Reported 11 days ago
OpenAI, Google, and Anthropic are experiencing significant challenges in advancing their AI technologies. OpenAI's new model, Orion, failed to meet performance expectations, particularly in coding tasks, while Google and Anthropic also reported issues with their upcoming models. The companies are struggling to find high-quality training data and justify the hefty costs of developing new systems. Despite these hurdles, experts suggest that breakthroughs may hinge on exploring diverse training approaches rather than relying solely on traditional scaling methods.
Source: YAHOO