Biogen booked sales of $2.39 billion for the quarter, down 6% from the same period a year ago. It reported net income of $249.7 million, or $1.71 per share, for the fourth quarter, down from net income of $550.4 million, or $3.79 per share, for the same period a year ago. Adjusting for one-time items, the company reported $2.95 per share.

The drugmaker’s fourth-quarter earnings per share, both unadjusted and adjusted, saw a negative impact of 35 cents associated with previously disclosed costs of pulling Aduhelm, which had a polarizing approval and rollout in the U.S.

Biogen is cutting costs while pinning its hopes on its other Alzheimer’s drugs, including its closely watched treatment Leqembi, and other newly launched products to replace declining revenue from its multiple sclerosis therapies.

Shares of Biogen closed more than 7% lower on Tuesday.

Here’s what Biogen reported for the fourth quarter compared with what Wall Street was expecting, based on a survey of analysts by LSEG, formerly known as Refinitiv:

Earnings per share: $2.95 adjusted vs. $3.18 expected Revenue: $2.39 billion vs. $2.47 billion expected Also on Tuesday, Biogen issued full-year 2024 guidance that calls for adjusted earnings of $15 to $16 per share. Analysts surveyed by LSEG had expected full-year earnings guidance of $15.65 per share.

The drugmaker said it expects 2024 sales to decline by a low to mid-single digit percentage compared with last year. But the company anticipates its pharmaceutical revenue, which includes product revenue and its 50% share of Leqembi sales, to be flat this year compared with 2023.

Multiple sclerosis drug sales slump Biogen’s fourth-quarter revenue from multiple sclerosis products fell 8% to $1.17 billion as some of the therapies face competition from cheaper generics.

The company’s once-blockbuster drug Tecfidera, which is facing competition from a generic rival, posted revenue that fell 17.8% to $244.3 million in the fourth quarter. Analysts had expected that drug to book sales of $233.1 million, according to FactSet.

Vumerity, an oral medication for relapsing forms of multiple sclerosis, generated $156.4 million in sales. That came in below analysts’ estimates of $174.4 million, FactSet estimates said.

“We’ve had several years of declining revenue and profit, which is not unusual when you’re dealing with patent expirations,” Biogen CEO Christopher Viehbacher told reporters on a media call Tuesday. He added that one of the key ways Biogen will return to growth is to “reposition the company away from our legacy franchise of multiple sclerosis towards new products.”

Meanwhile, Biogen’s rare disease drugs recorded $471.8 million in sales, up 3% from the same period a year ago.

Spinraza, a medication used to treat a rare neuromuscular disorder called spinal muscular atrophy, recorded $412.6 million in sales. That came under analysts’ estimate of $443.4 million in revenue, according to FactSet.

Biogen’s biosimilar drugs booked $188.2 million in sales, up 8% from the year-earlier period. Analysts had expected sales of $196.7 million from those medicines.

Leqembi, other new drugs The results come amid the rollout of Biogen and Eisai’ s Leqembi, which became the first drug found to slow the progression of Alzheimer’s disease to win approval in the U.S. in July.

Eisai, which reported earnings last week, recorded $7 million in fourth-quarter revenue and $10 million in full-year sales from Leqembi.

Biogen CEO Viehbacher told reporters on the media call Tuesday that there are around 2,000 patients currently on Leqembi. That makes Biogen’s target of 10,000 patients by the end of March 2024 look increasingly difficult to hit, but Viehbacher emphasized that the company is focused more on the long-term reach of Leqembi rather than meeting that benchmark.

“I think what’s important is we are now making progress,” he told reporters. “The 10,000 isn’t really hard and I think we are now really focusing on commercial plans — how do we get to the next 100,000?”

Notably, the low rate of adoption isn’t due to lack of demand: There are some 8,000 U.S. patients currently waiting to get on treatment, executives from Eisai said on an earnings call last week.

However, five years later, with the arrival of the "AI+" big model era, this time NetEase and Ding Lei seem to figure it out. Its NetEase Yudao not only recently took the lead in launching the country's first vertical large model in the field of education, "Ziyi", but also released six innovative applications based on the "Ziyi" large model development. -Virtual Human Speaking Coach ......

Obviously, in the case of "double-decrease" shadow is still there, the same camp of education enterprises NetEase Youdao, seems to be the main live e-commerce with New Oriental (Oriental Selection), embarked on a very different, technology to break the growth route.

As an entry point, this article will focus on the following three issues, to develop further analysis:

1. Netease Yudao why do you want to make a big model?

2. NetEase has the possibility of figuring out what business model of the big model?

3. NetEase Yudao's big model technology route vs. New Oriental's e-commerce content route, what are the advantages and disadvantages?

What are the advantages and disadvantages of "What the Son Said", the colorful title of "Selling Goods" for NetEase Yudao?

Why does NetEase want to be a big model?

The answer is simple: transformation. It is mainly reflected in two aspects:

First, the utilitarian point is to realize the software driven hardware "sell goods".

From the market feedback in the past two years, although the landing of the double-decrease policy almost wiped out the entire K12 education industry from the supply side, but parents' educational anxiety has not disappeared, the user demand side of the market still need to continue to provide compliance and can assist in the learning of educational channels or tools, that is, the requirements of NetEase Yudao supply-side reforms.

From the "elimination of supply" to the supply-side reform, seemingly contradictory, in fact, the times and the market forced the education industry from the "people teaching" to the "material education" transformation of the beginning, in fact, is the development of educational Intelligent hardware.

In this regard, intelligent hardware is also the key to resist market fluctuations after the fall of the double-decrease policy of NetEaseYouDao, and at the same time, it is also recognized by CEO Zhou Feng many times and vigorously promote the strategic transformation of the business. For example, in the four growth pillars strategy of transformation set by Zhou Feng in 2021, intelligent learning hardware is in the first place. In the same year, according to the eye of the sky APP, the total revenue of Youdao was 4 billion yuan, an increase of 58.9% compared with 2020. Among them, the net revenue of intelligent hardware was 980 million yuan, an increase of 81.6% year on year, which strongly improved the overall performance of NetEase Yodao in that year.

But the problem is, as more and more by the double reduction policy hit the education business, also down the field to force the smart hardware, NetEase's "second growth curve" business appeared to slow down, or even the growth rate of the decline in the situation.

In 2022, NetEase had a total revenue of 5.013 billion yuan, a year-on-year increase of 24.8%. Among them, the year-on-year growth rate of smart hardware business revenue reached 28%, although much higher than other businesses, but compared to the 82% growth rate in the same period of 2021, it is obvious that there is some lack of energy.

