SpaceX pursues access to US cell sites for wireless network, report says
Discussions cover at least 80,000 sites and include Crown Castle, potentially advancing plans for a fourth nationwide wireless carrier.
OpenAI broadens Decisions API access as decision-model rivals expand
The public version adds voice and visual capabilities to a product based on a reconfigured ChatGPT Luna model.
OpenAI has made its Decisions API available to all customers, broadening its challenge to TypeSafe’s Jev in software that quickly chooses between options. The wider rollout includes new voice and visual capabilities, following the product’s announcement on September 29.
Decision models return a choice along with a confidence rating. Developers have adopted them as alternatives to large language models in some tasks because they can process requests faster and at lower cost.
For its offering, OpenAI modified ChatGPT Luna rather than building a model specifically for decision-making, said Nikunj Handa, the company’s API product lead. He said the changes shortened inference times and reduced costs. The company also placed inference resources across regions to limit delays.
Speak, Ramp, Box and Datadog were among the initial customers testing the API. Engineers used it to substitute for OpenAI’s large language models and increase throughput. OpenAI also deployed it internally to improve responsiveness in its Dots voice assistant.
Competing products include Liquid’s D1 Ultra and Base Lab’s Elephant-1. TypeSafe’s newest Jev version supports images and smart-if statements. Handa said OpenAI customers also want more flexibility in using images as outputs from the Decisions API.