AI Startup Runway Expands Beyond Video Generation

Google Gemini Leads in AI Model Usage

Google's Gemini model is in high demand amongst AI customers. It led in the number of tokens handled by Vercel's AI Gateway in April. Google's Gemini model is popular due to its speed and affordability. In April, Google led with 21% of dollars spent on AI models, while Anthropic led with 61%. OpenAI's spend share tripled from March to April. The AI market is constantly changing, with new models and features being released. Google will likely unveil new AI models and tools at its upcoming conference.

AI Streamlines Employee Performance Reviews

Companies are using AI to write employee performance reviews. AI systems can pull data from across an organization to draft evaluations. AI can cut review-writing time by 40%. This matters because it can save companies time and money. For example, Boston Consulting Group uses an AI assistant to speed up the review process. Companies will continue to use AI to improve performance reviews.

x.AI Unveils Terminal-Based Coding Tool Grok Build

Elon Musk's AI company, x.AI, has launched Grok Build, a terminal-based coding tool designed to assist developers. This tool marks x.AI's first entry into the coding agent space and aims to compete with existing tools like GitHub Copilot. By integrating AI directly into the command line interface, Grok Build offers suggestions and automates repetitive tasks, potentially saving developers time and effort. The move is significant as it brings advanced AI capabilities to a traditionally manual process, enhancing efficiency for software development. While specific details on its performance and integration with other platforms are limited, Grok Build represents x.AI's strategic push into practical AI applications. Developers can expect more tools from x.AI that focus on streamlining coding workflows in the future.

AI Breakthrough Makes LLMs Faster Without Losing Accuracy

AI researchers have discovered a new way to make large language models (LLMs) run much faster without losing their accuracy. This advancement, unveiled in a recent study, shows that by adjusting the model's internal architecture-specifically the balance between attention layers and MLP layers-they can boost processing speed by up to 47% while keeping performance intact. The key insight is identifying an optimal ratio of MLP-to-attention parameters around 1.0 for LLaMA-style models. This is a significant improvement over existing open-source versions, which often have much higher ratios like 4.8. The researchers tested this approach across different GPU architectures and found consistent gains in efficiency, making it easier to deploy high-performing AI systems in real-time applications. This development could lower the cost of running LLMs while maintaining their accuracy, opening up new possibilities for businesses and developers. Look out for further studies on how these scaling laws can be applied to other types of AI models in the coming months.

OpenAI Brings AI Coding Assistant Codex to iOS and Android

OpenAI has integrated its powerful AI coding assistant, Codex, directly into ChatGPT for mobile users. This means anyone with the app can now write code in multiple programming languages by simply describing what they want. Whether it's debugging or building new features, developers can get real-time assistance right on their phones. The move marks a significant expansion of OpenAI's tools, making advanced coding support more accessible than ever. Developers who test this feature have praised its speed and accuracy, especially for complex tasks like code generation and debugging. This integration could save professionals hours of work and streamline the development process. For users already familiar with ChatGPT, accessing Codex is seamless-it works right within the app's interface. OpenAI plans to roll out updates with more features and improved performance in the coming months, making it an even stronger tool for developers worldwide.