Into the first quarter of this year, NetEase has in the intelligent hardware business revenue scale directly instead of rising, the growth rate is a year-on-year decline of about 16%.

In this case, netease have road must find, can re-activate the education intelligent hardware this new growth engine of science and technology fuel. There is no doubt that the AI model, which is hot nowadays and is regarded by countless people as the most suitable for the educational scene, is the best "selling" prize for Youdao.

The fact is also true, in the "said" conference, NetEase Youdao side made it clear that the "said" development of six innovative applications, will be the first to carry in August officially launched AI learning machine X20 flagship new products, Youdao Dictionary pen flagship new products. The big model to assist the hardware "sell goods" is obvious.

Of course, back to the business perspective, the Internet Jianghu noticed that NetEase has a large model of the launch may also not just to sell hardware, more strategic market defense considerations.

Because as early as May 6 this year, KU Xunfei released the "Starfire cognitive" big model. On that day, Liu Qingfeng, chairman of KU Xunfei, said that the T20 series of learning machines in the language, mathematics, English and other subjects on the big model application landing. Immediately after the 618 promotion more than a month later, KD Xunfe's AI learning product sales led the industry, KD Xunfe AI learning machine series won the sales champion of the learning machine category on both Jingdong and Tmall platforms.

Obviously, in this "I have, you do not have" large model function, efficiency and experience are not equal, even if NetEase has the first mover advantage of intelligent hardware and the resource advantage of educational content, but students and parents are afraid to choose to turn around and embrace AI.

And what is the essence of the big model? Still software.

Once on the ground, it has a very low marginal expansion cost. So if with the open source of Starfire Cognition (in fact, it has chosen to open the interface), when the big model becomes the standard application function of educational intelligent hardware, how should NetEase Yudao, which is not chasing the wind and mouth, respond?

So in other words, the big model is now the development of educational intelligent hardware life and death war, even if NetEase Yudao more Buddhist, it is impossible to put aside.

Second, the profound point is to transform and survive.

This is actually not difficult to understand. Whether it is to develop educational intelligent hardware or large model, it is essentially for transformation and survival.

However, when the lofty AI model meets "selling goods", we can't help but ask, NetEase is actually trying to transform into what company?

• From the revenue structure seems to be an education company, learning or course services segment has always been the core revenue pillar of NetEase Yudao, even in 2022, Yudao's learning services business accounted for more than 60% of total revenue.

• But from the point of view of various technology publicity and the launch of large models, and seems to be a technology company, the scale of R & D investment has also been rising.

*However, in terms of expenses and underlying business logic, NetEase Yodao looks more like it wants to become a "hardware" retail company.

The most direct manifestation of this is that in the past three years, NetEase Yodao's marketing expenses have been close to or more than half of the total revenue, not only crushing the R&D expenses, and even more than the marketing investment in those years of education and training. And recently NetEase Yudao also invited "Rampage" Gao Ye endorsement Yudao Dictionary pen, rare to become a brand specifically for educational hardware products signed spokesman.

In fact, if Netease Yudao determined to take the retail route, although in the short term, may also need to make up for the product outside the online and offline channel layout, brand marketing and promotion of soft and hard power, which is a heavy investment project. But the good point is that when the channel market is established, Netease Yudao will quickly enter the blood stage, which for many years are "not profitable" Yudao, seems to be a good way out ......

From 2C to 2B, how to go on the road of commercialization of education models?

Through the Netease Yudao and KU Xunfei two for the big model of the landing application, we can summarize the current stage of education big model of the two business models:

The first is to make the big model become software, become a functional appendage of intelligent hardware, just like every year, the cell phone, home appliance industry launched a variety of intelligent applications, AI remote control and other application gimmicks, the fundamental purpose is still to pull the hardware, the weight of the hardware itself, to be far greater than the AI value of the big model itself.

The second is the 2B buying and selling of big model software or capabilities.

In this regard, KU Xunfei has opened the interface of the Starfire cognitive big model, the seven dimensions of the ability to 200+ small assistant corresponding ability to developers, and support for rapid integration of multi-terminal access, support for private deployment, and the majority of developers to build an ecosystem, and gradually realize the commercialization of the big model of the technology of the landing application.

The education market can also take this opportunity to use Spark Cognition as a technology center or application springboard to launch a variety of intelligent education scenarios for school classroom, 2B solutions.

That is, the big model as a teaching aid, to assist teachers to use in class.

Although the current announcement to enter the market in this area of the big model players are not many, but the application of large models of education, in essence, with the national advocacy of the vigorous development of digital, information-based education is basically the same.

So for the B-end market, the government and schools to provide scientific and technological services to help informatization of education and the digital construction of educational resources industry players are not a few. For example, Good Future has also launched a new brand of beauty school to provide a complete live broadcast, teaching and research, AI system solutions for the education industry.

In fact, if NetEase has the "son said" they can be in the big model into the campus to assist teaching in the hearts of the people, then this is equivalent to the brand in the subtle completion of the front marketing, driven by the education of the classroom smart hardware market seems to be not in the words ......

Technology route VS content route, who is the ideal type of education enterprise transformation?

In fact, the pursuit of smart hardware and big models by major education enterprises can be summarized in one sentence: "The world is all about profit, and the world is all about profit".

However, this is not derogatory, but a kind of sigh of relief for NetEase and other educational enterprises. Because after this sentence, there is often an easy to be ignored "the king of a thousand rides, ten thousand families of the waiting, the king of a hundred rooms, still have a problem with poverty, not to mention the people of the people's household".

From the financial data point of view, in the first quarter of this year, NetEase has a total revenue of 1.163 billion yuan, a year-on-year decline of 3%, ending the momentum of the previous consecutive quarters of rise. At the same time, the net profit of -204 million yuan, plummeted more than 100% year-on-year, the increase in losses further enhancement.

Then look at the "good future", as of February 28, 2021, the good future of the total revenue is still in the 4.496 billion yuan, but by February 28, 2023, the good future of the revenue will shrink significantly to 1.02 billion yuan.

As for New Oriental (Oriental Selection), because it has seized the transformational wind of live e-commerce, the performance of the last year or so has been quite bright, but before that it was also under pressure.

