diagnose problems, uncover opportunities, and ship fixes: Self-driving mode: Turn signals in your product data (errors, rage clicks, failed queries, and more) into researched reports and pull reque…
- I measured whether AI writes hollower tests than humans. It doesn't
hn-ai· 17-jul
Article URL: https://driivai.github.io/voidguard/ Comments URL: https://news.ycombinator.com/item?id=48943442 Points: 1 # Comments: 0
- EU forces Google to share search data and open Android to rival AI companies
hn-ai· 17-jul
Article URL: https://apnews.com/article/eu-google-android-antitrust-184b3067120e56d858cb8c81aee26d45 Comments URL: https://news.ycombinator.com/item?id=48943496 Points: 2 # Comments: 0
- Show HN: Scribe, a CLI that builds AI agent memory from your repos and sessions
hn-ai· 17-jul
Article URL: https://getscribe.dev/ Comments URL: https://news.ycombinator.com/item?id=48943417 Points: 2 # Comments: 0
- Lightport – a maintained fork of Portkey AI gateway
hn-ai· 17-jul
Article URL: https://github.com/glama-ai/lightport Comments URL: https://news.ycombinator.com/item?id=48943532 Points: 1 # Comments: 0
- China Just Dropped Another Bomb on America's Frontier AI Companies
hn-ai· 17-jul
Article URL: https://gizmodo.com/china-just-dropped-another-bomb-on-americas-frontier-ai-companies-2000786670 Comments URL: https://news.ycombinator.com/item?id=48943385 Points: 3 # Comments: 0
- AI is changing what we can do. Who we become is still our choice
hn-ai· 17-jul
Article URL: https://humanistreview.ai/issue-1/appiah-ai-moral-character/ Comments URL: https://news.ycombinator.com/item?id=48943333 Points: 1 # Comments: 0
- Show HN: No AI, just plain old JavaScript and Python RAD
hn-ai· 17-jul
Article URL: https://jam-py-v7.github.io/jam-py-v7/ Comments URL: https://news.ycombinator.com/item?id=48943410 Points: 1 # Comments: 0
- Ask HN: Workflow Automation vs AI Agents?
hn-ai· 17-jul
A while ago, workflow automation tools (Zapier, Make, n8n) were booming. Now AI agents start eating into that space, OpenAI's Workspace Agents, Lindy, Claude's agentic features, etc. i understand both have tradeoffs, but i'm trying to understand real-world usage: 1. why workflow automarion over AI Agents? What does it still do better? 2. if you've using AI agents, what made you switch? What do you trust them with? 3. why not use either of them? feel free to give your opinion, im curious Comments URL: https://news.ycombinator.com/item?id=48943616 Points: 1 # Comments: 0
- Ask HN: I feel like I've lost my identity due to AI
hn-ai· 17-jul
I am a self taught software developer. I was previously an electrical engineer. For the past 10 years I’ve spent almost all my spare time reading and learning and lifting weights - software, electrical engineering, reading, and creating what I thought were novel projects - mostly in the software domain. For the longest time I’ve only been interested in ideas that haven’t really been done before, maybe that’s because I have adhd? But lots of the ideas I pursue and build are relatively novel or unique in some way, not derivative. Anyway, then LLMs happened. Now any idiot can just ask an AI for ideas or to think for them, or to build software for them in a single shot. Initially I was really excited by LLMs because I thought great, something to bounce ideas off. Now LLMs have far surpassed a…
- Xi pitches China as leader of new global AI order, challenging US dominance
hn-ai· 17-jul
Article URL: https://www.reuters.com/world/asia-pacific/chinas-xi-promotes-chinas-commitment-ai-access-speech-shanghai-conference-2026-07-17/ Comments URL: https://news.ycombinator.com/item?id=48943519 Points: 2 # Comments: 3
- 🔬 The Lab of the Future Should Feel Like a Data Center — Andy Beam & Rafa Gómez-Bombarelli, Lila Sciences
latentspace· 16-jul
Lila is betting that science, not the internet, is the last untapped source of training data. We went to find out what that actually looks like in a room full of robots.
- [AINews] Thinky's Inkling: 975B-A41B multimodal, new best American Apache 2.0 open model (with Inkling-Small, 276B-A12B)
latentspace· 16-jul
Thinky's first full LLM release is a banger and bonus: it's open weights!
- [AINews] Kimi K3 2.8T-A50B: the largest open model ever released; Opus 4.8-class at Sonnet 5 pricing
latentspace· 17-jul
a great week for open models continues.
