OpenAI Unveils Custom Chip for Next-Level AI Processing
In brief
- OpenAI has developed a new custom chip called "Jalapeño" in collaboration with Broadcom.
- This specialized hardware is designed to enhance the performance of large language models, which are used in tasks like chatbots and text generation.
- The chip aims to improve efficiency and speed, making it particularly suited for handling complex AI computations at scale.
- This development could be a game-changer for AI researchers and developers, as it provides a more powerful tool for training and deploying advanced models.
- By tailoring hardware specifically for large language models, OpenAI is addressing one of the key bottlenecks in AI processing-computational power.
- The chip is expected to become available by late 2026, which could significantly accelerate progress in AI capabilities across various industries.
- As AI technology continues to evolve, the integration of specialized hardware like Jalapeño will likely become more common.
- OpenAI’s move signals a shift toward optimizing AI systems not just through software advancements but also through purpose-built hardware.
- The future of AI processing may lie in such custom solutions, promising faster and more efficient computations for next-generation applications.
Terms in this brief
- Jalapeño
- A custom chip developed by OpenAI and Broadcom designed to enhance the performance of large language models. It aims to improve efficiency and speed for complex AI computations, particularly useful for training and deploying advanced models.
Read full story at The Decoder →, Digg AI →
More briefs
Google Pulls AI Feature from Google Earth
Google pulled its new AI feature from Google Earth less than a day after launch. The feature allowed users to generate AI images and overlay them onto real satellite and aerial imagery. Google had envisioned uses such as visualizing planning concepts or historical reconstructions, but users began generating fabricated disasters and false scenes. This matters because over 1 billion people use Google Earth. The feature was meant to be a tool for legitimate uses but was quickly abused. The company will likely rework the feature to prevent such abuses in the future.
Meta Enters Cloud Business
Meta is forming a cloud unit to sell its excess AI computing power to enterprise clients. This move puts Meta in competition with established cloud providers like AWS, Azure, and Google Cloud. The demand for AI compute resources is growing, with Meta investing $145 billion in infrastructure, and this new business could help generate returns beyond advertising. Meta will sell access to its compute infrastructure, and the company is weighing whether to lead with raw computing capacity or AI models. Next, Meta will start selling its cloud services to enterprise customers.
Google AI Spreads False Claim About Flock Cameras
Google's AI claimed Flock cameras hold $650 in gold and 23 pounds of copper. This claim came from a joke that was circulating online. The claim mattered because it was repeated as fact by Google's AI. The AI said each camera holds about 1 to 5 grams of gold and 23 pounds of copper. The Flock camera actually weighs about three pounds. Google's AI has now corrected the claim. Google's AI will likely face more tests of its fact-checking abilities.
AI-Generated Film Premiere
Higgsfield presented the world's first fully AI-generated feature film. The 110-minute film stars licensed likenesses of real celebrities. The film's success matters because it shows AI can make high-quality movies at a lower cost. This can help creatives around the world make their visions a reality. The film features four real entertainers and brings to life a story of struggling rappers in East London. The film's premiere marks a new era in entertainment. The future of filmmaking will likely include more AI-generated movies.
Rippling Unveils AI Spend Console
Rippling has created a tool to track and control AI spending. The tool maps how much employees and teams spend on AI. The tool was made after Rippling found it was spending too much on AI tokens. It was on track to spend 40% of its R&D budget on tokens. Spending was growing by 80% each month. The company discovered that 10-15% of employees were using 60% of the total AI spend. One engineer was spending $50,000 a month. Rippling will now help other companies control their AI spending.