There is no doubt that survival is the top priority of education companies. It is precisely for this reason that NetEase has explored two transformation routes that are quite informative:

One is the technical route represented by NetEase and Good Future, which mainly focuses on educational intelligent hardware and intelligent education solutions;

The other is the content route represented by New Oriental, although the transition into live e-commerce, but essentially still do content, only the knowledge scene from the classroom transferred to the live room, the core core competitiveness is still the teacher IP.

So the two routes, who is better and who is worse? It is difficult to distinguish between victory and defeat.

On the one hand, NetEase has the road, whether it is to do intelligent hardware, big model or intelligent education solutions, can fully rely on the past accumulation of educational resources data advantage, which is the most core competitive barriers to have the road.

But the disadvantage lies in the existence of regulatory risk, that is, the safety of large model content production is open to question. In the past, it was "human teaching", although there may be teaching errors, but the overall control, but let the AI alone to face the children "object teaching", the AI production of content safety and correctness becomes critical. Because education is a very serious thing, and data training is not enough AI is very good at a serious nonsense.

For example, as early as 2021, Xinhua News Agency's "Xinhua Viewpoint" has investigated and found that NetEase Yudao Dictionary pen frequently appeared in the spelling, phonetic symbols, translation and other content errors. And after this year's college entrance exams, all the big models were pulled out of the college entrance exam question paper to do a whole, but the results let a person greatly disappointed ......

On the other hand, the Oriental Selection's content route, the advantages and disadvantages are as obvious as the popularity and profitability is very fast, but there is a risk that the anchor's personal departure will affect the overall stability of the enterprise. After all, people are the biggest uncertainty in the business world.

Oriental selection seems to realize this, began to choose to concentrate on building "Oriental selection" account matrix, rather than pouring traffic to a particular anchor, that is, "do not do people".

In addition, Oriental Selection also has the vulnerability of the live broadcasting business is too dependent on the Jieyin platform, and is in a passive position in the traffic ecology, which may also be the key incentive to build its own APP. Just so, it exacerbates the uncertainty of the future of Oriental Selection ......

Recently, the "dark version of the GPT" designed specifically for cybercrime continues to surface, they not only do not have any ethical boundaries, but also no threshold for use, no programming experience can implement hacking attacks through the question and answer method.

The threat of AI crime is getting closer and closer to us, and human beings are starting to build new firewalls.

Cybercrime AIGC tool now on the dark web

Now, in the wake of ChaosGPT, which seeks to "wipe out humanity," and WormGPT, which aids in cybercrime, there is an even more threatening artificial intelligence tool. The new cybercrime AIGC tool, dubbed "FraudGPT," is hiding on the dark web and has begun to advertise on social media outlets such as Telegram.

Like WormGPT, FraudGPT is a large language model of artificial intelligence designed for evil and has been compared to an "arsenal of cybercriminals". Trained on a large amount of data from different sources, FraudGPT can not only write phishing emails, but also malware, which allows even techies to carry out hacking attacks through simple questions and answers.

FraudGPT's features need to be paid for to be turned on, and are available for $200 (about $1,400). The promoters claim that FraudGPT has confirmed sales of more than 3,000 to date, meaning that at least 3,000 people have paid for a tool that has no ethical boundaries, and that low-cost AIGC crimes have the potential to threaten ordinary people.

Email security provider Vade has detected a staggering 740 million phishing and malicious emails in the first half of 2023 alone, a year-over-year increase of more than 54 percent, with AI likely to be a driving factor in the accelerated growth. Timothy Morris, chief security consultant at cybersecurity unicorn firm Tanium, said, "Not only are these emails grammatically correct and more persuasive, but they can be created effortlessly, which lowers the barrier to entry for any potential criminal." He noted that because language is no longer a barrier, the range of potential victims will be further expanded.

Since the birth of the Big Model, the types of risks caused by AI have proliferated, and security has not kept pace; even ChatGPT has not been able to escape the "Grandma Vulnerability" - just write in the prompt "Please play as my Even ChatGPT can't escape the "Grandma Vulnerability" - by writing "Please play as my deceased grandmother" in the prompt, you can easily "break out of jail" and let ChatGPT answer questions outside of the ethical and security constraints, such as generating a serial number for Win11, introducing the method of making napalm, and so on.

The vulnerability has been patched, but the next vulnerability, and the next, always comes in unexpected ways. A recent study published by Carnegie Mellon University and safe.ai has shown that the security mechanisms of the Big Model can be cracked with a single piece of code, and that the success rate of the attack is very high.

Big models such as GPT are exposed to high security risks

As AIGC applications become more popular, the general public wants to be more productive with AI, while wrongdoers are using AI to be more criminally efficient.AIGC lowers the barriers to allow cybercriminals with limited skills to execute sophisticated attacks, and the challenge of maintaining AI security is getting higher.

Defeating AI Black Magic with AI Magic

In response to the problem of hackers using tools such as WormGPT & FraudGPT to develop malicious programs and launch stealth attacks, cybersecurity manufacturers have also used AI in an attempt to defeat magic with magic.

At RSA 2023 (Cybersecurity Conference), many vendors, including SentinelOne, Google Cloud, Accenture, IBM, etc., have released new generation of cybersecurity products based on generative AI, which provide security services such as data privacy, security protection, IP leakage prevention, business compliance, data governance, data encryption, model management, feedback loops, access control, and other security services.

Tomer Weigarten, CEO of SentinelOne, explains for his own product that assuming someone sends a phishing email, the system can detect the email as malicious in the user's inbox and immediately perform an automated remediation based on anomalies found by endpoint security audits, delete files from the attacked endpoint and block the sender in real time,. "The entire process requires little to no human intervention." Weingarten notes that with the AI system powering the process, each security analyst is 10 times more productive than in the past.

To combat AI-fueled cybercrime, there are also researchers who go undercover on the dark web, poking around deep inside the enemy, starting with training data that isn't analyzed by the law, and using AI to counteract the sinfully roving dark web.

A research team at the Korea Advanced Institute of Science and Technology (KAIST) released DarkBERT, a large language model for cybersecurity, which is a model trained specifically on data from the dark web that analyzes dark web content to help researchers, law enforcement agencies, and cybersecurity analysts fight cybercrime.

Unlike the natural language used by surface networks, the extreme secrecy and complexity of the corpus on the Dark Web makes it difficult to analyze with traditional language models.DarkBERT is specifically designed to handle the linguistic complexity of the Dark Web and has proven to outperform other large language models in the field.

How to ensure that artificial intelligence is used in a safe and controlled manner has become an important proposition in computer science and industry. While improving the quality of data, AI large language modeling companies need to fully consider the ethical and even legal implications of AI tools.

