AI 日报 — 2026-07-11

A recent PDF, "GPT-5.6 Sol Ultra produces proof of the Cycle Double Cover Conjecture," has been shared online. The document, accessible via an OpenAI URL, claims to present a proof generated by an AI model. This development has garnered some attention, with 8 points and 1 comment on a Hacker News thread.

Onboard-CLI is an LLM-powered, AST-based tool for visualizing codebases. It uses Tree-sitter for deep parsing across multiple languages, generating structural graphs displayed on a React Flow canvas. Key features include an interactive visualizer, architecture drift detection, and commands for impact analysis and owner tracking. The tool aims to help developers understand complex code, enforce architectural boundaries, and maintain code health.

Apple's lawsuit against OpenAI stems from escalating tensions, particularly highlighted by the strained relationship between Tang Tan, OpenAI's hardware chief, and his former boss, John Ternus. The lawsuit reportedly involves an iPhone engineer, Chang Liu, who Apple claims took proprietary information when he moved to OpenAI's hardware division. This suggests a deeper conflict beyond just personnel changes, hinting at intellectual property concerns.

A contractor automated their 40-hour work week using Claude, a language model. The job involves extracting information from websites, a task that would be extremely time-consuming manually. The contractor had Claude write a script that efficiently performs the job, consuming minimal tokens. The contractor initiates the process, and Claude completes the work, allowing for daily submission. This automation is effective for straightforward tasks requiring specific field expertise.

A user compared Claude Sonnet 5 and Fable 5 on two coding tasks. For a RAG Debugger, Sonnet was cheaper but slower, while Fable was faster and preferred. For a complex Clash Royale-style game, Sonnet took longer, cost more, generated more tokens, and required multiple fixes, despite being advertised as cheaper. The user noted that Sonnet 5's cost escalated quickly for longer coding tasks, suggesting that cheaper per token doesn't always mean cheaper per completed task.

Apple has sued OpenAI, alleging trade secret theft by former Apple employees now working at OpenAI. The lawsuit claims individuals like Tang Tan and Chang Liu stole confidential information, including unreleased technologies and product designs. Apple states Tan used insider knowledge to interview candidates, directing them to bring Apple hardware and revealing project codenames. Liu allegedly downloaded thousands of pages of technical files. Apple seeks injunctive relief and damages, asserting OpenAI ignored initial concerns.

GPT-Live is a trending topic, garnering significant attention with 408 points and 274 comments on Hacker News. The discussion revolves around OpenAI's introduction of GPT-Live, indicating public interest in this new development from the company. The provided URLs point to the official announcement and the ongoing conversation.

BuzzRadr reports Claude is beta-testing a new "reflection dashboard" feature. This tool helps users understand and refine their AI usage by tracking activity, identifying patterns, and offering insights into how

FableCut is a browser-based, Premiere-style video editor designed for AI agent control. Its unique feature is exposing the entire timeline as a JSON document, allowing AI agents (like Claude Code) to edit videos by modifying this project file. The UI hot-reloads live, enabling simultaneous human and AI collaboration. It offers extensive editing, visual, motion, and text features, including AI background removal and the ability to remake videos from a reference.

GeoSQL is a new skill for data scientists and analysts, enhancing Claude, Codex, and GitHub Copilot for geospatial data tasks on various platforms like PostGIS and BigQuery. It operates locally or self-hosted, offering a 4x improvement on geospatial tasks by incorporating a "map in the loop" for visual validation and correction. GeoSQL explores warehouse metadata, writes spatial SQL, includes cost checks, and validates geometry, with optional Dekart integration for map rendering.

Prismata is a defense mechanism designed to secure autonomous web agents against cross-site prompt injection attacks. These attacks exploit agents' interpretation of natural language, allowing malicious content to hijack tasks. Prismata enforces "contextual least privilege" by dynamically labeling page content and restricting agent capabilities, inspired by integrity models. It redacts content and limits agent actions without requiring developer annotations. Prismata significantly reduces attack success in various web agent attacks while maintaining utility for legitimate tasks.

Mistral's Robostral Navigate is an 8B model enabling robots to autonomously navigate complex environments using only a single RGB camera. It achieves 76.6% success on unseen R2R-CE benchmarks, outperforming multi-sensor approaches. Built in-house with simulation-trained data and token-efficient techniques, it generalizes across robot types and adapts to real-world obstacles. The model combines pointing-based navigation with reinforcement learning for continuous improvement, paving the way for unified embodied AI.

Kastor offers a vendor-neutral, versionable, and reviewable solution for defining AI agents. It uses a typed, declarative spec in HCL for agents, tools, and prompts. A Go toolchain allows Kastor to generate runnable projects for frameworks like LangGraph or reconcile agents on hosted platforms with state management and drift detection. This provides a "Terraform-style" approach to managing AI agents, addressing the current lack of a unified source of truth in agent development.

