12:19 - 10.01.2025
January 10, Fineko/abc.az. Elon Musk agreed with experts’ statements regarding the fact that there is little real data left for training AI models.
ABC.AZ informs that TechCrunch writes about this.
In December, Ilya Sutskever, co-founder of OpenAI and founder of AI startup Safe Superintelligence, said that the industry has reached a peak in data usage. In his opinion, AI agents, synthetic information and accelerated computing are the next stage in the evolution of artificial intelligence, which will lead to the emergence of superintelligence.
"The only way to complement [real-world information] is with synthetic data, when AI creates [training information]. With such materials, [AI] will sort of evaluate itself and go through this self-learning process," the entrepreneur added.
In 2024, AI startup Anthropic used synthetic data to train one of its flagship Claude 3.5 Sonnet models. Meta has refined its Llama 3.1 neural networks with the help of AI-created materials. OpenAI also uses synthetic information to train o1 - a "reasoning" artificial intelligence.
AI startups have started looking for new ways to scale due to the lack of high-quality data.