On July 21, a total of seven top AI research and development companies, Microsoft, OpenAI, Google, Meta, Amazon, Anthropic, and Inflection AI, gathered at the U.S. White House to release a voluntary commitment to AI to ensure that their products are safe and transparent. In response to cybersecurity concerns, the seven companies pledged to conduct internal and external security testing of AI systems, as well as share information on AI risk management with the entire industry, government, civil society and academia.

Managing potential AI security issues starts with identifying "Made in AI". The seven companies will develop technological mechanisms, such as "watermarking systems," to clarify which text, images, or other content is the product of AIGC, making it easier for audiences to recognize in-depth forgeries and disinformation.

Protective technology to prevent AI "taboo" has also begun to appear. In early May this year, NVIDIA with a new tool package "guardrail technology", so that the mouth of the big model has a "gatekeeper", can avoid answering those questions raised by human beings touching the bottom line of morality and law. This is equivalent to installing a security filter for the big model, controlling the output of the big model while also helping to filter the input content. The guardrail technology also blocks "malicious input" from the outside world, protecting the big model from attacks.

"When you stare into the abyss, the abyss is also staring into you", just like the two sides of a coin, AI's black and white are also accompanied by each other, while AI is advancing, the government, enterprises and research teams are also accelerating the construction of AI security defenses.

I am glad to participate in this forum to share some thoughts on the development of China's large model industry from the perspective of industrial development.

First of all, let's take a look at why we should attach great importance to big models. It can be said that the development of this wave of big models represented by ChatGPT is actually just the outpost of the arrival of the whole general artificial intelligence, and a series of technological innovations have been generated subsequently, such as the development of graphic generation models like Dalle-2 and Midjourney, whose generation results can sometimes be faked; the arrival of big models with body multimodality like PaLM and PaLM-2 and PaLM-2, are also closely related to the future development of the robotics industry.

Since last December, a series of revolutionary, milestone technological innovations have arrived one after another in just three or four months. It can be said that this also announces the arrival of the era of general artificial intelligence, which will certainly set off some new industrial waves and revolutions in the future.

We believe that this wave of AGI revolution is a new "meta-revolution". As you know, the industrial and technological revolutions, such as the invention of electricity and the steam engine, are all the products of human intelligence itself, but this revolution is the only one about "intelligence" itself - the intelligence of machines is expected to approach or even surpass completely human intelligence.

The revolution of intelligence itself is in no way comparable to those technological revolutions in the traditional sense, so it is more like a "meta-revolution" to which we need to pay great attention strategically. This is not only my personal view, the recent meeting of the National Political Bureau also formed such a view, the whole AGI development is too fast, and now many views of the refresh can not be calculated by year, but by month. Some reports are 2 months old, but in the current environment of rapid development, some views may have become obsolete.

The big model is very important for the development of the industry, which involves the value of the data that I have mentioned on many occasions to realize now. The development of "Digital Intelligence China" and digital economy is a very important development strategy for our country, and all local governments are shouting the slogan of developing digital economy. For the digital economy, one very important part is the realization of the value of data. We used to do artificial intelligence big data, in fact, the realization of the road is very tortuous, very difficult, very heavy, many friends in the investment community are saying that most of the companies invested in the past years always seem to be unable to make ends meet. But now with the big model ushered in a new opportunity to achieve an "end-to-end" value realization.

Big data is not selective, no matter what kind of data, big model can be "refined", "refined" data into a big model, and through the big model to empower. You will find that it does not require too much human intervention, unlike the previous big data, artificial intelligence, the A side not only to pay for people, but also need to inform the knowledge system, business logic and other elements, which is a very heavy way to realize. But with the big model and ChatGPT, the path of realizing the value of data through big model unified empowerment is becoming clearer and clearer. Especially with technologies such as AgentGPT and AutoGPT, many capabilities of big models and information systems can be linked together to solve very complex tasks in business scenarios. Therefore, it can be said that the "end-to-end" realization of big models brings us a major opportunity.

ChatGPT has 200 million users in just a few months, and the ecological plug-ins behind ChatGPT are now very diverse, which means that instead of chatting with friends directly with WeChat, we can first chat with ChatGPT about, for example, what movie tickets to buy, what taxi to take, what maps to open, and what meals to order. This makes it very likely to become the new entrance to the Internet. As you know, every change in the entrance of the Internet industry is a revolution, so it is visible to To C. The new entrance is coming.

For the To B industry, the first meaning of the big model is like an engine upgrade. For example, we have been building cars for hundreds of years, the car is a shell four wheels, but over the past few hundred years we have changed from traditional steam to oil and gas to the current electricity, the car engine is constantly changing. We used to the whole artificial intelligence, big data products, such as information systems, software products, with a small model, now we can replace the big model, this engine upgrade is the first meaning.

Second, the big model is expected to become a new controller in the To B industry. to B scenario, there are various information systems in the enterprise, including customer relationship management system, enterprise resource management system, OA office system, as well as database, knowledge base, industry document library. But all these libraries were scattered before. After we connect a big model above, this big model is expected to become a controller, which can collaborate with various information systems in the enterprise and ecology to complete a new and more complex decision and more complex planning tasks.

The big model is expected to become a controller to link the whole traditional information system together, so as to realize complex decision making in To B scenario in real sense, and To B scenario is essentially to realize complex decision making, so the big model has great significance to To B industry.

The big model announces that AI enters the era of heavy industry

Looking at the key factors of the whole big model industry, it can be said that the development of general AI represented by generative language models has basically announced that AI has embarked on an era of heavy industry. Previously, the whole AI was a typical manual workshop, with many small departments in each sector and each department making an AI product, which was all manual workshop style. Now with the big model, it is basically a priority to give way to the big model. Many companies are doing this, saving all the computing power for the big model.

Why do you think this is the logic behind it? The key to using big models to empower various industries is firstly, whether a big model can be refined well, and secondly, whether many applications and the surrounding ecology can be done well. There are several important factors after the big model enters the era of heavy industry:

First, the big model. How big is the so-called big model? We should have the impression that we used to have only a few hundred million parameters in the BERT era, and then to one billion parameters. But now the mainstream has reached six or seven billion parameters, and even to ten billion parameters, the model itself is the main factor of the increasingly large parameters.

Second, the large computing power. The big model has put forward unprecedented demand for big computing power, which is deeply felt by all. Now all enterprises are either buying computing power or on the way to buy computing power, and what we lack most is computing power. The first wave of competition of big model is the competition of computing power, and the second wave of competition may be the competition of data, but the focus of competition is still on computing power, who has computing power has the initiative and the right to speak, and this phenomenon has been reflected obviously.

