At the recent RTE2024 Real-Time Internet Conference, industry leaders conducted a deep analysis of the trends in the AI industry. With OpenAI significantly reducing API call costs and intensified price competition in the Chinese market, generative AI is driving industrial transformation at an unprecedented pace.
Zhao Bin, Founder and CEO of Agora, pointed out that the next 10-20 years of technological development will focus primarily on enhancing the application capabilities of large models at the edge. This transformation will unfold comprehensively across four main areas: terminal devices, software development, cloud services, and human-computer interaction:
Terminal devices will evolve into AI PCs and AI Phones.
Software development will transition from "Software with AI" to "AI Native Software".
Cloud services will fully support model training and inference.
Human-computer interaction will primarily be through natural language dialogue.
McKinsey's latest report predicts that the global generative AI market will grow rapidly from $67 billion in 2023 to $1.3 trillion by 2032, with an annual compound growth rate of 42%. In this context, Agora is actively expanding its reach and has announced a partnership with the large model unicorn MiniMax to create China's first Realtime API.
Image source: Picture generated by AI, authorized service provider Midjourney
There is also good news on the cost reduction front. Yangqing Jia, founder of Lepton AI, predicts that AI inference costs could drop to one-tenth of the current level within a year. Meanwhile, with advancements in model compression technology, the performance of smaller models has approached that of larger models, and the "open source + fine-tuning" solution will become the mainstream choice for enterprise-level applications.
However, industry experts also warn of potential risks associated with AI development. Tieshan Wang, an engineer at Hugging Face, noted that while fears of AI replacing humans are premature, negative impacts have already emerged in some areas, such as the societal and psychological effects of video forgery, which also present opportunities for innovation and entrepreneurship.
Wei Wei, partner at MiniMax, is optimistic about the prospects of multi-modal AI in the creative industry. He believes that with the maturity of multi-modal technology, AI will bring efficiency improvements to creators in text, voice, music, video, and other areas, driving the upgrade of related industries.