Intel Patents a Chip That Shuts Off Its Own Processing Units During AI Tasks
Running an AI model on-device is a power hog, and most chips don't adjust mid-task. Intel's new patent describes a processor that…
The unglamorous machinery underneath: training methods, inference speedups, model compression, and the scheduling tricks that make AI affordable to run.
Running an AI model on-device is a power hog, and most chips don't adjust mid-task. Intel's new patent describes a processor that…
AI models go stale the moment the world changes around them. IBM's new patent describes a system that watches for those changes…
Running a powerful AI model on a phone or laptop usually means accepting a watered-down version. Microsoft's new patent tries to change…
Most databases speak a language only developers know. IBM's new patent describes a way to automatically build the training material an AI…
Getting a general-purpose AI to give reliable answers in a specialized field, like finance, medicine, or law, is harder than it sounds.…
Training a large AI model can take days across hundreds of servers, and a single server crash can throw away all that…
Training a large AI model can take days or weeks, so a server failure halfway through is genuinely painful. IBM has filed…
AI models are packed with enormous grids of numbers, and storing them is expensive. Microsoft has filed a patent for a compression…
Compressing an AI model to run faster hits a wall whenever the model needs to add or subtract numbers, and Nvidia's new…
Getting an AI to answer questions correctly is one thing. Getting it to answer them completely, in natural language, without missing key…
AI is already being used to help analyze legal cases, but if the data those models trained on is skewed, so are…
Training an AI model across thousands of devices without ever pooling their private data is one of the harder problems in machine…
We read every patent application Big Tech publishes and send you the ones worth knowing. Plain English, free, every week.