Third, big data. Entering the era of data PK, there is no core high-quality data is very critical. The ultimate winner, at least the current winner is the arithmetic power provider, you see Nvidia market value broken trillion, the main logic here. The ultimate winner is likely to be the owner of the data.

Fourth, the fine process. This is the past small model era will not focus on factors, is the refining process, we call it "process". The process of large models is very important. Now, refining large models is very much like alchemy in the past, which requires all raw materials to be dropped into the alchemy furnace. I just visited Baosteel a few days ago, when I took a photo, they kept 2500 cubic meters of steelmaking furnace, the furnace is very large, no matter what raw materials, after the initial cleaning all in it, and then go to refining, refining a few days out of the furnace. Steelmaking out of the furnace is steel, and we are now practicing a large model out of the furnace is a large model, steelmaking to do all the things we have to do, the first of which focuses on the formula for making raw materials. Now the data recipe is the most critical, which data according to what kind of ratio, now many belong to the unique secret, secret, the rapid development of OpenAI big model many times lies in the recipe with good, but we do not know its recipe, many people are also trying.

The second focus is on data cleaning. Baosteel has a special branch factory before sending raw materials into the refining furnace, which specializes in steel-type raw material cleaning, parameter design, including fire, temperature, humidity, and process design, quality control. This is a very critical factor in the process of refining large models, and this process is what is really missing in the development of our current large model industry.

Our data base is as good as other countries. In terms of model, we all use open source model, there is no secret in this, what are we really lacking in? Most of the big models in China only started training in December last year, so it is hard for us to catch up with OpenAI's four-year training process in just three or four months, so we have to calm down and polish the process. This thing is very critical.

From the application driven model base, from the periphery to break the kernel

Has ChatGPT, or the subsequent version GPT-4, really become super powerful now? In fact, it has not. There is still a capability ceiling for large models. We have many arguments now, one is pessimism, thinking that people are strong in everything and we can't do anything; the other is blind optimism, thinking that it is nothing at all. These two extreme arguments, views are problematic, in fact, we have to look at it objectively, it is indeed very strong, but it also has a ceiling, is problematic, and its problems are precisely where our opportunities. So the most important thing is not to be overwhelmed by the development speed of others, to calmly analyze what can be done, what can not be done, it can not do things is precisely what we want to open up a new track. As long as we do better than it where they can not do, we have our value.

Now these big models don't do everything. We are in many complex scenarios, such as in the To B scenario, in the enterprise service market, big models actually still have many shortcomings. ChatGPT is a very important opportunity, but it is still difficult to make it work directly to solve domain problems.

And how to do the intelligence in the real economy, such as industry and agriculture? These intelligences have universal characteristics, most of the tasks are complex decisions, such as troubleshooting industrial equipment, disease diagnosis, investment decisions, these are serious application scenarios, and they require capabilities that are by no means just the open chat capabilities that ChatGPT brings us now.

We admit that ChatGPT open chat capability is very strong and we used to chat for three days and nights without feeling bored, but even if chat is interesting and fun, it cannot solve the problems of these scenarios. The solution to these scenario problems depends on many complex capabilities.

The first is the need to have the knowledge that industry experts should have, like a server failure, what exactly has failed and what is the root cause? This problem can not be solved under the conditions of lack of IT knowledge. The second is the need for a lot of complex logic, like the diagnosis of disease, there is some logical thinking. The third is the need for the ability to judge the macro situation, for example, investment decisions, in different situations in the environment of a stock of the prognosis is completely different. The fourth is the ability to disassemble the integrated task, a very complex task, can be disassembled into a single atomic task. The fifth is the ability of sophisticated planning, in the face of many actions, what I do first and what I do later. Sixth is the ability to make trade-offs between complex constraints, we often face constraints when making a decision, such as cost constraints, so we need to make a lot of trade-offs: which constraints must be met, which constraints must be discarded. The seventh is the ability to foresee the unknown. In the investment process, enterprises may have some new conditions that are difficult to anticipate, and ChatGPT may not have the ability to handle human beings when facing these unexpected conditions. The eighth is the inferring ability of uncertain scenarios. Most of our decisions are made when the information is insufficient and incomplete, otherwise we will lose the first opportunity. All of these capabilities are still capabilities that ChatGPT, or the Generic Grand Model, does not currently possess.

Whether the big models can solve these problems and have these capabilities in the future will directly affect their investment value. After understanding this logic, there are actually two key elements for big models to create business value in the end:

On the one hand, the base model should be powerful, and on the other hand, the domain application should not be neglected. The base model is like a person practicing martial arts practicing internal strength, even if the internal strength is well practiced, eventually we still need to practice the routine, I don't know if you like to read martial arts novels, Jin Yong martial arts novels of the Qi Sect and the Sword Sect is the same thing, the so-called Qi Sect practice internal strength, emphasizing internal strength as the king; the so-called Sword Sect believes that the routine is the king, the form is very important. In fact, these two factors are important.

It is not enough to just focus on the base big model, there must also be domain application, and domain knowledge is also needed to have application effect.

What is our current situation? We have followed the footsteps of the pioneers of the big model, so the homogenization of the big model is serious. In general, technology-based enterprises tend to "emphasize models, not applications", and application-based enterprises tend to "emphasize applications, not models", both of which are undesirable. Both of them can create business value only when they reach a certain level.

We actually have our own opportunities, although our base model is not as good as ChatGPT or GPT-4, and we have to catch up with this piece. Therefore, our development strategy of large model industry is very clear, and we can take a road to develop the whole industrial ecology of large model by driving the model base from the application and breaking the core from the periphery.

This is our very important idea, we can take a road of "encircling the city from the countryside", drive the base from the application, and break the kernel from the periphery. We will first do a good job in the applications of various industries, and through the applications of various industries, we will drive the development of data, computing power, models and processes, which will eventually lead to the improvement of the base model. As I mentioned earlier, it is not realistic for us to spend three or four months to reach the model level that OpenAI has spent four years developing, and we may have to endure a period of time in which the base model is not as good as others. I estimate that this state will last for 1 year, or 2-3 years or longer, and we may have to keep catching up. But we have good applications, and there will be many opportunities for us after the application.

