Meta Unveils New AI Models To Reduce Human Involvement In Development
Meta, the parent company of Facebook, has introduced a new suite of AI models designed to reduce human involvement in the development process. Announced on Friday, the models include a “Self-Taught Evaluator” developed by Meta’s research division, which leverages AI-generated data to train itself, eliminating the need for human input at that stage.
The new model also employs the “chain of thought” technique, similar to recent models from OpenAI, to enhance its ability to solve complex problems in fields such as science, coding, and mathematics. This approach allows the AI to break down and process tasks more effectively.
Meta’s innovation could lead to the creation of autonomous AI systems capable of self-improvement, learning from their own errors without relying on traditional, expensive methods like Reinforcement Learning from Human Feedback (RLHF). This development signals a major step toward more cost-efficient and independent AI systems.