- The Archaeologist’s Copilot
martin-fowler· 16-jul
When people think of legacy modernization, most folks aren't imagining the target environment will be Java 8. But this was the challenge facing Nik Malykhin when he needed to run a Java 1.5 codebase on today's hardware. His early use of LLMs gave plausible answers that did not hold up in the codebase. Progress came when he grounded the process in evidence, using AI to support analysis, validation in a stable Docker environment, and gradual refactoring protected by tests. The main takeaway is practical: AI was most useful when constrained by evidence, clear roles, and a step-by-step modernization strategy. more…
- Fragments: July 13
martin-fowler· 13-jul
Some more of my notes from Thoughtworks Future of Software Development Retreat. When we had our first retreat in Utah early this year, nobody had heard of Harness Engineering. This time we had a whole session on it. When comes to the guide side of harnesses, most of the discussion is about context management. While context windows have increased is size as models get more sophisticated, that doesn’t mean that models will properly focus on the right bits. Models typically only focus attention on part of the context, and to get the best behavior, we need to manage that focus. One attendee keeps their context small, limiting the agents.md file to less than 200 lines On the sensor side, we see more attention on computational sensors. Two patterns from one participant was shifting to languages…
- DSLs Enable Reliable Use of LLMs
martin-fowler· 14-jul
LLMs generate code incredibly fast, but to ensure they generate exactly what is intended, they need clear boundaries. Abstractions and Domain-Specific Languages (DSLs) provide a strong harness that guides LLMs right from the start. Unmesh Joshi describes how the example of Tickloom - a domain model and DSL for illustrating distributed system behavior - shows how we can use an LLM as a partner to iteratively build a DSL and as a natural language interface to use it. Such a DSL can act as the key source of truth for software systems in the world of LLMs. more…
- Building Service Topology at Scale: Architecture, Challenges, and Lessons Learned
netflix-tech· 13-jul
- Inkling: Our open-weights model
simonw· 16-jul
<p><strong><a href="https://thinkingmachines.ai/news/introducing-inkling/">Inkling: Our open-weights model</a></strong></p> Mira Murati's Thinking Machines Lab just released their first open-weights model. Inkling is "a Mixture-of-Experts transformer with 975B total parameters, 41B active" - an Apache-2.0 licensed multimodal model trained on 45 trillion tokens of text, images, audio and video.</p> <p>They're also promising Inkling-Small, a 276B (12B active) model, but that's still being tested and the weights will be released "once that work is complete".</p> <p>The <a href="https://thinkingmachines.ai/model-card/inkling/">model card</a> is much shorter than I've come to expect from US AI labs. It links to even shorter <a href="https://thinkingmachines.ai/training-data-documentation/">Tra…
- Kimi K3, and what we can still learn from the pelican benchmark
simonw· 16-jul
<p>Chinese AI lab Moonshot AI <a href="https://www.kimi.com/blog/kimi-k3">announced Kimi K3</a> this morning, describing it as their "most capable model to date, with 2.8 trillion parameters". It's currently available via their website and API, but an open weight release is promised "by July 27, 2026".</p> <p>Moonshot are calling this the first "open 3T-class model" (I guess they're rounding 2.8 trillion up to 3 trillion), taking the crown from <a href="https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro">DeepSeek's 1.6T v4 Pro</a>. Their <a href="https://www.kimi.com/blog/kimi-k3#full-benchmark-table">self-reported benchmarks</a> have K3 mostly beating Claude Opus 4.8 max and GPT-5.5 high, while losing out to Claude Fable 5 and GPT-5.6 Sol.</p> <p>A few highlights from the <a href="https:…
- Quoting Linus Torvalds
simonw· 16-jul
<blockquote cite="https://lore.kernel.org/linux-media/CAHk-=wi4zC+Ze8e+p3tMv8TtG_80KzsZ1syL9anBtmEh5Z40vg@mail.gmail.com/"><p>I realize that some people really dislike AI, but this is an area where I'm willing to absolutely put my foot down as the top-level maintainer.</p> <p>Linux is not one of those anti-AI projects, and if somebody has issues with that, they can do the open-source thing and fork it.</p> <p>Or just walk away.</p> <p>AI is a tool, just like other tools we use. And it's clearly a useful one.</p> <p>It may not have been that "clearly" even just a year ago, but it's no longer in question today.</p> <p>There are other questions around AI (like what the economy of it will actually look like in the end), but "is it useful" is no longer one of those questions. Anybody who doubt…
- Mermaid to ASCII art (mermaid-ascii)
simonw· 16-jul
<p><strong>Tool:</strong> <a href="https://tools.simonwillison.net/mermaid-ascii">Mermaid to ASCII art (mermaid-ascii)</a></p> <p>After building the <a href="https://simonwillison.net/2026/Jul/16/grok-mermaid/">Mermaid to ASCII tool based on Grok Build's Rust code</a> I learned that there's an older, more fully-featured Go library called <a href="https://github.com/AlexanderGrooff/mermaid-ascii">AlexanderGrooff/mermaid-ascii</a> that implements a similar pattern, so I had Claude Fable 5 compile that one to WebAssembly as well so I could compare the two.</p> <p>This one includes support for colors!</p> <p><img alt="Screenshot of a Mermaid diagram editor web app. A row of tab buttons reads: Flowchart, Multiple links, Subgraphs, Multi-line labels, Colors (selected, highlighted blue), Sequenc…