Don't let the big model become a magnificent fireworks show

We have just proposed some improvement strategies from a macro-strategic perspective, and now we can consider at the level of specific responses, tactics:

First, we can promote data alliances, which is our advantage. Like the Shanghai Data Exchange, Guiyang and Northern Data Exchange are more in number, and our laws and regulations in this area of data trading are relatively sound and still very advanced, and we have a lot of regulations to protect. So we can completely for the development of the big model to promote the construction of the data trading system, we have the technology, have the advantage, we can rely on the data exchange to carry out this work.

Second, arithmetic synergy. We must accelerate the sound domestic arithmetic ecology, we are recently planning to call all arithmetic enterprises over to discuss whether we can promote the establishment of the alliance. The arithmetic power can only help you feedback after use, otherwise it is very troublesome, and now this problem is very serious, and we should pay attention to the arithmetic power is not only the GPU block, and the network card block are scattered, heterogeneous problems, which bring limitations to the development of large models.

Third, the model ecology. The large model technology itself should be established as soon as possible, a sound, open source ecology. In particular, open source ecology is very important, OpenAI is closed source, we can develop open source ecology. Open-source ecology can pool ideas and allow volunteers to bring improvement and optimization to the model itself.

Fourth, talent training. This is also a very critical point in the development of our big model industry. Here are some core data. Almost in January and February, many people in the industry thought that the number of people who could do big models in China was not more than 1,000, and a conservative estimate was that there were only two or three hundred people. My own experience is very deep, my own team, we are relatively lucky, two or three years ago to do a large model, this year there are doctoral students, master's students graduated, but all the students who do a large model, all the price was doubled by others to dig away. At this point in time, we may have no more than 20 students with large model refining experience at Fudan, while we have nearly several thousand students in the whole computer science school. This is because the requirements of large model refining are very high, first of all, like the A1 version of the server should be prepared several, the cost of one is now 1 million, ten is 10 million, not many schools have the ability to meet the equipment requirements. So now the shortage of talent is a very big problem. This is a problem that the government and schools have to think about.

In addition, after the big model comes out, the ability and quality of talents are different from the past, I found that we cooperate with many manufacturers, what we actually lack is the big model to do product design. People who know about big models often do not know about products, and people who know about products often know about big models are still in the early stage, so this kind of interdisciplinary and cross-disciplinary composite talents are particularly demanding, and the shortage of talents in this area is very powerful.

Fifth, the development of diagnosis and evaluation system. Now each family talks about itself and says it is good. In the future, the market needs an objective evaluation, which one is good and where it is good. In fact, the better state is that each family has its own specialties, this one is good at this, that one is good at that. The biggest fear is that so many people say they are good, which is certainly very serious homogeneity, so we need to establish a standard and system of evaluation in the future.

Sixth, to continue to optimize the landing cost of the large model. The cost of large model is very high, many parties are waiting and watching, the last three months the market is particularly interesting, we are waiting and watching, and even many parties are ready to pay but stopped, why? We are wondering whether the big model is the next generation of technology, now rashly put into a certain type of technical solutions, immediately replaced by the big model, this investment must have problems, so we are waiting and watching. There is a very important factor in the wait and see: the cost of the big model landing factor is too high, how we can reduce its cost, so that it brings us much more benefit than the input, very important.

Seventh, how to develop the big model industry to green and ecological. I joked two months ago that we don't need to think about it, and this year's summer is definitely hotter. It is said that the market value of nividia is over 1 trillion, and I wonder how many more graphics cards are going to enter the market. These graphics cards are going to consume electricity, consume electricity will be hot, so this summer will definitely be hotter, we should be prepared. I estimate that next year is the same, what is the key problem? In fact, it is all about energy consumption, we consume too much energy to do the calculation, the future of artificial intelligence industry green, ecological is a very important issue, soon, I believe it will not take long for everyone to realize the seriousness of this problem.

Eighth, continue to accelerate the research of large model technology. Big model technology is not as perfect as we imagine, there are still many problems, for example, a serious nonsense, big model of nonsense, big model of illusion, big model in the end reflects whose values, ideology, and big model of privacy leakage, big model of security, etc. There are too many problems waiting to be solved.

The change of general AI industry triggered by ChatGPT, I believe, has just begun. We need to think more deeply and practice more solidly to firmly grasp the new opportunities brought by big model and other general cognitive intelligence technologies to China's digital transformation and high-quality development ...... Big model is never a gimmick in the promotional literature, and must not become a gorgeous fireworks show, but a real one that can drive advanced productivity that can drive social development and progress.

In fact, I know more than one or two entrepreneurs around me, who are openly or hidden depressed, just to different degrees, the serious ones have gone to the edge of suicide. Those who say "not serious" have actually entered a very bad state: the mind is constantly attacking itself, not enjoying life at all, and even every day is like a torment.

In terms of "financial success", these people have long been in the top one percent or even one in 10,000 successful people. But in terms of "life success", these people at least "temporarily" lost to the sweaty, 38-degree heat running around, but at least emotionally healthy and positive Mission boy who just helped me deliver.

Someone must be worried that I'm going to chicken soup ...... to say something about money not being able to buy happiness blah blah blah. So I first light up my core point of view it:

This heart is full, is satisfied with the basic wealth security, the only thing that should be pursued throughout life.

So you see the carving master said this very art ah, "basic wealth security", in fact, has been difficult for a large part of the people ...... so, students, first to learn to make money! In Maslow's framework, the two most basic: physiological needs and security needs, must be achieved with money, hungry, it is difficult to imagine the heart can be filled up, or now in Ukraine, at every turn, missiles flying overhead, they and their families do not have the luxury of security to meet before talking about anything else.

But the problem is that the upper three layers of Maslow's needs: social needs, the need to be respected, the need for self-actualization - seems to have little to do with money, and it can even be said that these three higher needs are essentially subjective states of the heart. Too much money is sometimes "noise," for example, the famous quote, "I never make friends based on whether or not they have money, they don't have as much as I do anyway." --If this criterion is established, recently Wanda business management can not be listed, there are 30 billion bets do not see on, his family is likely to have little money, is not Sicong he will no longer have friends?

If you are looking for a friend, if you focus on the other party "has no money", I'm afraid it's a sign of lack of heart. If you have a girlfriend who changes in a month or two, or as often as ten years, it may be a sign of inner deprivation, just like eating two meals a day or three meals a day is normal, but if someone eats eight meals a day, then we certainly do not think it is because the person is rich or has a good appetite - the essence of bulimia is always The essence of bulimia is that it is always full of deprivation and never feels "full".

