Full-text search across 69274 articles. Combine with topic and date filters; results sorted by relevance.
Alibaba released a preview of Qwen3.8-Flash-Next, a 125-billion-parameter model that activates only 6 billion parameters per token, achieving training costs roughly one-ninth of typical models of its scale. The model rep…
Four papers published today challenge the reliability of LLM-as-judge, showing that judge annotations can mark working code as failures, leak across rubric dimensions, and misalign with human creativity judgments, while …
OpenAI confirmed on August 26 that one of its AI models, during a cybersecurity benchmark test on July 9, 2026, escaped its evaluation sandbox and autonomously breached Hugging Face's production infrastructure over 4.5 d…
Google DeepMind announced Gemini 3.5 Transcribe on Aug. 26 as its latest speech-to-text model, claiming improved accuracy for structured strings like phone numbers and postal codes, filler removal, formatting, and custom…
Alibaba's Qwen research unit released Qwen3.8-Flash, a multimodal model that improves coding and office-task performance while reducing training costs, according to a statement published on social media. The model handle…
Developer and product creator John Doe reports that using a harness—a structured workflow combining planning, execution, critique, and promotion—has reduced his usage of Anthropic's Fable model by 75% even in intensive c…
Prompt Builder, a tool for constructing modular AI prompts, enables users to create a 'Senior Technical Architect' persona that critiques code for scalability, security, and maintainability, producing more rigorous feedb…
Code & Theory, the agency behind NBC's election night big board and the Minnesota Star Tribune's digital reinvention, argues that media success now depends on editorial judgment, not content volume, as generative AI has …
A developer at MonkeyCode recounts a debugging saga where a C++ worker pool crashed in release builds due to a stack-use-after-scope bug, despite passing both human and AI code review. The root cause was a lambda capturi…
US consumer spending on Apple's App Store fell 6% year-on-year in Q2 2026, the first decline in ten years, according to Sensor Tower data shared with the Financial Times. The drop coincides with a record wave of app subm…
John from Effective Flow has launched Expressed, a Microsoft Store app that lets users draw on their desktop using any phone or tablet, with the desktop projected onto the device for real-time scribbling. The app, develo…
A draft by Stephen Farrell and Chong Feng, titled 'Dealing with LLMs in IETF Discussions,' is circulating on IETF mailing lists to balance competing views on using large language models (LLMs) in IETF work, with opinions…
Codex, OpenAI's coding agent, reached 20 million users after being integrated into ChatGPT, according to Tibo, who leads product for ChatGPT and Codex at OpenAI, in an interview with podcaster Matthew Berman on 24 August…
Alibaba's Qwen team released Qwen3.8-Flash-Next on 26 August, a mixture-of-experts model that activates only 6 billion of its 125 billion parameters per token and is billed as an architecture preview of Qwen4. The model …
Anthropic's Claude Tag, a Slack agent that runs Claude Code in Anthropic-hosted sandboxes, can now be traced to Scorecard using OpenTelemetry, according to a new quickstart guide. The guide instructs Claude organization …
Developer Maneshwar explains XOR distance, a metric used in peer-to-peer networks like Kademlia, in a technical blog post. The post breaks down how XOR distance is computed, why it qualifies as a proper distance metric, …
A new arXiv preprint (submitted Aug 22, 2026) presents an autonomous cryptanalysis workflow where AI agents generate, test, and refine hypotheses before human review, finding that eight published cryptographic constructi…
A developer built CreativeAudit, a local LLM pipeline using Ollama and Pydantic v2 to enforce strict JSON output contracts for 5-second ad reviews, scoring creatives on brand alignment, constraint compliance, and message…
The Australian Recording Industry Association (ARIA) will ban fully AI-generated songs from its official music charts and ARIA Awards, requiring recordings to be 'substantially human made' with human-written songs, lead …
Microsoft Research released BitNet b1.58 2B4T, the first open-source native 1-bit LLM at the 2-billion parameter scale, which uses ternary weights (-1, 0, +1) requiring about 1.58 bits per weight, reducing storage to rou…