OpenAI Unveils Developer Tools to Boost Generative AI Model Creation

OpenAI released new developer tools to enhance the use of its AI technology in model creation. This gives it an edge in the race for generative AI.

On Tuesday, OpenAI released many new tools that will make it easier for developers to make models based on its AI technology. The company that makes ChatGPT is competing with big tech companies to stay ahead in the generative AI race.

The Microsoft-backed startup said that developers would only need to follow one set of instructions to make AI voice apps using a new real-time tool that is now available for testing. Before, developers had to do at least three steps: first, they had to transcribe the audio, then run the generated-text model to get an answer, and finally, they had to use a separate text-to-speech model.

Businesses that use OpenAI’s services to make their own AI apps bring in a lot of money, so adding new features is a big part of the business’s strategy. Technology giants, like Alphabet, which owns Google, are also stepping up the competition by using AI models that can process different types of data across their businesses. These models can handle video, audio, and text.

Reuters reported last month that OpenAI thinks its sales will go from an estimated $3.7 billion in 2024 to $11.6 billion next year. The business is also raising $6.5 billion in funding, which could put its value at $150 billion. As part of Tuesday’s rollout, OpenAI added a fine-tuning tool for models that developers could use to make the responses their models gave using text and images better.

Enhancing AI with User Feedback

As part of this fine-tuning process, people can provide feedback on the model’s answers and show it examples of good and bad answers. OpenAI said that using images to fine-tune models would make them better at understanding pictures, which would lead to uses like better visual search and better object detection for self-driving cars.

The startup also showed off a tool that lets smaller models learn from bigger, more experienced ones. It also showed off “Prompt Caching,” a feature that reuses parts of text that AI has already processed, which cuts some development costs by half.