OpenAI has announced a no-code ChatGPT Builder platform for businesses.

Announced at DevDay, OpenAI’s inaugural developer conference in San Francisco, the application can generate personalised versions of ChatGPT for certain use cases without requiring any coding. These AI agents, donned simply GPTs by OpenAI, will become available on the GPT Store over the coming weeks.

One potential use case is for businesses to develop internal-only GPTs for their workers.

An OpenAI blog wrote:

We’re rolling out custom versions of ChatGPT that you can create for a specific purpose—called GPTs. GPTs are a new way for anyone to create a tailored version of ChatGPT to be more helpful in their daily life, at specific tasks, at work, or home—and then share that creation with others.”

OpenAI suggests that the ChatGPT Store will eventually involve monetisation functionality, paying creators for the use of their GPT depending on how frequently their GPTs are used — as well as unconfirmed (for now) secondary factors. Both the creator and OpenAI will get a cut of the revenue produced by the GPT’s deployment, while OpenAI is responsible for the online housing of the user’s GPT and the processes of collecting the money.

Other GPT features include the decision of whether to provide each bot with access to OpenAI’s Code Interpreter tool, web browsing, and DALL-E.

OpenAI presented a case study at DevDay that entailed a GPT being created to assist attendees during the conference. The GPT platform automatically suggested a name for the bot, “Event Navigator”, and created a draft profile picture for it using OpenAI’s AI-powered image generation service DALL-E.

Event Navigator then parsed a PDF that included the conference’s schedule, which formed the basis of its answers. This is because the builder interface includes a “Knowledge” section for uploading personal data, such as that conference schedule, that the bot will then consider in its interactions with people. “Actions” enables GPTs to plug into third-party services to access databases and emails, with Canva and Zapier among its initial lists.

OpenAI’s interface enables users to design how their GPT interacts with other people before it’s finished. It can also be designed with autogenerated starter prompts. The Event Navigator agent was designed to be helpful and succinct while preventing scheduling conflicts.

“It’s very silly, but that thing I built on stage is something I’ve been waiting like a decade to build,” said OpenAI CEO Sam Altman.

GPTs will be exclusive for ChatGPT Plus and OpenAI enterprise subscribers. After the GPT Store launches, only agents from verified users will be approved by OpenAI. To begin with, GPTs can be accessed via shareable web links.

The Turbocharged GPT-4

OpenAI stated that the “Turbo” version of GPT-4, currently in preview, will be able to parse novel-length prompts and provide insightful answers and is also built online data running until April 2023.

The original version of GPT-4 could only parse around 50 pages of text and was also limited to September 2021, albeit OpenAI released a feature that allowed ChatGPT users to leverage the internet for current data. The Turbo GPT-4 will be better informed about current events.

Turbo GPT-4 will be three times cheaper for developers to use than previous iterations, and OpenAI affirms that it will emulate the likes of Microsoft and Google in protecting customers against copyright lawsuits.

At DevDay, OpenAI also announced it had reached an astonishing 100 million weekly users, that over 90 percent of Fortune 500 businesses are building tools on ChatGPT and that it was introducing lower prices for developers leveraging its models in their applications.



from UC Today https://ift.tt/51ISQA7