If 2023 has given us enough "surprises", 2024 will undoubtedly give us the privilege of seeing more "historical turns". The word "more" is used because the world is not singular. Breakthroughs in different fields will eventually converge into an irreversible torrent of the times. Today's human society is on the eve of change. The forces of progress and backwardness pull each other to form a solid "dam". Behind the dam is a huge amount of energy that is constantly accumulating. Although the old and backward forces are still stubbornly resisting, cracks have begun to appear in this dam. Streams of water began to gush out one after another. The result will definitely come sooner than expected! We are bound to usher in a whole new world. But it is not destined that everyone will be able to see tomorrow's sun as they wish, provided that they first walk out of tonight.
I added the previous paragraph after I finished writing this article. It's just a little bit of my personal feeling in the past two days. The following is the text for today. On February 16, without any news revealed in advance, OpenAI suddenly released its first Wensheng ** model: SORA. If there is one word to describe people's reaction after seeing Sora in the past two days, I think it would be "ecstatic". Does anyone still remember? It has only been a little more than a year since OpenAI first released ChatGPT on November 30, 2022. But now, not only has it been iterated to GPT-4Turbo, but SORA has also come without warning. Looking back at GPT35, it will give birth to a strange feeling of witnessing "a naughty boy grow into a young man overnight". It's rare that even Lao Hu feels "restless" this time.
Unlike before, I didn't write an article to express my opinion first, but set aside more time to read technical articles from industry insiders. After all, SORA now only invites professional users to participate in the experience. But what can be clearly felt is,After the release of SORA, the social repercussions brought by this time are very different from those in the past. First of all, the focus of discussion is not only on the advancement of SORA in terms of technology, but also on the beginning of a more profound reflection. Not only did we rethink the development path of original technology, but we also began to rethink education and beyond. People are starting to try**:What kind of future should we plan for, give ourselves and our children a chance. Reflection and self-examination have always been indispensable qualities on the path to progress. This was originally a good thing, but I didn't expect it to trigger another kind of social repercussions. Some people say that this is "growing the morale of others and destroying their own prestige". Criticizing those who put forward reflection is "creating anxiety", "kneeling in front of the West for a long time, even the knees have taken root", "ulterior motives". It's so much that even Lao Hu can't stand it. On the 18th, he issued an angry denunciation: This kind of person who insists that SORA "doesn't mean anything" with his eyes open is too irresponsible, and they are pouring ecstasy soup into the majority of netizens in doing so. What is the difference between this and the Qing Dynasty's accusation that foreign goods are "strange and ingenious"?
Those who are familiar with Lao Hu know that except for **, Lao Hu rarely participates in the discussion of the same thing or the same topic repeatedly. People's traffic is expensive. But from Lao Hu's two long articles before and after, and the content of ** was specially made, it can be seen that although Lao Hu is old, his sense of smell is still sensitive. So,What makes Lao Hu "restless"? What exactly does SORA mean? I think that with Lao Hu's "shrewdness", he must be able to see the future one step ahead of many people.
The developers of SORA are very low-key to call SORA a "Wensheng ** model". But my opinion is,The revolution brought about by SORA can allow humanity to begin to come into contact with the realm of "gods" like never before.
Judging from the several paragraphs released by SORA, it has completely jumped out of the traditional process of completing the final process through 3D modeling and then lighting rendering, but directly presents the world that people can understand. I studied engineering and have been using professional modeling software such as CAD since college. After work, I taught myself 3D animation. I'm very aware that a lot of the photorealistic effects of movies are modeled by huge computers and then rendered frame by frame. Even if it is an ordinary beam of light, it is a kind of existence with volume. This ensures that no matter what angle you look at the light, you will see the same results. If it's a beach, a fountain, waves and hair are the most troublesome. To achieve realistic results, you need to use more special effects. As for movies like Alita: Battle Angel, build a skeletal model first, then add muscle, fat tissue, and **. This ensures that even every subtle expression of the character appears "advanced". This comes at the cost of a huge time investment and increasing costs. For example, Weta Studios, the team behind the project, said that the company had nearly 800 people working on the project, using 30,000 computers for computing and production, and the team spent a total of 43.2 billion hours.
