Skip to content

While humans are busy rolling around, AI has completed the biggest evolution in recent years

For a long time, the term AI (artificial intelligence) became less exciting. People have been unable to avoid it, but at the same time they have found that both the technological evolution of AI and commercial applications seem to have encountered bottlenecks. People haven't experienced the amazingness of AlphaGo for many years, and the industry has never experienced business opportunities like the popularization of voice assistants. Even many investors only reluctantly turned around and looked back when there were no new stories. Opportunities for AI. However, in 2022, when all human beings are busy rolling around, AI is undergoing one of the biggest evolutions in recent years. AI suddenly became popular on October 18th. StabilityAI, an artificial intelligence company that became popular because of the launch of the Stable Diffusion text-image AI generation model, announced the completion of a seed round of financing of 101 million US dollars, bringing the entire company's valuation to 1 billion US dollars, becoming the so-called "Unicorn" companies. This is only two years after StabilityAI was founded. Source: StabilityAI official website Even in terms of the development of the technology Internet industry, the growth rate of StabilityAI is amazing. This speed is a microcosm of the explosive growth of the global AI industry since 2022. At this moment, the StabilityAI's Stable Diffusion open source model is popular. The world is less than 2 months away. This rapid advance is a real revolution, especially against the backdrop of weakening global economic expectations. Like all revolutions, this AI revolution didn't happen overnight. People have always had a dream to use artificial intelligence (AI) technology to expand the boundaries of existing human intelligence, knowledge and creativity, but the learning ability brought by the complex structure of the human brain far exceeds the ability of humans to build AI, so AI can only Through a variety of specific deep learning models to single-point breakthroughs in some specific fields, such as alphaGO used to learn Go, or help find pulsar candidates through astronomical big data. AIGC, that is, content creation based on AI capabilities (including text, pictures and videos, etc.) is also an important category. Before 2022, due to the limitations of core technologies, this field has been tepid because AI does not The magic of turning stone into gold, it does not have the ability of human beings to create out of thin air. AI's "deep learning" training is not self-awareness of self-learning. It is a process of collecting a large number of samples to allow AI to summarize the rules from the massive data, and then reproduce the content based on the rules according to human instructions. It is also affected by the core algorithm, Hardware conditions, database samples and many other limitations. A 2018 work generated by a neural network, trained by researcher Robbie Barrat on a large number of nude paintings, mostly women.