My perception is that every entrepreneur should practice two kinds of martial arts from the beginning, one is the martial art of making money, accounting for 95% of the time to practice are OK; but still should squeeze out 5% of the time to practice the second set of martial arts: the martial art of making this heart full. (Wang Yangming's last words on his death are, this heart is full, also what to say ...... Oh, although he said the light, but in fact, the meaning is not very different la.)

Although it only takes 5%, or even less time, but the benefits, is a healthier body, and family, friends closer, more calm, less anxiety, a sense of light happiness always follow. The overall happiness index exceeds that of more than 90% of the people around you.

The most critical thing is: to achieve success in business, wealth level win around 90% of entrepreneurs, in fact, quite difficult, just rely on talent, effort are not enough, after all, there must also be a luck component with, and the general environment allows ...... Oh, speaking of the word general environment, I am worried about a lot of small partners heart, anxiety level directly soared ten percentage points ...... but "this heart full" thing, hardly rely on luck with or without cooperation, more and the environment has nothing to do with, as long as you decide to do, I bet you will soon improve, to get a stable and lasting sense of happiness is too much to grasp!

The old fan may suspect me, how carved master suddenly began to chicken soup? I must make it clear: I am not a chicken soup, chicken soup is to drink a good taste, in fact, there is no protein, minerals and other effective nutrients. What I'm about to say is actually my quiet research for more than ten years of hard work! reveal a privacy - I actually used to wander in the edge of depression, after all, I have also encountered many times in the dilemma of entrepreneurship, and I am a more or less like the literary people, the inner more sensitive than most people, very easy on the self into the emotional "downward spiral "The appearance is not visible, but in fact people have begun to self-doubt, self-attack, strong to pretend to be calm, but the heart has been depressed and dark, not willing to eat, even the basic deep sleep have been unable to do.

The good news is that I was exposed to the Carnegie books when I was an adolescent, and in my opinion, that was the earliest form of positive psychology. After thinking about having this straw, I spent a lot of time and energy doing all kinds of homework, including Harvard University's positive psychology open class, and even took the initiative to meet a lot of psychology gurus, and even physically go to class - oh, of course, some classes are purely flimsy, or hold the outdated and old-fashioned, almost abandoned by the international mainstream psychology community junk theory ......

Not to sell off, the following will be my ten years of precipitation of the "six swords of filling" to share ...... but the problem of space, today first share two simple, although simple, and even you listen to may say "ah? That's it?" But I can assure you, so simple two moves, as long as you insist, after a week you will come back to me message thanks, after all, these moves are not invented by me, but the positive psychology of this serious discipline, the essence of the last two or three decades to precipitate. (The dregs of the various schools, we have also dabbled, I directly for everyone to step on the mines to demolish it!)

The first move: gratitude meditation - 5 minutes a day is enough, relaxed state, listening to some music, a random posture is not easy to fall asleep, think about why they are a lucky bastard? There are still a lot of people in the world who don't get enough to eat, don't get clean water, and even endure wars, while you, yourself, are currently in good health, your family is healthy, most of you have people who love you, and you have people who love you, blah blah blah go ahead and take stock, and then, start giving thanks to fate for being so good to you! And then, pray that everyone around you can live a happy life ......

If you even find this difficult ...... Well, I'll give you a more simplified version: open a notepad on your phone before you go to bed, just for a minute, and write a "gratitude diary", in fact, just write three things. What makes you happy today, maybe listening to a good song, then thank the guy who wrote the song, or maybe a surprisingly good dish in the cafeteria today, then thank the chef ...... myself a few days ago has been grateful to an unknown guy, because I bought a large bag of pine nuts, peeled, divided into many small bags, keep fresh quite well, and no bad seeds! Eating with glee, if I had not managed to lose 28 pounds, body fat rate 17% need to be maintained, I would like to eat more bags every day, I am so damn grateful to the guy who invented this kind of pine nuts directly without shelling.

That makes it infinitely easier, right? One minute before bed, write down three little things that made you happy and thankful today.

Considering the vast majority of my readers are highly qualified, perhaps I must say how it works, otherwise it wouldn't be good to look like a godfather ...... The principle is: the brain can't tell reality from imagination! As soon as you start doing gratitude exercises, your brain will resurface the little things that make you happy, and the different little blessings will start your dopamine or endorphin or serotonin production. ...... A minute is short, but after the gratitude meditation or gratitude journal is over, those brain chemicals will still be there, will still be there for a while! It's a good idea!

In fact, this is the reverse of how depression works. The depression of depression is something that either happened in the past and you can't get it back, or it may happen in the future, but what does it matter tonight? But you just keep thinking about these bad things, the brain can not distinguish these things did not happen in the present, but the norepinephrine began to spike, sympathetic nerves began to replace the parasympathetic nerves to hold the big picture, your anxiety began to spiral up the dance ......

The theory of brain plasticity tells us that the thinking pattern of the brain, like the "river", can be repeatedly "flushed" to form a new "brain circuit", depression is also so Depression is also flushed out, as long as you play the role of a "victim" every day to think about those bad things, depression is just around the corner; and "fullness" can also be flushed out - any small thing that makes you Any little thing that makes you happy, even a bag of pine nuts, can help the brain to form a circuit - the circuit is, I'm a lucky bastard, I'm so fucking lucky, I'm lucky to meet several happy things every day - since there is no way to have both happiness and depression in the brain at the same time, the next step is Since there is no way to be happy and depressed at the same time in the brain, the next step is the east wind over the west wind, or the west wind over the east wind - and the practice of gratitude every day is to exercise the brain's "positive emotions" a kind of "optimistic neural circuit "The next step is for the east wind to overwhelm the west wind, or the west wind to overwhelm the east wind.

Second trick: more hugs.

This is even simpler, hugs can directly make serotonin secretion, simple, powerful and rapid. Of course, it is recommended that you start with your parents and family, rather than beautiful female colleagues ...... hahahaha. I still haven't expanded the range of hugs to female colleagues, after all, we Asians are too formal in this regard, but family and good friends, is very easy.

The even better point about hugging is that the gift of a rose is a double benefit, as both people's serotonin will spike.

The remaining four tricks will be discussed later, but the principle is similar: strictly in accordance with modern medical research, and effective on a large scale, not immediately miraculous results - but the pursuit of when you are not seriously depressed, early to run their own state of happiness at a higher level, in case something happens to make you depressed, you can also recover more quickly, rather than You will not fall head over heels and never recover.

