A brief background of the important news
In the early morning of Tuesday (May 14), Beijing time, OpenAI released its new flagship model, GPT-4o. GPT-4o's "o" stands for "omni", which means "omni" and is derived from the Latin word "omnis".
GPT-4o can perform real-time inferences on audio, visual, and text, with greatly improved speed and quality. In the past, GPT-3.5 had an average latency of 2.8 seconds for voice conversations and 5.4 seconds for GPT-4, and they had to be interrupted and restarted.
GPT-4o, on the other hand, can react to audio input in 232 milliseconds, which is comparable to the reaction speed of a person-to-person conversation, and can also interrupt and continue the conversation at will.
In addition, GPT-4o can also understand real-life scenarios in real time through the mobile phone camera, and then make feedback, such as identifying the mathematical equation 3x+1=4 on the whiteboard, and talking about the solution process step by step.
In other words, GPT-4o can hear, see, and speak like a human. Humans talking to it is like making a video call, and it's free to use.
So, what kind of industrial changes will GPT-4o bring?
The following is Wang Yuquan's important news commentary:
At 1 p.m. Eastern Time in the United States on Monday, which is 10 a.m. Western Time, OpenAI held its own press conference on time and released its new product called GPT-4o. o is the meaning of omni, that is, the almighty GPT is finally here.
On the surface, the press conference was too shabby, less than half an hour, and it was held in a larger living room, with a dozen people sitting scattered under the stage, which looked shabby, but it gave me a sense of shock that was definitely no less than GPT-3.5 back then, because we saw that a new era was born.
Why? I believe that many people have this feeling, that is, this time GPT is really available to ordinary people, because they integrate text, voice, pictures, and videos together, and can interact with you in various ways.
For example, they have made some scene designs, you can chat with GPT, and you can ask it to change the tone of voice, tone, and attitude when chatting. You can also ask it to watch a video, read out a mathematical formula on the video, and then teach you to derive step by step, not directly to tell you the conclusion, but to teach you to deduce step by step to reach a conclusion. Is it already possible to be a tutor for your child?
In fact, this is an important sign that GPT is available, why do you say that? Because we are all integrated in our daily lives.
For example, the teacher teaches the students to write on the blackboard, to communicate with the students in language, and to constantly discuss the things written on the blackboard, that is, it does need text, audio, video, including scene interaction, etc., and now these things are integrated in a unified way.
Of course, there's certainly a lot of room for discussion when we talk about how the technology behind it is implemented, but now it seems that it all is seamlessly connected, and it can even interrupt the conversation of GPT in real time, and GPT is able to continue, especially like a conversation between people and people.
Because in the past, human-computer interaction was particularly afraid of interruption, especially in the past, generative artificial intelligence, its content was to be continuously generated, and after you interrupted, it would be broken later, and now there is no problem to look at, so its intelligence has been significantly improved.
Of course, some people are saying that this is just integration, and we still expect a larger model and faster computing speed, so how can we play integration now? Isn't OpenAI strong?
We believe that integration is a form of strength. Just like the launch of the iPhone back then, in fact, it is also the result of integration, but what is the difference between its integration and the integration of other mobile phones? It's that the integration is very seamless, and the user experience is very silky.
In OpenAI, we saw that the user experience of the iPhone was very silky, so we said that this is an epoch-making product.
After OpenAI launched GPT-4o, I believe there must be a lot of discussion in China, and I guess it focuses on two points.
The first point is that there are a lot of companies that are going to die. For example, Character.AI, it is characterized by the ability to set the identity of the conversation when talking to users, which can be a young woman, a middle-aged man, and so on. Wouldn't it be more convenient now that GPT can change its identity with just one sentence and be able to turn into the character you want to talk to you without having to set it in advance? This is very important.
The second point is that there must be a lot of discussion, that is, when China will be able to achieve such a level, including what is the technical strength behind this. Because many people will think that this is not synthesis, right? This is what the Chinese are good at. But to achieve such a silky synthesis, there are actually a lot of technologies behind.
At this time, we saw that China is slowly making efforts on large models and has caught up with the pace of the world, but now OpenAI has come out with a more comprehensive thing, and the gap has been widened again, and it is also suspenseful when we can keep up again.
There will be a lot of discussion on these two topics, but this is not the most important thing I want to talk to you about today. There was a word that was mentioned many times in this short press conference, and it was called collaboration.
So what does collaboration mean? Why OpenAI does such an integrated thing, not to emphasize the technical performance, but to emphasize the ease of use of users, because they think that artificial intelligence should form a collaboration with people, and the real sense of collaboration is not the use of computers like we use, but I have formed a partnership with it, and this kind of cooperation is called collaboration, which is equal collaboration.
Of course, we saw some other videos later, where two GPT-4os can talk to each other, one when the eye recognizes the environment, and the other when the brain analyzes the environment. So what does that mean? It means that a lot of the future of collaboration will be machine-to-machine collaboration.
Then let's remind you that you need to act quickly at this time, because when the user experience has reached this level, and the consideration of collaboration has reached this depth, you need to quickly introduce GPT and generative AI into your workflow.
I even think that in the future, the most important position in a company will no longer be the technologist of artificial intelligence, but what? Attention, which many people may overlook, may be HR experts. Because human resources are the job determination work. Of course, this human resources expert must understand the professional skills of artificial intelligence, especially what we call the ability boundary.
Why? Because basically AI has come this far, as long as you can define the position, it can achieve it for you.
In other words, if you want to recruit someone, you used to go outside to find it after you described the job setting, but now what is it? You have to understand technology, and after you set up your position, tell GPT directly, please become such a person, and it will become, you have such workers in an instant, and even different workers have different positions, and you have to connect them.
At this time, you will find that the human resources experts and human resources leaders in the company have become very important, which is a particularly interesting point.
Now that artificial intelligence has developed to this point, don't feel that this is a distant black technology, but know that this is an advanced technology that can be used immediately, and arm yourself with science and technology in order to be invincible in the future competition.
So remind all businesses, all entrepreneurs, all entrepreneurs, to act now and take a good look at what GPT-4o can do? And then quickly rethink their work tasks, workflows, and job design, so that AI can be involved, so that their work can be truly energy efficient, and even achieve many results that were not achieved before, which is something that every enterprise needs to think about.
In addition, this year is the 8th anniversary of the science and technology training camp, and the friends of the innovation map have prepared the 8th anniversary benefits for everyone, from May 1st to May 31st, that is, if you buy a member of the science and technology training camp this month, you can get two months of school hours.
Interested friends should seize the time to scan the QR code in the event poster below, sign up, and let us understand the future together in the science and technology training camp.
The above is today's content, more detailed industry analysis and underlying logic, I will share it in the science and technology training camp. Welcome to pay attention to the global outlet WeChat account and sign up.
Wang Yuquan has to be heard, and we'll see you tomorrow.