Former Facebook and Oculus employee Daniel Habib acted immediately when OPENAI, a San Francisco-based business creating artificial intelligence tools, announced the release of ChatGPT in November 2022.
Four days after ChatGPT’s debut, Habib leveraged the chatbot to create QuickVid AI, which largely automates the process of coming up with ideas for YouTube videos. Creators choose the subject of their video and the category they want it to go under, and QuickVid queries ChatGPT to generate a script. The script is then voiced by other generative AI tools, who also produce images.
Tens of thousands of people utilised it every day, but Habib had been utilising unauthorised access points to ChatGPT, which limited his ability to advertise the service and prevented him from charging for it in a formal manner. On March 1, OpenAI declared the availability of API access for ChatGPT and Whisper, a speech recognition AI product the company has created. Habib connected QuickVid to the official ChatGPT API in less than an hour.
He claims that all of the great unauthorised tools that were essentially simply toys for your own personal sandbox are now available to a large number of users.
The statement from OpenAI may signal the beginning of a new AI gold rush. What was formerly a cottage industry of enthusiasts working without the proper licencing can now turn their tinkering into legitimate businesses.
Hassan El Mghari, who owns TwitterBio, a service that employs ChatGPT’s computing capacity to generate Twitter profile text for users, believes that this release makes it much easier and more economical for businesses to integrate AI capabilities into their products.
Moreover, OpenAI updated its data retention guidelines, which should reassure companies considering trying out ChatGPT. The business has announced that it will no longer retain user data for more than 30 days and has pledged not to utilise user data to train its models.
David Foster, a partner of the London-based data science and AI consultancy Applied Data Science Partners, asserts that this will be “essential” in encouraging businesses to use the API.
Foster believes that the reason why they haven’t adopted the technology yet is because they were concerned that ChatGPT’s training models might eat up sensitive client information or important business data. “OpenAI’s decision to essentially say, ‘Hey, you can use this now, risk-free for your organisation,’ shows a lot of dedication. You won’t find information on your company in that general model, he claims.
According to Foster, this policy shift allows businesses to feel in charge of their data rather than relying on a third party—OpenAI—to regulate where it goes and how it’s utilised. He claims that you were effectively constructing this stuff using someone else’s architecture and data usage policy.
Because to this and the decreasing cost of access to big language models, a proliferation of AI chatbots is predicted in the near future.
Access to OpenAI’s higher-powered GPT3.5 API, which it debuted in June 2020 and which could produce convincing language when instructed but did not have the same conversational strength as ChatGPT, is 10 times more expensive than access to OpenAI’s lower-powered GPT3 API.
The Targum language translator for videos, which was informally based off of ChatGPT at a December 2022 hackathon, was created by Alex Volkov and is much faster and cheaper. “Normally, that doesn’t happen. The API world typically sees pricing increases.
This might cause a fresh wave of innovation and alter the economics of AI for many organisations.
It’s an incredible time to be a founder, adds Habib of QuickVid. Every app out there will have some sort of chat interface or LLM [large language model] integration due to how affordable and simple it is to incorporate. Individuals will need to adjust very quickly to communicating with AI.
