A brief background of the important news
Competition in the generative AI space is intensifying.
On May 14, OpenAI just released its new model GPT-4o, and Google held the 2024 I/O Developer Conference the next day, May 15, and released more than a dozen generative AI products. For example, Project Astra, which benchmarks the speech-visual interaction model GPT-4o, Veo, benchmarks the video generation model Sora, and the new version of search implanted in Gemini, etc.
Demis Hassabis, CEO of Google Deepmind, explains the demo video of Project Astra. In the video, a staff member holds a Google pixel phone, turns on the camera, points at the desk, computer screen, blackboard, window and other places, and directly interacts with the AI in language, allowing the AI to say what it sees and solve some procedural problems.
Global outlets
, Like31
OpenAI's GPT-4o has the same features, but the difference is that OpenAI's technicians made an interactive demonstration directly at the conference, while Google only showed the demo video.
So, will Google be able to push back against OpenAI? What does the future hold for them?
The following is Wang Yuquan's important news commentary:
At 10 a.m. EST on Tuesday, the same time the day after OpenAI's launch, Google held its own Google I/O conference, an annual conference dedicated to developers, for years.
This year's Google I/O conference seems to be in full swing again, whether it is from the improvement of search, a large number of photo searches are getting smarter and smarter, or their own large models, providing more and more options for large models, as well as their own hardware, are vividly displayed.
Part of it was explained by Demis Hassabis, the founder of DeepMind, and it is very similar to yesterday's OpenAI demonstration, that is, a mobile phone camera can recognize various external objects, and can interact with people through the camera, and when you ask it what the situation is, it will answer you. Even if you ask it what kind of program it is, it will recognize it.
But overall, my feeling is that Google is really stressed this time, why do you say that?
Because this comparable presentation to OpenAI is not real-time, but a video shown by DeepMind's Hassabis.
We know that the characteristic of DeepMind is that it is still relatively academic. But we all know that Google has a flaw in the past, that is, DeepMind has never been able to integrate with Google's own AI team, let alone its other departments.
It is said that there was a rumor in the past, which should be true, that is, the new CEO of Google once wanted to integrate the DeepMind team with Google Brain, which is another AI R&D team of Google, because there is no need for one company to have two AI R&D teams.
But the founder of DeepMind directly approached the founder of Google, saying that the condition you promised when you bought me was to remain independent, and I still want to be independent, but as a result, "the emperor interferes in politics", and the founder of Google stood up and said, or don't merge. So we're kidding that Google's future is really worrisome.
First, on the surface, AI has made a lot of progress, but it may take time to integrate it into its own business. Including we said that some of the presentations are just video demonstrations, not real-time operations, and they are not integrated with the business.
The second is that Google's biggest revenue business right now is still search. On the surface, this search has also been improved a lot, but the problem is that it is a left-handed and left-handed fight, which means that its artificial intelligence search will definitely harm the interests of traditional search.
Third, Google needs to complete a revolution like Microsoft, so that advanced AI products are implanted into its entire product line.
Think about it, although it is an in vitro product and an in vitro company, how resolutely Microsoft implants OpenAI's products into its own system, which can reflect the power of artificial intelligence everywhere, which is very important.
So why can't Google do it? The reason is that the "Emperor Taishang interferes in politics" just mentioned, although the CEO is asked to go all in artificial intelligence, but the various departments of artificial intelligence cannot be connected, which is a big problem.
So on the surface, Google is booming, and it seems that it is still leading the progress of artificial intelligence, especially since even the Transformer that gave birth to the biggest progress in artificial intelligence is from Google's research laboratory, but this is precisely Google's problem.
Because OpenAI is now focusing more on the optimization of functions and the tuning of performance, it is no longer a breakthrough in basic principles. However, Google is actually good at breaking through basic principles, and it is not Google's advantage when it comes to performance tuning and user experience improvement.
If the entire artificial intelligence has entered the stage of performance tuning, Google may become "get up early in the morning and catch up late" and slowly be left behind by other competitors.
So it looks like OpenAI has a launch conference, and Google has made a similar product, and OpenAI should be nervous, but we think that on the contrary, OpenAI has taken the lead again, because it has taken another step forward in performance tuning and user experience, and Google is just not good at it.
In addition, the search that used to be the source of his huge income and profit has become a heavy burden for himself, whether he wants a strong man to break his wrist, this matter is still very difficult to decide, especially the CEO does not have such a great decision-making ability, the emperor is staring at it, staring at income, profits, and performance every day, and it is unlikely that you want to let him break his wrist as a real strong man.
So we say that on the surface it looks like a boom, but Google's performance is likely to go down in the next few years.
Of course, this is a boon for the AI industry, or for us AI users. Because it can be seen from the press conferences of OpenAI and Google, the main theme for a period of time in the future will enter the era called terminal intelligence, that is to say, whether it is a PC terminal or a mobile phone terminal, more and more people can use artificial intelligence.
If you look at the comparison of the data volume of the Internet and the wireless Internet, you can see that once you enter the era of mobile terminal intelligence, the number of users will skyrocket.
I'm also trying out OpenAI's new GPT-4o today, and I found that especially at noon, I can't even connect, and GPT directly tells you that the connection is interrupted. Because there are too many users, the server is too tight.
It can be seen that there is a phenomenon that is likely to occur in the future for a period of time, that is, the number of visits to GPT through mobile terminals will skyrocket, and the number of artificial intelligence used on mobile phones will skyrocket.
Of course, we believe that there is a high probability that a large number of mobile phone-based artificial intelligence apps will be born in the next two years. I believe that many of these AI-native apps were born in China, why? Because we are good at developing mobile apps, it is very worth looking forward to.
Artificial intelligence will change the pattern of the entire industry and the territory of the industry, and at the same time, it will also bring us a lot of new opportunities, and it will bring a lot of new experience to users, which is really an era worth looking forward to.
Regarding the OpenAI and Google press conference, I will do more analysis and explanation on AI technology and industry trends in tonight's science and technology training camp course, and welcome interested friends to sign up.
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.