But all this is being broken by the arrival of Sora. Take the following scenario, which many authors are citing, as an example. In the traditional model, creating a city, even if it is just a part, is a staggering amount of modeling work. Not to mention the complex lighting, reflected light from pedestrians and streets. The creative team needed to arrange the body movements of each animated character and match the movement of the camera. For the uninitiated, it is not easy to achieve an impeccable realism in a seemingly ordinary picture. Trying to produce a high-end video with a few descriptive words was completely unthinkable before. So,When many filmmakers saw this sample film provided by Sora, they felt more shocked than words.
So, how exactly does Sora do it? Although I didn't see too many technical details revealed by the creative team. But I can try to sum it up in one word, and that is -Slack off. Yes,If there is one thing that humans are best at compared to other animals, I think it is "lazy". Because of the laziness to copy over and over again, man invented the printing press; Because he was too lazy to pass on the book, mankind invented the telegraph**; Because they were too lazy to fight with a knife and a gun, mankind invented machine guns, artillery and atomic bombs; Because they are too lazy to drive themselves, humans have invented the ...... of self-driving technologyThis laziness is not about lying on the spot, but about constantly discovering ways to free up labor and make production more efficient. And the human brain itself has a strong talent for "laziness". Think about it, when you're looking at something carefully, the content next to the object will fall into your sight. But you don't feel like you're seeing them. In fact, it's your brain that selectively ignores this information. What happens if you don't? There is such a scene in the movie "Rain Man", Raymond accidentally knocks over the toothpick box, but can instantly calculate the number of toothpicks in the box. Such "talented" just shows that the brain does not activate the "lazy" talent in time. If the human brain is in a state of "diligence and hard work" for a long time, it will soon "shut down" due to information overload, or choose to rest in some aspects that should not be rested. So, Raymond in the movie is going to be sent back to a nursing home.
The biggest difference between computers and people in the past is that computers don't choose what is important and what should be ignored. As a result, AI has a completely different perception of the world than people. Everything that is seen in the human eye comes from the projection of light on the retina. In other words, the image that humans see is two-dimensional. However, based on the understanding of the world, human beings reduce two-dimensional images to three-dimensional worlds in the brain. WhileSORA's revolutionary breakthrough is reflected in the fact that through the study of the human world, AI has begun to understand the human understanding of the world. It may seem a bit awkward to say that. In other words, AI has generated "imagination". If we say that many human imaginations still need to be described in words, words, etc. This imagination of AI can be reproduced in its entirety. With a few simple descriptions, AI can create a "world" for you. This is very similar to what Buddhism says, all things are born from the heart, and a thought blooms and a thought falls. As I wrote in a previous article, humans will eventually have two lives, one physical and the other digital. Silicon-based and carbon-based life will dominate the planet at the same time. What Sora shows us is that humans have begun to have the ability to create worlds and create digital life. And this has always been the power of "God".
In the Buddhist cosmology, it is said that with Mount Meru as the center, from the beginning of the color realm to the wind wheel under the earth, including the four continents, the sun, the moon, the six heavens of desire, and the Brahma of the color world, it is a small world. A thousand small worlds, a small thousand worlds. A thousand small worlds are one in a thousand worlds. It is a thousand worlds, covered with four meditations, nine heavens and four empty heavens, which is a great world. A big world. There are a billion small worlds. A world of 3,000 is just a world in which a Buddha has transformed sentient beings. There are countless immeasurable Buddhas in all worlds, so there are countless immeasurable worlds of three thousand worlds.
In the eyes of ordinary people, there is an insurmountable barrier between the world of human consciousness and the real world. But Sora is like a small hole drilled into this wall. Let people see,In the future, AI will be able to evolve the world that anyone can imagine. Countless small worlds together make up our "big world". Even if Einstein is reincarnated, he will be amazed by the magic of AI! I make a bold prophecy here. As the boundaries between the conscious world and the real world are broken, it is possible to help solve one of the greatest puzzles of theoretical physicsReconciling general relativity and quantum mechanics. Of course, that's not what I'd like to talk about today.
