In this emergency episode of Big Technology Podcast, Alex Kantrowitz interviews Brad Lightcap, the COO of OpenAI, about the newly released GPT-5. Lightcap explains that GPT-5 is a next-generation model that dynamically chooses whether to reason about a problem, improving upon previous models by abstracting the model selection process for users. The discussion covers the improvements in writing, coding, and health-related tasks, as well as its increased speed and accuracy. Kantrowitz presses Lightcap on whether GPT-5 represents an exponential or incremental increase in capabilities, with Lightcap emphasizing the importance of reasoning time and structured problem-solving. They also discuss the shift towards post-training methods, the definition of AGI, and the potential impact of GPT-5 on average users, health care, and enterprise applications, including its coding and problem-solving abilities. The conversation also touches on pricing, profitability, and the future development of AI models.
Sign in to continue reading, translating and more.
Continue