OpenAI Co-Founder: AI Training Data Is Running Out
According to him, the growth of computing power due to improvements in algorithms, hardware and software is outpacing the total amount of data available for training AI models. The researcher compared existing information to fossil fuels, which will eventually run out:
“Data is not growing because we only have one internet. […] It was created in some way, and now we are using it, and we have reached the peak, there will be no more of it – we will have to deal with the information that is there,” Sutskever emphasized.
OpenAI co-founder believes that AI agents , synthetic data and faster computing are the next stage in the evolution of artificial intelligence, which will lead to the emergence of superintelligence.
AI agents are neural networks that are capable of independently performing operations on the Internet, for example, booking a flight or hotel.
In October, AI startup Anthropic released an updated version of the Claude 3.5 Sonnet model. It can interact with a computer like a human: move the cursor, click buttons, and type text.
In December, Google introduced its first AI agent that can operate independently on the Internet.
Earlier, it became known about similar developments from OpenAI.