Tag
18 articles
Learn how Sakana AI's new tool, Sakana Marlin, uses artificial intelligence to automate weeks of business research in just hours. This explainer breaks down how AI research assistants work and why they matter for decision-making.
Salesforce AI Research introduces VoiceAgentRAG, a dual-agent memory router that cuts voice RAG retrieval latency by 316 times, enhancing conversational AI responsiveness.
Researchers from Meta, Cornell, and CMU introduce TinyLoRA, a 13-parameter fine-tuning method that achieves 91.8% accuracy on GSM8K using Qwen2.5-7B.
Microsoft is hiring leading AI researchers from the Allen Institute for AI and the University of Washington to strengthen its Superintelligence team led by Mustafa Suleyman.
Yann LeCun introduces LeWorldModel (LeWM), a new framework targeting JEPA collapse in pixel-based predictive world modeling. This advancement could significantly improve how AI systems learn meaningful representations from raw visual data.
Learn to build a basic AI research agent that generates, verifies, and revises mathematical proofs using natural language processing techniques.
Thinking Machines Lab has secured a multi-year deal with Nvidia worth at least one gigawatt of compute power, along with a strategic investment that positions the company at the forefront of AI research.
Learn how to use the open-source CiteAudit tool to detect hallucinated references in scientific papers, a growing problem in AI research.
This article explains Yann LeCun's critique of the term Artificial General Intelligence (AGI) and introduces his proposed framework of Superhuman Adaptable Intelligence (SAI), which focuses on measurable performance and adaptability in AI systems.
Balyasny Asset Management has developed an AI research engine using GPT-5.4 and rigorous model evaluation to transform investment analysis at scale. The system combines advanced language models with agent workflows to automate research processes and enhance decision-making efficiency.
Anthropic's new study reveals that while AI poses theoretical risks to certain jobs, real-world unemployment has not yet risen significantly, with young workers showing early warning signs.
OpenAI's GPT-5.2 Pro has helped extend single-minus amplitudes to gravitons, advancing quantum gravity research through AI-assisted mathematical derivation and verification.