When Mila Mulatti, chief technology officer of OpenAI, released the desktop version of ChatGPT and the new flagship model GPT-4o, the development of OpenAI entered a truly singular moment.
OpenAI's 26 minute press conference still caused a lot of waves.
Whether ChatGPT-4o is completely free or ChatGPT for PC desktop is launched, it shows us that OpenAI is leading the way.
However, if you just blindly boast about the leadership of OpenAI, it will certainly not be of much benefit.
The reason is that OpenAI does have a comparative advantage in many aspects. This kind of leadership is not only reflected in the gap between OpenAI and Google, but also reflected in the gap between OpenAI and Chinese AI players.
Through the OpenAI press conference, we need to see more profound changes taking place in the current AI industry, and thus find new opportunities suitable for the current development trend.
If we have to find out the enlightenment of this short conference of OpenAI, I would like to see it as a direct reflection of AI on the Internet.
From this point of view, AI will really change from being confined to the laboratory to being recognized, used and accepted by more and more people.
If we regard OpenAI before ChatGPT-4o as a preacher, then OpenAI after ChatGPT-4o is more like a practitioner.
In a word, OpenAI is opening AI.
In the future, we will see more and more scenes and industries where AI begins to appear. We will see more and more players begin to join in the process of popularizing and landing AI.
With this in mind, we may be able to better grasp the profound significance of OpenAI, a 26 minute event.
one
The "king bomb" thrown out by such a conference as OpenAI is actually the launch of ChatGPT-4o, a completely free application.
On the surface, one of the most important reasons why OpenAI will be completely free is that its user growth has entered a bottleneck period. In order to expand more users, OpenAI has started to obtain new users through free ways.
Since the launch of ChatGPT last year, the traffic peak will be reached in May 2023, and the network visits will reach 1.8 billion times, but the traffic will decline later. In March 2024, the number of visits to ChatGPT remained stable at 1.8 billion, with no significant growth.
When OpenAI further reduces the threshold of use through ChatGPT-4o, which is completely free of charge, the first immediate effect is the growth of users.
For OpenAI, which has fallen into the bottleneck period of user growth, it is undoubtedly a very good attempt.
If we regard ChatGPT-4o's free service as an attempt by OpenAI to increase the number of users, then ChatGPT-4o's application of more scenarios to the end mode and code generation capabilities will expand the growth of OpenAI's users into a broader field.
In fact, this is similar to the development of the Internet industry.
As we all know, as early as the Internet was not popular, people's use of the Internet was not as popular as it is now. Even in many cases, people must pay for the use of the Internet, just like people's use of AI now.
Obviously, it is difficult for the Internet to release greater development potential and achieve an era called "Internet" through payment.
Therefore, in order to let the Internet release greater development potential and make the Internet truly an era, what must be done is to make the Internet services charged in the past free, and really use this way to achieve the maximum popularity of the Internet. On this basis, we will look for the "Internet+" business model.
It can be said that free has truly made the Internet an "infrastructure", and free has also promoted the continuous maturity and improvement of the "Internet+" model.
The same is true for AI.
Undoubtedly, OpenAI shows us the power of AI and its positive impact on efficiency improvement through ChatGPT.
However, in many cases, people's perception of AI is still completed under the condition of charging, which is still constrained by charging.
In such a situation, AI is difficult to become an infrastructure, and even more difficult to release new and greater development potential.
Therefore, if we must find the specific expression of AI on the Internet, it will be completely free, and it will promote the popularization of AI to the maximum extent and breed more new business models, which is undoubtedly the most direct expression.
When OpenAI opens the door to free, we may also see more players join it in the future, which will really bring people to a new stage of development where everyone can use AI and everything can use AI.
two
In addition to the free ChatGPT-4o, OpenAI showed us more about GPT-4o's model capabilities, benchmarking, model security and limitations at the press conference.
In terms of model capability, Before GPT-4o, the average delay of ChatGPT voice mode conversation was 2.8 seconds (GPT-3.5) and 5.4 seconds (GPT-4).
Now with GPT-4o, OpenAI has trained a new model end-to-end across text, visual and audio, which means that all inputs and outputs are processed by the same neural network.
At the traditional benchmark level, GPT-4o achieves GPT-4 Turbo level performance in terms of text, reasoning and coding intelligence, and is at a high level in multilingual, audio and visual functions.
On five MMLUs (common sense questions), GPT-4o scored a new high of 87.2%, far surpassing Google Gemini Pro 1.5, Ultra 1.0, and its own GPT-4T and GPT-4.
In terms of model security and limitations, GPT-4o has built-in security in cross pattern design by filtering training data and refining model behavior after training. And created a new security system to provide protection for voice output.
According to the evaluation of network security, CBRN, persuasiveness and model autonomy, GPT-4o scores no higher than medium risk in any of these categories, and the team continues to reduce the new risks found.
If we want to find the information that these features of ChatGPT-4o convey to us, in the final analysis, it is to enable more users to use ChatGPT-4o more quickly, more safely and more efficiently.
In simple terms, the evolution of OpenAI on ChatGPT-4o is precisely for the better commercialization of its products.
