Recently, Sam Altman, founder of Open AI, revealed the latest breakthroughs of the GPT big model in an in-depth conversation with Garry Tan, president of YC. He announced that Open AI is about to launch an open source model, which has attracted widespread attention in the technology circle.

Altman revealed that this upcoming open source model will far exceed industry expectations, and its powerful local operation capabilities will greatly promote the popularization and innovation of AI technology. He emphasized that the launch of this model marks that AI technology has entered a new stage of development.

Altman also revealed some information about GPT-5. He said that GPT-5 is expected to be released this summer. It will be a model that supports multimodal input, including voice, images, code and video. Although GPT-5 has not yet fully realized Open AI's ultimate vision for future models, it is undoubtedly an important step towards this goal.

The fully multimodal model that Open AI is pursuing will have deep reasoning capabilities, be able to delve into and generate real-time video, and write a lot of code at the same time to instantly create new applications for users. Altman pointed out that when Open AI realizes this vision, it will bring a revolutionary computer interface.

Altman further stated that the capabilities of current AI models have far exceeded the application level of existing products, and there is a huge "product overflow" space. This means that even if the capabilities of AI models are no longer improved, they are enough to support the development of a large number of new products. He mentioned that the cost of GPT-3 has dropped significantly in a short period of time, and this trend will continue, making the cost of using AI models lower and lower.

ChatGPT's memory function is also a highlight emphasized by Altman. He believes that this will enable AI to develop into an entity participant that understands users, runs continuously, and actively provides help. At present, some users have set ChatGPT as an operating system, connecting their lives to multiple data sources, achieving unprecedented convenience.

Altman also quoted the views of Greg Brockman, president of Open AI, calling this year the "year of the intelligent agent." He described the AI ​​agent as a "third-level AGI" (L3), which can perform hours of work in front of a computer like a junior employee. Altman pointed out that most of the work in the world that is done in units of a few hours in front of a computer may be replaced by intelligent agents.

In order to more clearly depict the development path of AGI, Altman divided it into five levels: chatter (L1), reasoner (L2), actor (L3), innovator (L4) and organizer (L5). He believes that with the continuous advancement of technology, AI will gradually develop from a simple chat tool to an intelligent agent that can act autonomously, innovate and even organize work.

Finally, Altman called on entrepreneurs to seize the opportunity of technological change. He believes that now is the best time to start a business in the history of technology. AI will improve people's quality of life faster and more profoundly, just like the invention of the transistor. Entrepreneurs should seize this historic opportunity to create new value for the world. Because when industry technology undergoes huge changes, large companies often find it difficult to adapt, while small companies have more advantages because of their fast iteration speed and low cost.