OpenAI is holding a conference starting today! The conference signals OpenAI’s ambition to expand beyond a consumer sensation to becoming a provider of a hit developer platform, and CEO Sam Altman has teased attendees with the promise of “some great new stuff.”

The one-day event, in a desolate area of San Francisco near City Hall, is attracting hundreds of developers from around the world who are eager to learn more about the latest advances in generative AI.

The burgeoning AI sector has been a bright spot for San Francisco’s economy, which has struggled to bounce back from the pandemic that hit the city hard.

After toiling in relative obscurity for years, OpenAI kicked off the generative AI craze last November by releasing a chatbot that became one of the world’s fastest growing consumer applications. The chatbot, which uses a powerful AI model called GPT-3, can generate realistic and engaging conversations on any topic, based on past data.

Generative AI can, using past data, create brand new content like fully formed text, images and software code. This opens up new possibilities for creativity and innovation, as well as challenges for ethics and regulation.

Backed by billions of dollars from Microsoft, OpenAI has become for many the default version of generative AI, helping users spin up term papers, contracts, travel itineraries and even entire novels. The chatbot has also spawned a number of viral social media accounts, such as a parody of Elon Musk and a fake Donald Trump.

OpenAI is expected to announce updates focused on slashing costs for its developers, as well as new vision capabilities, Reuters previously reported.

Cost-cutting addresses a major concern for partners whose spending on OpenAI’s powerful models could pile up quickly as they try to build sustainable businesses by developing and selling AI software. OpenAI charges developers based on the number and complexity of requests they make to its models, which can quickly add up for large-scale applications.

The vision capabilities, which will enable OpenAI’s software to analyse images and describe them, will enable developers to build applications with new uses in fields from entertainment to medicine. For example, developers could create apps that can generate captions for photos, identify objects and faces, or create realistic images from text descriptions.

Another announcement could be an ability to fine-tune GPT-4, its most advanced AI model, which the company previously said would be coming in the fall. Fine-tuning allows developers to customize the model to their specific needs and data, improving its performance and accuracy. GPT-4 is expected to be even more powerful and versatile than GPT-3, which already has 175 billion parameters and can generate text in 45 languages.