share_log

OpenAI祭出大杀器,Sora再度引爆!AI概念集体狂欢,全年主线又稳了?

OpenAI unleashed a giant killer, and Sora detonated again! AI concept group carnival. Has the main line stabilized again throughout the year?

Gelonghui Finance ·  Feb 18 22:34

Musk “can't sit down” anymore!

At the beginning of 2024, OpenAI's newly released text-generating video model Sora shocked the world once again.

Undoubtedly, this is another major breakthrough after ChatGPT.

Today, as the first day of the A-share market opening after the holiday, AI concept stocks set off a wave of fun across the board. Many stocks such as Huichang Communications, Chinese Online, Hanyi Co., Ltd., Danghong Technology, Wanxing Technology, and Insai Group rose and stopped.

big

Sora comes out of nowhere

As soon as Sora came out, they were shocked.

On February 15, local time, Open AI released Sora, the first Wensheng video model. It can quickly produce realistic videos up to one minute based on users' text prompts.

Open AI calls it a “world simulator.”

According to reports, Sora is capable of generating complex scenes with multiple characters, specific types of movement, and accurate details of subjects and backgrounds. The model not only understands what users ask for in prompts, but also how these things exist in the physical world.

OpenAI said that Sora is a model foundation that can understand and simulate the real world, and believes this feature will be an important milestone in achieving AGI (General Artificial Intelligence).

big

Judging from the video Sora generated, the extremely “realistic” and “imaginative” images are amazing. Among them, the highly detailed background, complex multi-angle footage, etc., made netizens call “reality doesn't exist anymore.”

Everyone says that to actually see the new industrial revolution coming, the next ten years will be the craziest.

big

Musk, who has always been at the forefront of the trend, “couldn't sit down” when he saw this.

Immediately after Sora came out, Musk posted a post on X showing some of the generated videos Tesla first launched last year.

Over the past few days, he also mentioned several times about the overall situation with OpenAI and Tesla's efforts in generating real-world video.

In one of the posts, Musk responded to the “Tesla should make a video game” and stated that he had wanted to do that for a long time.

He believes Tesla's “real-world simulation and video generation are the best in the world,” but until unsupervised fully automated driving (FSD) is introduced, the company won't be able to make games for the public.

big

Musk also said that Tesla has been able to generate real-world videos using precise physics principles for about a year, but the generated videos created by these systems aren't very interesting because the training data comes from their cars.

He said it looked like a regular Tesla video, although it was actually created with a “dynamically generated (don't remember) world.”

“Where Tesla's video generation surpasses OpenAI is that it predicts extremely accurate physical phenomena, which are critical for autonomous driving.”

In other words, what OpenAI made, he made a long time ago.

As we all know, Musk and OpenAI had controversial historical roots. Until now, Musk has also been highly critical of OpenAI's transformation from a non-profit company to a for-profit company.

big

So last year, he also launched his artificial intelligence company “xAI,” and launched his first product, the Grok Big Model, in November. Right now, though, it seems like it's still a bit difficult to catch up with ChatGPT.


AI multi-modality surges


Since ChatGPT started, big domestic models have not stopped catching up for a moment.

Since last year, big tech companies, startups, and research institutes have joined the AI model. According to incomplete statistics, more than 10 A-share listed companies have previously laid out multi-modal large-scale models or AI video-related research and development.

Among them, in the field of video generation, many A-share listed companies such as Hikvision, Jiadu Technology, Kunlun Wanwei, Dahua Co., Ltd., Fluorite Network, Wanxing Technology, Insai Group, Tolos, and Danghong Technology are already active.

Regarding Sora's influence, 360 Group founder and chairman Zhou Hongli said earlier on Weibo,Sora means that the implementation of AGI will be shortened from 10 years to two or three years.

big

Zhou Hongli said that she shared the top ten trend predictions for the big model a few years ago, and she verified several of them before the year was over.

From Gemini and Nvidia's Chat With RTX to OpenAI's release of Sora, everyone was shocked. Zhou Hongyi shared a few opinions on how to view SoraGenerally speaking, AGI will soon be realized, just a few years ago.

First,Technology competition ultimately competes for talent density and deep accumulation.
Second,AI may not disrupt all industries that fast, but it can inspire more people to be creative.
Third,I've always said that on the surface, the level of development of large domestic models is close to GPT-3.5, but in reality there is still a gap of one and a half years from 4.0. And I'm sure OpenAL still has some secret weapons in its hands.
Fourth,The best thing about the big language model is that it's not a fill-in-the-blank machine, but rather a complete understanding of the world's knowledge.

Guohai Securities believes thatSora showed that the big model evolved faster than expected.

As of December 2023, the number of ChatGPT users is 180 million, the average number of monthly visits to the website is 1.7 billion, and OpenAI's valuation exceeds 100 billion US dollars. The release of Sora will open up the price and valuation space for OpenAI users.

The CITIC Securities Research Report pointed out that if Sora becomes a super application in the field of video creation, it will bring huge demand for network equipment upgrades. Meanwhile, Google's latest multi-modal model Gemini 1.5 and Nvidia's local chatbot Chat with RTX have also been released one after another recently.

The development trend of AI has become more clear, and the AI industry's mature business model and a virtuous cycle of “revenue - capital expenditure” are gradually being formed. This is expected to ease the market's concerns about the continued investment in computing power equipment such as optical modules, and thusOpen up valuation space for leading manufacturers.

Disclaimer: This content is for informational and educational purposes only and does not constitute a recommendation or endorsement of any specific investment or investment strategy. Read more
    Write a comment