OpenAI says upcoming 'Spud' AI model major step toward AGI
What's the story
OpenAI is gearing up to unveil its next AI model, codenamed 'Spud.' The move comes as the company looks to double down on its core products after facing competition from Anthropic in the enterprise market. OpenAI President Greg Brockman has teased the improvements we can expect from this new model during the Big Technology podcast. He claimed that "two years worth of research" has gone into developing Spud, which he says will set a new benchmark for AI models.
Model evolution
Spud will be a lot smarter
Brockman hinted that Spud, rumored to be released as GPT 5.5, is going to be a lot smarter than its predecessors. He even coined the term "big model smell," possibly referring to the fresh scent of a new car. The OpenAI chief also said that Spud will better understand contexts for tasks without requiring detailed user explanations. This could make interactions with AI models more intuitive and seamless for users.
Launch timeline
Spud is currently in its pre-training phase
Reports suggest that Spud is currently in its pre-training phase and could be launched later this year. The new AI model is crucial for OpenAI as it tries to manage its finances by focusing on core products. Despite the shift in focus, the company has decided to pause work on an erotic mode for ChatGPT and shut down its video generation platform Sora.
Strategic shift
AGI is the goal
OpenAI's decision to shut down Sora surprised many, especially after the company had signed a billion-dollar deal with Disney for the video-generation platform. However, Brockman said that the company made this decision to focus on achieving artificial general intelligence (AGI), a state where AI can think like humans. He explained that there has been much debate about how far text models can go and claimed they have definitively answered that question: "It is going to go to AGI."