A new project, Colibrì, enables running the 744B-parameter GLM-5.2 Mixture-of-Experts model on consumer machines with 25 GB RAM. It achieves this by streaming experts from disk, keeping only the dense part of the model (9.9 GB) resident in RAM. The pure C engine, with zero dependencies, utilizes techniques like MLA attention, DeepSeek-V3-style routing, and native MTP speculative decoding for efficient operation, despite cold starts being slow due to disk reads.

Since Claude Code's February 2025 launch, US software development job postings on Indeed have increased by approximately 15%. This growth contrasts sharply with a 7% decline in overall job postings during the same period. This trend suggests that agentic AI might be altering the typical relationship between AI exposure and job posting growth, as reported by Guillermo Gallacher from Indeed Hiring Lab.

Rowboat is an open-source, local-first desktop AI coworker for Mac, Windows, and Linux. It indexes user work into a knowledge graph, offering features like an email client with AI drafting, background agents, a built-in browser, and a meeting note-taker. Rowboat supports various AI models, integrates with popular products, and stores all data locally as Markdown, emphasizing long-lived knowledge and user control over data.

A user reports that the new 5.6 Ultra model is performing exceptionally well, particularly when compared to other AI models. They provided a prompt asking for a visual representation of nerves connecting teeth to the brain. The user noted that Fable flagged this request as unsafe, and Opus was unhelpful, highlighting 5.6 Ultra's superior performance in handling such requests. They have since switched their 20max plan from Claude to Chat due to 5.6 Ultra's effectiveness.

A new article aims to simplify the understanding of world models within the ML community. The author proposes a classification framework for different world model approaches and identifies emerging trends. Feedback is requested on the framework's completeness, clarity, and technical accuracy.

A user leveraged Claude to develop a suite of personalized Mac desktop applications for finance, accounting, and CRA filings, completely replacing QuickBooks. These custom apps offer real-time data synchronization and automated information flow, eliminating duplicate entries and providing a more intuitive system tailored to the user's workflow. The user is now curious about other innovative projects people have created with Claude.

The Strix Halo, despite being perceived as slow, offers significant value due to its low power consumption and versatility. Running at a worst-case scenario cost of $0.48 daily, it can handle 50tps on Qwen 3.6 35B while being compact and quiet. While Nvidia cards are faster, the Strix Halo's efficiency, size, and ability to host other services make it a compelling alternative, especially when considering total power budget compared to high-end GPUs like the A6000.

Databricks' benchmarks indicate that pi-coding-agent, utilizing minimal tools, is approximately two times cheaper and achieves a higher pass rate compared to CC/Codex. Additionally, GLM 5.2 performs on par with Opus 4.8 high, surpassing GPT 5.5 high and xhigh in coding tasks. This analysis, from a developer of large language models, suggests GLM 5.2 is competitive for coding, though it lacks native image input and built-in tools like Playwright found in CC.

OpenAI's ChatGPT is being promoted for ambitious work projects. The article highlights its potential applications, while the Hacker News discussion shows 27 comments and 80 points, indicating significant public interest and engagement with the topic. The community is actively discussing the implications and uses of ChatGPT in professional settings.

OpenAI has retracted its recommendation for SWE-Bench Pro, a coding evaluation benchmark. This decision follows concerns about the benchmark's reliability and its ability to accurately assess coding capabilities. The company is now advising against its use, suggesting that it may not effectively differentiate between signal and noise in evaluating coding performance. The announcement has generated discussion online, with 56 points and 20 comments on Hacker News.

Apple's lawsuit against OpenAI could significantly delay or even end OpenAI's hardware ambitions. This legal challenge, described as another "controversy and messy divorce" for the startup, specifically threatens the future of the ChatGPT device. The situation highlights the potential consequences of challenging established tech giants, echoing the sentiment of "don't poke the bear."

Dr. Ben Bernanke, former Federal Reserve Chair and Nobel laureate, has joined Anthropic's Long-Term Benefit Trust. This independent body ensures Anthropic responsibly develops AI for humanity's long-term benefit. Bernanke's expertise in economics and navigating financial crises will help the Trust understand AI's impact on economies and workforces, advising Anthropic on critical decisions and potential risks. He joins other trustees with diverse backgrounds.

BuzzRadr users are actively discussing GPT-5.6, a new model from OpenAI. The conversation centers around its deployment safety, as detailed in a provided PDF, and its technical specifications, available through the OpenAI API documentation. With 317 points and 196 comments, the community is deeply engaged in analyzing the implications and capabilities of this latest iteration.

OpenAI is expanding its focus beyond individual users to families, hiring a product manager to build experiences for households. This shift reflects a broadening user base, with more older adults and parents adopting ChatGPT. Experts suggest this move signals AI's integration into daily family life, necessitating new trust and safety measures, especially for younger users. The company has faced lawsuits regarding harm to children and is implementing safeguards, aiming to avoid past social media mistakes.