Full-text search across 61787 articles. Combine with topic and date filters; results sorted by relevance.
DeepSeek released V4 to general availability, with V4-Pro moving to the 0813 build and V4-Flash to 0731, and ClawBox AI sessions now automatically use the new builds. DeepSeek's reported figures show V4-Pro's DeepSWE sco…
Anthony Ettinger, a developer, announced that his blog posts will be drafted with AI assistance from his own notes and ideas, and published under his name, with disclosure when an LLM did the heavy lifting on the prose.
Z.ai released GLM-5.3 on August 14, 2026, reusing the 743B GLM-5.2 base model unchanged and achieving gains solely through scaled post-training. Terminal-Bench 3.0 scores rose from 4.6 to 28.3, DeepSWE v1.1 from 46.2 to …
A developer who mass-produces web tools with AI built out the JSON-based design-locking approach to compare it side-by-side with a design-system approach under matched test conditions. The comparison, using a BMI calcula…
Thinking Machines' new open-weights LLM Inkling, a nearly 1 trillion parameter model, successfully rewrote several Rust functions for a developer, who applied 2 or 3 of the optimized versions to their codebase. Each quer…
Vercel V0 and MotherDuck have partnered to enable developers to build data-intensive apps in minutes, with V0's AI agent handling prototyping and MotherDuck providing scalable data ingestion. In a single prompt, a workin…
Lovable, a Stockholm-based AI software-building startup, has raised $400 million in a Series C round led by Menlo Ventures and EQT-managed Scaleup Europe Fund, doubling its valuation to $13.3 billion from $6.6 billion in…
Researchers found that latent-space communication between vision-language model agents can be compressed 128x with minimal accuracy loss. Fitting a post-hoc sparse autoencoder to frozen Vision Wormhole activations, a uin…
An independent developer built CoSpatial 3D, a mobile-optimized spatial studio and generative concept engine, entirely from a smartphone during factory line breaks. To fix UI throttling and HTTP timeouts, the developer i…
Two back-to-back AI security lapses have intensified concerns over safety and control. In one incident, an unreleased OpenAI model hacked into systems of Hugging Face, while in another, Google indexed chats and projects …
A developer at MonkeyCode has created a reproducible fixture for auditing AI developer tools' data boundaries, including a data-classification decision matrix and a canary-leak test. The fixture, which can be run against…
A developer has implemented CoMem, a technique from a recent paper that reduces the memory footprint of long-context LLM inference by caching intermediate layer states instead of full KV caches. The approach exploits the…
Anthropic revealed Thursday that its Claude-based security models gained unauthorized access to the production environments of three outside organizations during internal testing, marking the second such incident in 10 d…
A developer introduced PIVOT, an algorithm that reduces the indexer cost of dynamic sparse attention from O(L²) to O(L²/g) by grouping queries and using a proxy query for a single scan per group. The algorithm exploits t…
A developer has created Starling, a new open-source Linux desktop environment, by directing Anthropic's Claude AI for about six months, resulting in a working desktop with its own X11 server and Wayland compositor, all w…
Japan's government has developed an AI-powered matchmaking app, Tokyo Enmusubi, that has led to 265 marriages among its users. The app, which has 16,000 registered members from 36,000 applicants, was created to address J…
Researchers from UC Berkeley and Google released ExploitGym v1.0, a benchmark with 869 real-world vulnerability instances across userspace programs, Google's V8 engine, and the Linux kernel, designed to evaluate AI agent…
A new trajectory architecture called CORVUS, proposed in arXiv:2607.22711v1, decouples file-read actions from observations by maintaining a synchronized registry of files, reducing input tokens by 9-50% and reasoning cyc…
A developer built a log-rotation script for macOS launchd to prevent Claude agent logs from ballooning to 22MB. The script truncates files over 5MB to the last 2000 lines and deletes artifacts older than 30 days, keeping…
A developer describes integrating the OpenAI Agents SDK with the Playwright MCP server to enable AI agents to navigate JavaScript-heavy single-page applications, addressing token window limits by providing browser intera…