OpenAI CEO Altman, who dropped out of Stanford to start his career, has received a lot of attention and welcome from students.
Yesterday, Jensen Huang personally delivered the world’s first Nvidia DGX H200 supercomputer to OpenAI. The photo of the three (Huang, Altman, Greg Brockman) caused a stir.
After the group photo, Altman did not idle, he appeared at the Nvidia Auditorium of Stanford University to deliver a speech, with about over 1000 students lining up.
Judging from the on-site photos, the whole speech process was fully attended, with some students who didn’t get a seat even standing in the corridor:
Stanford University computer science master’s student Wanrong He summarized the key points of Altman’s speech.
Altman shared his insights into the future development of AI, including the following three aspects:
First is the rapid progress of AI.
The next AI model after GPT-4 (like GPT-5) will be more powerful, which shows the astonishing speed of progress in this field. Altman believes that GPT-5 will be smarter than GPT-4, and GPT-6 will be smarter than GPT-5. We are far from reaching the top of the curve.
-
Although it has become easy to replicate existing models like GPT-4, real innovation lies in defining the next paradigm shift in AI capabilities.
-
Altman compared the potential of AI to the transformative impact of the iPhone on mobile computing.
Second is the path to AGI.
-
OpenAI’s mission is to realize AGI, and open sourcing AI may not be the best way to achieve this goal.
-
Free, ad-free ChatGPT is the way OpenAI pursues its goals while positively impacting society. In order to make AI accessible to everyone, the focus should be on making AI computing more affordable and widely used on a global scale to eliminate inequality.
-
Injecting emotions into ChatGPT, a tool designed to enhance human capabilities, is unnecessary.
-
There is no need to worry about super AI because each new model is considered deficient compared to future models, which allows for continual progress.
Finally, he talked about the opportunities for innovation and start-ups.
-
Focusing solely on resolving the current limitations of AI may prove to be futile, as future models such as GPT-5 and GPT-6 may render these efforts obsolete.
-
Breakthroughs like Sora can redefine entertainment, creating personalized interactive experiences that blend film and game quality.
-
AI start-ups have huge potential, but relying solely on AI cannot guarantee their success. They must still adhere to basic business principles.
-
Although AI may surpass humans in certain fields like chess, people still prefer to watch human players compete. However, Altman pointed out a counterexample, like teenagers preferring to talk with an AI therapist rather than a human one.
In summary, Altman’s insightful speech provided a transformative vision for the future development of AI, emphasizing the rapid progress of AI, pursuing AGI, innovative methods of utilizing AI potential while addressing social challenges, and the evolving dynamics between humans and AI.
In addition, Stanford Daily also summarized some of the key points of Altman’s speech.
Altman told everyone that since the establishment of OpenAI in 2015, OpenAI’s mission has not changed. It remains committed to ensuring that general artificial intelligence benefits all of humanity, but its corporate structure has been constantly adjusted and will continue to be adjusted. OpenAI’s current structure includes a for-profit subsidiary, which issues stocks to raise funds but is still constrained by the mission of the non-profit organization.
Because the computing power and talent costs required for OpenAI research are high, Altman says he doesn’t care about burning through $500 million, $5 billion, or $50 billion a year – he believes AI will ultimately create much more value for society, as long as a way can be found to pay the bills.
Since its release, everyone has spoken highly of ChatGPT, but Altman says that ChatGPT is at best a bit awkward, GPT-4 is the model that everyone will have to use in the future because it is the dumbest, but they believe in the power of iterative deployment.
Altman believes that to be prepared for technological progress depends on responsible development and iterative deployment, even if the model is not perfect.
Although the prospect of AGI is frightening, Altman believes it will be the scaffold that allows society to achieve greater heights. “People will bring surprises to life with better tools. All of history shows that when more leverage is given to humans, they do more amazing things.”
Altman believes AGI can bring high-quality education, disease treatment, entertainment, and space exploration to society. He believes AGI can replace lawyers and doctors, making it easier for people to access legal and medical services, thus helping the poorest people in the world, not the wealthiest.
Altman believes that although new technology may initially shock society, society adapts quickly. Altman stated that GPT-4, which they launched a year ago, caused two weeks of panic. People thought it was a crazy thing and the world was forever changed. But now, what people care most about is – where is GPT-5?
© Copyright notes
The copyright of the article belongs to the author, please do not reprint without permission.