Full-text search across 7151 articles. Combine with topic and date filters; results sorted by relevance.
Researchers have demonstrated that large language models can serve as selective surrogates for GPU kernel performance evaluation, accurately forecasting kernel runtime without requiring repeated compilation and execution…
White House Office of Science and Technology Policy Director Michael Kratsios accused Chinese AI startup Moonshot AI of using distillation techniques to copy Anthropic PBC's Fable model for its Kimi K3 open-weights model…
Researchers have developed a failure-driven self-improvement loop for AI training that leverages failed trajectories to boost performance, challenging the traditional focus on successes. The approach improved the OpenCUA…
Researchers have found that continuous reward models in reinforcement learning are oversensitive, assigning different scores to equally good responses, which can lead to bad policies. They propose evaluating reward model…
Punjab's plan to teach artificial intelligence to primary school students is stalled because thousands of government schools lack computers, internet, and trained teachers, according to the Punjab Education Curriculum, T…
Researchers propose MILP-Evo, a closed-loop framework that uses LLM-guided program evolution to automatically design mixed-integer linear programming (MILP) solver logic, such as cut selectors and branching rules, by gen…
A developer known as workofart has created a PyTorch-like framework for training AI agent harnesses, achieving a 45-minute experiment cycle on terminal bench tasks by enforcing deterministic inference via SGLang on NVIDI…
Dean Ball will join OpenAI on July 6 as Head of Strategic Futures, a new team focused on frontier AI policy including catastrophic risk, recursive self-improvement, labor market impact, and government relations. Ball wil…
NVIDIA announced that TSMC is integrating NVIDIA accelerated computing and AI across semiconductor design and manufacturing to improve turnaround time, energy efficiency, yield and operational productivity. TSMC is using…
A developer rescued a Retrieval-Augmented Generation (RAG) assistant from memory leaks and deployed it on a Render free-tier instance with a 512MB RAM limit. The engineer re-engineered the pipeline by replacing heavy PyT…
A developer published a plain-English guide to transformer architecture, explaining how large language models like GPT-4, Claude, Gemini, and Llama process text from tokenization to autoregressive generation. The guide c…
Researchers used deep learning to screen 19.3 million fragments from 2,897 prion-related proteins, identifying 1,179 candidate antimicrobial peptides called prionins. Of 75 synthesized prionins, 59 inhibited bacterial pa…
Millions of people are confiding in AI therapy chatbots, but a KFF Health News investigation reveals that mental health data is dangerously exposed. Researchers found 1,575 vulnerabilities in popular Android mental healt…
Researchers discovered that artificial neural networks like CLIP contain "multimodal neurons" that respond to abstract concepts (such as "Spider-Man") across different input modalities like images and text. The team deve…
Leanpub has launched the book 'Systems Thinking for Agentic AI: A Software Architect’s Guide to Building Reliable LLM and Agent Systems' by software architect Ediz Najim. The book teaches software engineers and architect…
Guardian, an open-source user-space firewall for AI agents, has released v0.1.0, intercepting and evaluating agent actions with a deterministic policy engine. In testing, it reduced prompt-injection attack success from 1…
Researchers developed a new method for out-of-scope intent detection using MiniLM embeddings, treating it as a one-class classification task with multi-cluster boundary learning. The approach outperformed existing baseli…
Shrijith Venkatramana, building git-lrc, explains how scaling laws discovered by OpenAI, Google, DeepMind, and Anthropic made large language models like ChatGPT, Claude, and Gemini possible. The 2020 OpenAI paper 'Scalin…
Tiling splits large images into smaller patches during computer vision inference, making small objects like cars and people in aerial imagery easier to detect by preserving detail that would otherwise be lost when the fu…
Large language models (LLMs) like GPT-3.5 lack a capability the author calls 'strong generalisation,' a mix of situational awareness, out-of-distribution generalisation, symbol grounding, and adaptive world modelling tha…