AI News Archive - Browse Page 130 of 282
Browse AI news articles covering LLMs, tools, research, and industry trends
AVLLMs Mirror VLM and VideoLLM Sequential Flow in Audio‑Visual Tasks
Models that process both sound and sight are often treated like alien minds. New research reveals a much more boring, and more useful, truth.
vLLM uses custom GPU kernels, TorchInductor and CUTLASS for portable inference
Portable inference across diverse hardware is a brutal optimization problem. vLLM attacks it with a triple threat: custom GPU kernels for raw...
Claude Fable declines basic biology queries; Opus 4.8 responds
Anthropic confirmed it. Their new Claude Fable 5 model is refusing to answer even elementary biology questions.
Tech oligarchs face loyalty test in Trump‑era Washington over past Democrat ties
For tech billionaires, backing Democrats used to be smart business. Now it's political kryptonite in Trump's Washington.
Run DiffusionGemma on NVIDIA GPUs for high‑throughput text generation
The line between experimentation and production has never been thinner. DiffusionGemma is changing what’s possible for high‑throughput text...
SynIB Introduces Information Bottleneck to Boost Multimodal Synergy
Most AI models are lazy. They take the easy way out. Confronted with data from different sources, like an image paired with text, they'll grab the...
Datadog engineers start AI coding firm Niteshift, backed by Hoffman, Pomel
Handing your proprietary code to OpenAI carries a unique risk. It's not just about a bug.
Model 5 tops penalized PR-AUC, recall and F1-score in scoring model training
When you train a credit scoring model, you need a tool that ranks risk across a portfolio, not just flags defaults at a single point.
Matmul Enables Dropless MoE Training; Grouped‑GEMM Kernel Drives Speed
Dropless Mixture-of-Experts training has long been haunted by a simple truth: you cannot skip the tokens, but you can pay for them in compute.
Decart’s world model simulates hours of photorealistic driving
Decart is selling a machine that hallucinates highways. The two-year-old company just banked $300 million in new funding.
Armstrong predicts 20% of AI workloads will stay on latest‑gen models
The demand for intelligence is effectively infinite, but the cost of delivering it is not.
NVIDIA Nsight Designer Streams ONNX Editing and TensorRT Engine Build
The ONNX graph is a labyrinth of nodes, each one a decision point. But when you’re chasing FP8 performance, the path isn’t just about layout, it’s...
AI moves beyond automation to plan, optimize and execute business initiatives
Forget automation. The real sales pitch has pivoted to strategy. Software vendors now hawk systems that don't just complete tasks—they decide which...
Understanding AgentOps: Discipline and the agentops.ai Platform Explained
Standard logging gives you a timestamp and a line of text. It tells you step 7 failed, and leaves you guessing why.
NVIDIA FLARE Auto-FL Enables Agent-Led Coding in Controlled Experiments
The most tedious part of federated learning research isn’t the thinking, it’s the iterating.
Multiverse reduces inference cost by favoring low‑cost prefill over decoding
Most AI cost analysis misses the point. The expensive part isn't starting a conversation with the model. It's letting it finish.
Grab, CJ ENM, LiveKit praise Gemini 3.5 Live Translate for quality and accuracy
Real-time translation software has always sucked. Users gritted their teeth through a slow, robotic mess or paid out the nose for human interpreters.
Apple's top AI concept mirrors vibe coding, using Shortcuts as a model
Google and Microsoft slap "AI" on every feature launch. Apple's Shortcuts app has simply done the work for years.
NVIDIA Nemotron Speech and Agent Skills Speed Clinical ASR Evaluation
In clinical settings, accuracy in speech recognition isn’t just a metric, it’s a matter of patient safety.
AI‑enhanced lessons in Sierra Leone: teachers lead impact study
Chalk dust and a Google AI model mixed in thirty Sierra Leonean schools. Test scores rose. But the real story was who wielded the chalk: the teachers.