Having said all that,If there is only SORA in the world, 2024 cannot be said to be the first year of change in any way. At about the same time that OpenAI released Sora, Meta and Google also released their latest models. It's just that Sora's performance is too eye-catching and suppresses the limelight from others. But in no way that the achievements of other giants can be underestimated. For example, Google's next-generation AI model, Gemini 15. A significant breakthrough has been made in the ability to understand cross-modal ultra-long texts. gemini 1.5. The amount of information that can be stably processed is up to 1 million tokens. Equivalent to 1 hour of audio, more than 30,000 lines** or 700,000 words. For that reason alone, it has managed to crush all the large models. I would like to draw everyone's attention,It's not simply "storing" the information, it's "understanding" what it's trying to say. One is a student who can only memorize by rote, and the other is a PhD with creative problem-solving skills, and the two are completely incomparable!
So here's the problem. Some people will say, isn't it just that AI is used to make two segments**, and it can still be eaten as food? Not to mention, it can not only be eaten as a meal, but also change the pattern of the world. Let me give you three examples. Let's start with the capital market. At present, OpenAI's valuation is expected to exceed $80 billion; And shortly after the release of ChatGPT early last year, OpenAI was valued at about $29 billion. AI is inseparable from the training of massive amounts of data. Last week, Nvidia's market capitalization briefly exceeded 18 trillion US dollars, surpassing Amazon and Google in a row to become the third largest company in the world by market capitalization. (56 words have been deleted here).The direction of capital is the direction of the world's development trend.
Let's talk about it from the perspective of technology application trends. On February 18, Chen Mingyong, the founder of OPPO, issued an internal letter, in which he mentioned: "In the next five years, the impact of AI on the mobile phone industry can be compared to the replacement of feature phones by smartphones in the past." From the perspective of the development stage of the industry, AI mobile phones will also become the third stage of the mobile phone industry after feature phones and smart phones. A peek at the whole panther. If this is true in the field of mobile phones, what will happen to other fields?
Finally, let's talk about the relationship with the individual. In our time, everyone is being pushed by new technology, whether they want to accept it or not. The most obvious ones are the emerging industries of live streaming, ride-hailing and food delivery. But we can think about it, with the gradual maturity of Wensheng ** models such as SORA, Runway, and Meta, new jobs will inevitably be created. Anyone can generate content from descriptions. This content can directly generate revenue for the author. SORA bulldozes the gap in academic qualifications. A Ph.D.'s work can be a complete defeat in front of a migrant worker. The final competition is who is more imaginative. Who created the world is more popular. But only if you have at least the opportunity to use these advanced "artificial intelligence", not "artificial retardation", or "far ahead"!
As we all know, there was an important period in the history of mankind. Between about the 8th century BC and the 2nd century BC, new currents of thought were born in both the East and the West. It gave rise to the major religions and philosophies that have influenced it to this day. This period is often referred to as the "Axial Age". My view is that humanity is entering a new "Axial Age". This time it will last shorter than the last, but the impact will be more far-reaching. This time humanity will recognise the universe and will possess, at least partially, the ability to "create". Since then,Different people will live in completely different futures.
In the last "Axial Age", there was a "hundred schools of thought" in China; The ancient Indian Shamanism developed a number of religious and philosophical schools; The Kingdom of Judea inaugurated the study of religion; Ancient Greece gave birth to numerous academic schools and philosophers. In the last "Axial Age", Western civilization had Socrates, Indian civilization had Shakyamuni, and Chinese civilization had Confucius. So, what will we contribute to the world in the next "Axial Age"?
Just as I was writing this, I saw a news that OpenAI announced on social platforms that it had lifted all daily limits on GPT-4 Turbo and increased the rate limit by 1x. It can now process up to 1.5 million tpm of data per minute. This tells us againThe future has accelerated.