Through this, we can also see that the development of AI is on the road of the Internet.
As we all know, in the Internet era, we have experienced DOS system, Windows system and later iOS system, Android system and other operating systems.
If we summarize and define these operating systems, in the final analysis, one of the ultimate goals of their continuous upgrading is to make the Internet more commercialized and better integrate the Internet with business scenarios.
When the Internet has evolved in this way, we can see the portals, "Internet+" applications and various applications born in the era of mobile Internet.
It can be said that the most direct result of the continuous iteration and upgrading of the Internet is the commercialization of the Internet.
In fact, AI is also following such a development context.
In the final analysis, AI will finally have to settle down in commercialization to truly unleash its development potential to the maximum.
Otherwise, the so-called AI is still just a beautiful story told to the capital market. When the enthusiasm of capital ceases, especially when the development of AI really needs self blood, its development will still face problems of one kind or another.
Whether it's Robin Li of Baidu or 360 Zhou Hongyi , and even Zhu Xiaohu of Jinsha River Venture Capital have expressed such a view on AI on different occasions.
In the final analysis, the views they expressed are actually a central idea.
Such a central idea is actually to let AI focus more on scenarios and applications, and more on commercialization, rather than just staying in the laboratory, and not just limited to niche existence in limited scenarios.
Through the release of ChatGPT-4o, we can see that OpenAI is exploring and practicing the ways and methods of commercialization. Starting from this, we will see more new signs of AI's development along the path of the Internet.
three
Looking back at the development of the Internet, it is easy to see that one of the most important reasons why the Internet has achieved such great development and become a way of life is that it has been popularized to the maximum extent.
Today, the Internet has become almost everyone's lifestyle.
The advent of a new era dominated by live broadcast and short video has brought the popularity of the Internet to a new height.
For AI, in order to obtain new development and become a new way of life, it is necessary to re interpret the path of the Internet in the past, and it is necessary to re interpret the popularization action of the Internet in the past.
If we want to find the new enlightenment provided by OpenAI, it may be another aspect worth paying attention to.
When OpenAI appears in front of people with ChatGPT chat robot, it only exists in such a scenario as chat, and it only solves the problem of simple logical reasoning.
Today, ChatGPT is no longer a simple chat robot. It can not only chat conversations, pictures of students, but also videos of students. It can even interact with people as fast as people react.
OpenAI said that GPT-4o ("o" stands for "omni") is a step towards more natural human-computer interaction - it accepts any combination of text, audio and image as input, and generates any combination of text, audio and image output.
GPT-4o can respond to audio input in as short as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time.
Its performance in English text and code matches that of GPT-4 Turbo, and its performance in non English text is significantly improved. At the same time, the API is also faster, and the cost is reduced by 50%. Compared with existing models, GPT-4o is particularly good at visual and audio understanding.
In the final analysis, OpenAI is to improve the threshold of communication and exchange between people and AI by constantly improving the interaction ability of ChatGPT, and constantly make OpenAI products more popular.
If we summarize and define the interaction improvement of OpenAI and the past development of the Internet, their common ultimate goal is to make AI more popular and truly make AI a new way of life.
Therefore, if we want to find the new enlightenment from the OpenAI launch conference, it is undoubtedly another aspect that deserves our most attention to keep AI popular and moving forward in the direction of the Internet.
As OpenAI CEO Altman said in his blog after the press conference:
"I am very proud that we can provide the most advanced models in the world for free in ChatGPT without advertising or other interference. At first, OpenAI's vision was to develop AI technology and use it to bring benefits to the world. However, the reality is that we have developed AI technologies, and others have used these technologies to create outstanding results that benefit the world. As an enterprise, we have many services that need to be charged, but this does not prevent us from supporting the provision of free top AI services to billions of users around the world. "
"The newly launched voice (and video) mode is the best computing interface I have ever used. It gives people a feeling of AI in the movie, which makes people wonder that it is true. Achieving a response speed and expression ability similar to human beings marks a major change. The initial ChatGPT has demonstrated the potential of language interfaces, and this new technology has made a qualitative leap in experience. It is quick, intelligent, interesting, natural and practical. Before, I had never felt so natural when I talked to a computer; But now, I finally have this feeling. As we gradually add personalized options, access to personal information, the ability to perform operations on behalf of users and other functions, I can really foresee an exciting future: we can use computers to complete more things that were previously unimaginable. "
epilogue
If you have to look for the information delivered by a 26 minute press conference like OpenAI, it is clear that OpenAI is on the road of the Internet.
It can be said that through such a press conference, OpenAI has completed the perfect transformation from an AI preacher to a practitioner.
In fact, AI needs such a free, universal and commercial competition.
At this point, OpenAI starts to open AI.
The reason is that only after such a popularization and commercialization, the development of AI is not only a laboratory presence, not just a niche presence, but also an Internet like presence that can be deeply integrated with thousands of scenarios and industries.
With this as a starting point, ChatGPT can truly become what Nvidia CEO Huang Renxun calls "the iPhone moment", and GPT-4o is undoubtedly the singular moment of OpenAI.
- End -
Author: Meng Yonghui, senior writer, columnist, industry observer, well-known KOL, digital economist.
