A research paper from UC Santa Cruz and accompanying writeup discussing how AI researchers found a way to run modern, billion-parameter-scale LLMs on just 13 watts of power. That's about the same as a 100W-equivalent LED bulb, but more importantly, its about 50 times more efficient than the 700W of power that's needed by data center
June 24, 2024 by Isidor Nikolic, @isidorn
In 2017, Google researchers introduced the concept of transformers in their seminal paper, "Attention is All You Need" . These types of models, which prioritize information similarly to human focus, are the foundation of today’s modern language models, such as the ones that power GitHub Copilot.
The same way that transformers were transformative (sorry) for AI, we believe GitHub Copilot Extensions will ultimately deliver on the full promise of AI
The AI boom has allowed the general consumer to use AI chatbots like ChatGPT to get information from prompts demonstrating both breadth and depth. However, these AI models are still prone to hallucinations , where erroneous answers are delivered. Moreover, AI models can even provide demonstrably false (sometimes dangerous) answers . While some hallucinations are caused by incorrect training data, generalization, or other data harvesting side-effects, Oxford researchers have target the problem from another angle. In Nature, they published details of
TL;DR: Anthropic has been on a roll of late, playing catch-up with OpenAI's offerings with new features. On Thursday, the company launched Claude 3.5 Sonnet, touting it as a leap over previous iterations while also claiming it beats the recently launched GPT-4 Omni on several metrics. The company also released Artifacts, a new workspace within Claude that displays things like code snippets as a
AMD is officially launching its latest ROCm (Radeon Open Compute) platform, now featuring support for multiple graphics cards in clusters.
