Full-text search across 70246 articles. Combine with topic and date filters; results sorted by relevance.
Anthropic's Claude Fable 5 scored 161 on the Epoch Capabilities Index, surpassing OpenAI's GPT-5.5 Pro at 159, with dominant performances in math and software engineering benchmarks. The model achieved 88% on FrontierMat…
Apple introduced Xcode 27 at WWDC 2026, featuring extended coding agent integration, a customizable toolbar and themes, a redesigned project creation experience, and DeviceHub for unified simulator and device management.…
Anthropic detailed how its Claude Code feature dynamically generates JavaScript harnesses to coordinate teams of AI agents for complex tasks, addressing issues like agentic laziness and goal drift. The system uses strate…
A developer found that GPT-5.4 produces 6.5x more output tokens per code fix than Claude Opus 4.6, with no improvement in correctness. The over-editing tax costs an extra $1,650 per month for a 50-engineer team making 40…
Meta is testing facial recognition software from Rank One Computing, a firm that sells surveillance tech to U.S. government and police, for its smart glasses. License documents show Meta is authorized to use Rank One's f…
Developers often tell small, strategic lies to keep work flowing smoothly, such as overconfidence in estimates or downplaying uncertainty, according to interviews with engineers. The pressure to sound confident and meet …
Anthropic researchers introduced Natural Language Autoencoders (NLAs), an unsupervised method that generates natural language explanations of LLM activations by training two modules to reconstruct activations through a t…
Silicon UK's AI For Your Business podcast explores autonomous money and AI spending in an episode titled 'Autonomous Money: Are We Ready to Let AI Spend for Us?', examining accountability, audit trails, and governance ch…
A user on Hacker News reports that Anthropic's Claude AI assistant frequently flags legitimate biology research questions, making it nearly unusable for immunology queries. The user notes that the same guardrails can be …
A developer warns that AI-generated code often creates naming inconsistencies because the AI invents new conventions each session. The solution is to define explicit naming rules for handlers, hooks, and props, which the…
Check is a new command-line tool that lets users prefix any command with 'check' to deterministically verify whether it will run successfully before execution, returning an exit code 0 or 1. The tool costs $0.04 per chec…
Three vulnerabilities in LangGraph's checkpointer, disclosed on June 12, allow attackers to chain SQL injection into remote code execution on self-hosted servers using SQLite or Redis backends. The flaws, discovered by C…
Shares of Zhipu AI surged up to 48% on Monday after the Chinese firm announced it would open-source its latest large language model, GLM-5.2, later this week. The move aims to attract users seeking cheaper alternatives t…
A developer built five Claude automation agents that save hours weekly. The agents handle tasks like clipboard-based knowledge storage, email action item extraction, error diagnosis, meeting note summarization, and pre-P…
Researchers applied a coding agent to formalize a numerical analysis textbook in Lean 4, introducing a three-dimensional quality audit framework beyond kernel acceptance. The audit revealed unfaithful formalization patte…
Researchers at MIT and other institutions released BEAVER, a large-scale enterprise benchmark for evaluating LLM text-to-SQL capabilities, containing 9,128 queries from private data warehouses across 19 domains. The benc…
Remote-first company Entire held its team summit in Tokyo with 32 employees to align on strategy. The company shipped fixes across Entire CLI and Entire Web, including new experimental commands 'entire blame' and 'entire…
Developers can enforce coding conventions and prevent unwanted commits by using git pre-commit hooks, which run automated checks before allowing a commit. This approach reduces the need to embed strict rules in agent ins…
Researchers Zhao et al. published a paper on arXiv 2025 introducing Absolute Zero, a method where a model proposes its own tasks and a code executor grades them, enabling reasoning reinforcement learning with zero human …
A developer outlines a workflow for designing a mobile app with AI, using tools like Mobbin for references and Daisy to generate cohesive screen sets from a single prompt. The process emphasizes validating demand, iterat…