- Spot birds not golf
simonw· 17-jul
<p>Suggestion for hyperscalers feeling pressure over data center water use:</p> <p>Buy up a few exclusive country clubs, convert the golf courses into public parks, pay for guides and binoculars to get the previous members into birdwatching - help them embrace a more sustainable hobby!</p> <p>Google <a href="https://sustainability.google/reports/google-2026-environmental-report/">used 10.9 billion gallons in 2025</a>, so about 30 million gallons per day.</p> <p>The Coachella Valley has <a href="https://www.cvwd.org/167/Water-Conservation">120 golf courses each using ~800 acre-feet per year</a>, which is ~750,000 gallons per day.</p> <p>So Google buying up 40 of those courses (1/3) should do the trick.</p> <p>Tags: <a href="https://simonwillison.net/tags/ai-energy-usage">ai-energy-usage</a…
- Mermaid to Unicode box art (grok-mermaid)
simonw· 16-jul
<p><strong>Tool:</strong> <a href="https://tools.simonwillison.net/grok-mermaid">Mermaid to Unicode box art (grok-mermaid)</a></p> <p>While <a href="https://simonwillison.net/2026/Jul/15/grok-build/">exploring the codebase</a> for the newly open-sourced Grok CLI coding agent I came across <a href="https://github.com/xai-org/grok-build/blob/b189869b7755d2b482969acf6c92da3ecfeffd36/crates/codegen/xai-grok-markdown/src/mermaid.rs">xai-grok-markdown/src/mermaid.rs</a>, a "self-contained terminal renderer for Mermaid diagrams" written in Rust.</p> <p>I figured it would be fun to try that out in a browser via WebAssembly. Here's <a href="https://github.com/simonw/tools/pull/293#issue-4897479396">the prompt</a> I ran in Claude Code for web (Fable 5), and this is what the resulting tool looks lik…
- Firefox in WebAssembly
simonw· 16-jul
<p><strong><a href="https://developer.puter.com/labs/firefox-wasm/">Firefox in WebAssembly</a></strong></p> This is absurdly cool: Puter compiled Firefox to WebAssembly such that the whole browser runs in another browser.</p> <p>Here's my blog, running in Firefox, running in WebAssembly, running in Chrome:</p> <p><img alt="A Chrome window. The tab has the Firefox UI and has loaded my blog. On the right is the Chrome network panel showing that it loaded resources that include a 233MB gecko.wasm and an 18MB chrome-assets.tar.zst" src="https://static.simonwillison.net/static/2026/firefox-wasm.webp" /></p> <p>They chose Firefox/Gecko because it has strong single-process support. The project used an estimated $25,000 worth of Claude Opus and Fable tokens, but took advantage of a Claude Max sub…
- Quoting Thibault Sottiaux
simonw· 16-jul
<blockquote cite="https://twitter.com/thsottiaux/status/2077630111499882637"><p>On file deletions. We’ve investigated a handful of reports where GPT-5.6 unexpectedly deleted files. </p> <p>What we have found is that this most commonly occurs when:</p> <ul> <li>Full access mode is enabled and codex is run without sandboxing protections, including without auto review being enabled</li> <li>The model attempts to override the $HOME env var to define a temporary directory.</li> <li>The model makes an honest mistake and mistakenly deletes $HOME instead.</li> </ul></blockquote> <p class="cite">— <a href="https://twitter.com/thsottiaux/status/2077630111499882637">Thibault Sottiaux</a>, describing a pretty gnarly Codex bug</p> <p>Tags: <a href="https://simonwillison.net/tags/codex">codex</a>…
- xai-org/grok-build, now open source
simonw· 15-jul
<p><strong><a href="https://github.com/xai-org/grok-build">xai-org/grok-build, now open source</a></strong></p> xAI's <code>grok</code> CLI tool faced severe community backlash yesterday when it became apparent that running the command in a directory could upload that <em>entire directory</em> to xAI's Google Cloud buckets. One user <a href="https://x.com/a_green_being/status/2076598897779020159">reported</a> running it in their home directory and seeing it upload "my SSH keys, my password manager database, my documents, photos, videos, everything".</p> <p>I've not seen an official explanation for why it was doing this, but xAI did respond to the feedback (<a href="https://twitter.com/elonmusk/status/2076739687658496209">Musk</a>: "As a precautionary measure, all user data that was upload…
- How I tricked Claude into leaking your deepest, darkest secrets
simonw· 15-jul
<p><strong><a href="https://www.ayush.digital/blog/the-memory-heist">How I tricked Claude into leaking your deepest, darkest secrets</a></strong></p> I've <a href="https://simonwillison.net/2025/Sep/10/claude-web-fetch-tool/">been impressed</a> by the way the Claude <code>web_fetch</code> tool is designed to avoid data exfiltration attacks. Ayush Paul found a hole in that design.</p> <p>To recap: regular Claude chat is at risk of <a href="https://simonwillison.net/2025/Jun/16/the-lethal-trifecta/">lethal trifecta</a> attacks, because it has access to private data (in the form of memories of your past interactions) and has a tool for accessing online content which can both read hostile instructions and exfiltrate data through the URLs it accesses.</p> <p>Anthropic's protection is that <cod…