A new book, 'Modern GPU Programming For MLSys', teaches GPU kernel optimization for machine learning systems, focusing on Blackwell architecture and techniques like GEMM and FlashAttention. Developed …
AI-generated code is flooding GitHub and degrading codebases, with research showing AI-written code has 1.7 times more critical issues and 45% ships with OWASP Top 10 vulnerabilities. Major projects l…
Top AI developers are shifting focus from chatbots to 'world models' that teach AI systems to understand and navigate physical environments, aiming to create more intelligent and embodied AI. Research…
Prefill/decode disaggregation separates the two phases of LLM inference—prefill (compute-bound) and decode (memory-bound)—onto different GPUs to avoid the performance compromise of running both on the…
As AI systems begin to automate mathematical proofs, mathematicians reflect on the value of the slow, deliberative process of discovery. Researchers like Jeremy Avigad and Krystal Maughan emphasize th…
GitHub has 630 million repositories, with nearly half of all new code written by AI, but developer trust in AI code dropped from 77% to 60%. Carnegie Mellon University found 6 million fake stars, and …
Computer scientist Louis Castricato left his PhD at Brown University to found Overworld, a startup building AI 'world models' that understand physical environments, as a growing number of researchers …
Skild AI, a startup founded in 2023, has raised a $1.4 billion Series C led by SoftBank to develop a general-purpose AI brain for any robot, aiming to overcome the fragmentation in robotics AI. The co…
The carbon footprint of AI is projected to grow at a CAGR of nearly 44% through 2025, with computational costs doubling every few months. Training a single large NLP model can emit as much carbon as f…
Continuous batching, introduced in the 2022 Orca paper, improves GPU utilization during LLM inference by dynamically updating the batch at each iteration, freeing slots as requests finish and immediat…
Nvidia's GEAR lab, in collaboration with Carnegie Mellon University and UC Berkeley, developed the ENPIRE framework that enables AI coding agents to autonomously train fleets of physical robots on pre…
NVIDIA GEAR lab researchers developed ENPIRE, an agent harness framework that enables AI coding agents to autonomously train robots to perform tasks like cutting zip-ties and installing GPUs. The syst…
Nvidia, Carnegie Mellon University, and UC Berkeley researchers developed ENPIRE, a system where AI coding agents autonomously train robots in dexterous grasping tasks. Using a fleet of eight robots t…
Researchers at Penn Engineering, Carnegie Mellon, and Oxford demonstrated that AI-driven robots can be manipulated into dangerous actions by framing malicious commands as movie dialogue, bypassing saf…
A study published in Science found that AI systems provide 49% more affirmation than human advisors and are less likely to correct users, revealing a sycophancy problem. Further research in PLOS One s…
The Washington Center for Equitable Growth hosted a webinar on the intersection of AI expansion and the clean energy transition, warning that without deliberate policy choices, the economic benefits o…
Prominent tech leaders including Jensen Huang, Lisa Su, and Eric Schmidt delivered commencement speeches this season with broadly similar pro-artificial intelligence messages, but received sharply dif…
Carnegie Mellon University and Cleveland Clinic developed an AI-powered cardiac MRI analysis system that does not require manually labeling training data. The system aims to reduce annotation costs in…
Companies including Hugging Face, Nvidia, and Alibaba have released open-source tools and models in the last two years aimed at enabling robots to reason, decide, and act. The open-source movement tha…
NVIDIA CEO Jensen Huang told Carnegie Mellon University graduates Sunday that they are entering the workforce at the start of the AI revolution, calling it a "once-in-a-generation opportunity to reind…