At the recent Snowflake Summit, OpenAI CEO Sam Altman delivered a strong warning to business leaders across various industries: companies must "act now" in the rapidly advancing field of artificial intelligence (AI). He pointed out that waiting for technology to "stabilize" is no longer a wise choice, and those who can iterate quickly will gain future competitive advantages.
Altman emphasized that the success or failure of enterprises in the AI field depends on their iteration speed. He said, "Companies that can quickly adapt and reduce error costs will ultimately become winners." At the same time, Snowflake CEO Sridhar Ramaswamy agreed with this view, mentioning that businesses should approach AI with curiosity rather than cautious observation.
Image source note: Image generated by AI, image authorization service provider Midjourney
In the exchange, Altman also warned that many traditional business assumptions have become obsolete, and AI is redefining the rules. Companies that are still观望 have already missed opportunities. He frankly stated that AI is not an unreachable dream but a tool that can be immediately put into use today.
According to Altman's introduction, over the past year, the reliability of AI models has significantly improved, leading to rapid growth in OpenAI's enterprise business. Large enterprises have gradually realized that AI can accomplish tasks once thought impossible. Ramaswamy added that background information and computing power are key to enhancing AI reliability. He particularly mentioned that the concepts of "retrieval" and "memory" are no longer empty terms but essential components of building AI infrastructure.
He used OpenAI's newly launched coding agent Codex as an example, stating that this tool gives people a glimpse of the potential of general artificial intelligence (AGI). Currently, Codex functions like an intern who can work for hours, but in the future, it might evolve into a senior software engineer who can work continuously for days.
Regarding the timeline for achieving AGI, Altman did not provide a specific answer, but he noted that AI capabilities are growing at an "astonishingly smooth exponential curve." He believed that if someone from 2020 saw ChatGPT, many would mistakenly think it is AGI. Ramaswamy then used the analogy of a submarine to emphasize how the definition of AGI evolves over time.
Finally, the two leaders also discussed how they would utilize these resources if they had 1000 times more computing power to conduct AI research or breakthroughs in RNA expression, aiming for revolutionary advances in disease treatment, showcasing the immense potential of AI in scientific discovery.