OpenAI's New Model Could Revolutionise Science And Coding. Here’s How
Meet OpenAI o1-Preview.
By Sukhbir Cheema — 13 Sep 2024, 05:53 PM
OpenAI has released its latest series of AI models, called OpenAI o1-Preview
The new generative AI chatbots are designed to enhance the way AI systems tackle complex problems.
With a focus on improving accuracy and providing more reliable responses, these new models have been built to "think" longer before responding, potentially making them better suited for challenging tasks like science, coding, and mathematics.
Unlike earlier versions, which often delivered impressive but inconsistent results, OpenAI o1-Preview has been trained to refine its problem-solving process.
The models are designed to explore various methods, analyse their reasoning, and identify potential mistakes before giving a final answer
This upgrade addresses a long-standing issue with previous models, which were criticised for generating persuasive yet incorrect content, also known as "hallucinations".
Early tests suggest that o1-Preview is less prone to these errors, though OpenAI admits there is still room for improvement.
OpenAI hopes to take advantage of the new models to raise more money.
OpenAI is in an ongoing fundraising effort that could value the company at USD150 billion (RM650 billion). If successful, this would place OpenAI among the world's most valuable private companies.
According to reports, current investors include tech giants Microsoft and Nvidia, and sources suggest the company may receive an additional USD7 billion (RM30.2 billion) from MGX, a United Arab Emirates-backed investment fund.
OpenAI's rising valuation is tied to its rapid development of cutting-edge AI technology, which has garnered significant attention from both investors and business leaders. Despite the high expectations, OpenAI CEO Sam Altman has acknowledged the limitations of even their most advanced models, emphasising the importance of continued refinement.
As OpenAI pushes forward, the company is also strengthening partnerships with AI safety institutes in the US and UK, ensuring that its innovations are both reliable and secure.
