Just over a year ago, OpenAI announced the release of its API for access to the company’s GPT-3 language model for advanced natural language processing. Previously, access to this API required joining an extensive waitlist, which is now a thing of the past.
Despite the infinite compute power of the cloud and all the machine learning and artificial intelligence that thrives on that compute power, technology just falls short of our expectations. Which is exactly what happened when one developer gave Microsoft's Bing Speech Recognition API a whirl.
Facebook just announced the upcoming beta release of PyTorch 1.0. The AI framework fuses the research oriented capabilities of PyTorch with the production and scaling functionality of Caffe2 and ONNX. Facebook has already implemented PyTorch 1.0 to power many of its AI products.