The data human learning platform is launched: www.shujurenclub.com
01
Why is it called 4o
GPT-4o, where "o" stands for "omni", which means omnipotence. It has not only reached unprecedented levels in text processing, but has also made major breakthroughs in image and speech processing. GPT-4o's ability to perform inferences on audio, visual, and text in real-time, providing human-like response times, is a huge step forward in the field of AI.
GPT-4o is a step towards more natural human-computer interaction, accepting a combination of text, audio, and images as input, and generating any combination of text, audio, and image outputs, "GPT-4o is particularly good at image and audio understanding compared to existing models." ”
02
What GPT 4o brings to the table
Among them, the focus is on multimodal and real-time capabilities, which not only raise the ceiling of AI technology, but also provide new ideas for future research directions and application scenarios. Similar to the scene in human life, it is actually a multimodal scene, for example, when you are chatting with your family, you not only have to use words to express, but your eyes will also observe and obtain information, which is a typical multimodal scene.
Whether it's GPT-3.5 at the beginning, GPT-4 at this time last year, GPTs at the end of last year, or Sora at the beginning of this year — OpenAI has proven once again that it won't disappoint. Although competitors such as Google, Claude, Character AI, and Perplexity are grabbing more new users and capital, OpenAI still proves that it has the ability to lead the "high ground" of technological innovation. It has to be said that ChartGPT is constantly raising people's expectations for large models, but the final application remains to be seen, and it also depends on the actual implementation.
I took a photo of the Xiaomi car and let GPT recognize it
03
What will happen to the big model in the future
There's a popular term in the development industry, and it's called PMF. In the process of implementing new technologies and products, it is necessary to find the best fit with the market (PMF), that is, the precise combination of products, target markets and business models. Many entrepreneurs actually have a hard time clearly defining their products, defining their target customer groups, and how to effectively integrate the two in the early stages. Therefore, most of them are holding hammers to find nails, and finally find that there are many difficulties in landing and cannot solve the actual problems of users in the application scenario.
Although the current technological progress of large models provides unprecedented opportunities, it is also accompanied by problems such as insufficient technical stability and unclear understanding of market demand. In particular, consumer-oriented (To C) entrepreneurship may face greater challenges, including the requirements for understanding human nature and operational capabilities, but many domestic startups will not give up on the To C market, and will give priority and vigorously invest in the To C market. At this stage, the entrepreneurial path for enterprise customers (To B) is actually a more feasible choice, especially when you need MVP applications and attempts in the early stage, you can give priority to doing some copilots that assist the workflow, and you should also distinguish whether you are solving an itch or a pain point.
Finally, I am still firmly confident in the development of artificial intelligence, and I hope that everyone can pay more attention, apply and practice. Looking forward to finding PMF in the future, and the killer application will appear as soon as possible!