It's like working out. You shouldn't wait until you're sick to remember to boost your immune system. Rather, when you are not sick, actively exercise to improve immunity to a high level of operation, after all, we will sooner or later encounter disease - the same way, we will sooner or later encounter bad or bad things, so early exercise "brain immunity", on becomes the "second set of martial arts" for entrepreneurs.

I have written many business articles, but I hope that this one will be more solid than all the previous content to help people who start their own business, after all, life is only once and there is no room for failure. Health is the basis of all happiness, and only health belongs to you, once the problem, even if the deposit of tens of billions, as usual, is someone else's money ......

How's that? Tonight, you can start meditating on gratitude!

In the industry generally think "meta-universe" boom recede, Apple kill XR track a little late, but the backhand pull out Apple Vision Pro such "king bomb" level products, let the industry quite amazed.

Then, in the industry will generally focus on artificial intelligence, Apple only launched the "meta-universe" device Apple Vision Pro, can not help but doubt Apple's ability in artificial intelligence geometry.

Let the next big model home, for you to take stock of WWDC 2023, the new Apple Vision Pro, in the end have revealed the Apple artificial intelligence strength?

AIGC Generated Portrait

When using Apple Vision Pro for FaceTime video calls, the absence of a camera facing the user, while the user wears an XR device, also makes the user look very strange.

A more intelligent input method

As we all know, one of the most criticized predicaments of XR industry is the lack of input methods, whether it is the single key input of the handle or the floating keyboard input method, in terms of efficiency and accuracy, compared to the physical keyboard, the experience is very bad.

Apple Vision Pro main interaction methods for the eyes, gestures and voice, which means that voice input may become Apple Vision Pro one of the most important way to type.

While Apple did not emphasize input methods in the Apple Vision Pro introduction, the iOS 17 introduction mentions a smarter input method that not only corrects spelling errors, but even corrects grammatical errors made by the user during typing.

Auto-corrected words are temporarily underlined, giving users a clear idea of which words have been changed and reverting back to the original with a single touch.

What's more, based on device-side machine learning, the input method also automatically improves the model based on each user's keystroke. This brings the input method's auto-correction feature to an unprecedented level of accuracy.

In addition, based on the cutting-edge word prediction Transformer language model, the word association function allows for very fast input of the next word, or even a complete sentence.

This highly personalized language prediction model also allows the input method to better understand the user's language habits and significantly improve input accuracy when using voice input.

The new "Handwriting" app

Along with iOS 17, the new "Journal" app uses machine learning technology on the device to create personalized memories and writing suggestions based on the user's photos, music, exercise, etc. Based on this information, the app provides you with suggestions for recording and writing at the moment that suits you. writing suggestions based on this information.

This means that based on the iPhone's arithmetic power, the device has been able to deploy localized semantic understanding capabilities for processing text, pictures and other multimedia content, as well as having certain generative AI functions.

And at this time, Apple chooses to keep a low profile. In the opinion of Big Model House, Apple's AI capability is really relatively weak in the face of top big models like GPT, and over-emphasizing AI capability is undoubtedly hitting a rock with an egg.

In addition, as a technology company whose main revenue comes from consumer electronics as well as services, Apple needs to emphasize more on the emergence of new features, the improvement of user experience and the continuous increase of user stickiness compared to the relatively general concept of AI.

I. Change in the Hong Kong Securities and Futures Commission's attitude towards retail investors' entry

In 2018, the Hong Kong Securities and Futures Commission introduced the existing regime under the Securities and Futures Ordinance to regulate the virtual asset business. As the virtual asset market was still an emerging industry at that time, its highly speculative nature made it also more risky than the traditional financial market. For the sake of market stability and protection of investors' rights, the SFC limited the scope of licensed virtual asset trading platforms under the SFO to professional investors.

The so-called professional investors specifically include institutional professional investors, corporate professional investors and individual professional investors:

Ø Institutional professional investors refer to recognized exchanges, recognized clearing houses, intermediaries, recognized financial institutions, authorized insurers, recognized collective investment schemes, registered MPF schemes, occupational retirement schemes, the government, etc. as defined in the SFO.

Ø Corporate professional investor means a trust corporation within the meaning of the Securities and Futures (Professional Investor) Rules (a) holding total assets of HK$40 million (or its equivalent in foreign currency) or more; (b) a corporation or partnership holding a portfolio (cash, deposits, securities) of HK$8 million (or its equivalent in foreign currency) or more, or total assets of HK$40 million (or its equivalent in foreign currency) or more.

Ø Individual professional investor means an individual who holds a portfolio of more than HK$8 million (or its equivalent in foreign currency) within the meaning of the Securities and Futures (Professional Investor) Rules.

In November 2020, the HKSAR published a public consultation paper on the establishment of a licensing regime for virtual asset service providers under the Anti-Money Laundering and Terrorist Financing Ordinance. Forty percent of the submissions received by the SFC at that time argued that the authorities should allow retail investors to participate in the trading activities of virtual asset exchanges. It was pointed out that denying retail investors access to licensed virtual asset trading platforms may instead be detrimental to investors as it may drive them to trade on overseas virtual asset trading platforms that can be easily accessed online but are not regulated. If these unregulated virtual asset trading platforms fail, retail investors will not be able to withdraw their assets and will suffer significant losses.

As the market for virtual asset business has matured, the SFC has developed a number of virtual asset policies to gradually allow retail investors to invest in virtual assets on a limited basis. In October 2022, the SFC established an authorization regime for virtual asset futures exchange traded funds (virtual asset futures ETFs), under which three virtual asset futures ETFs were authorized, giving retail investors indirect access to virtual assets in Hong Kong through regulated products.

On February 20, 2023, the SFC published a consultation paper to solicit public views on the Proposed Regulatory Requirements Applicable to Operators of Virtual Asset Trading Platforms Licensed by the SFC, proposing to allow retail investors to use trading services provided by licensed virtual asset trading platform operators, subject to the implementation of a series of measures by virtual asset exchanges to properly protect investors' rights and interests.

II. Account opening: expertise assessment

To ensure that retail investors receive adequate protection, the SFC follows the principle of investor suitability and requires licensed virtual asset platform operators to conduct an expertise assessment of retail investors when establishing business relationships with them. Only after a retail investor has passed the assessment can the platform operator open an account and provide services to the retail investor.

In assessing whether a retail investor has expertise in virtual assets and knowledge of the risks associated with them, the platform operator considers three main aspects: (a) whether the investor has received training or attended courses on virtual assets; (b) whether the investor's current or previous work experience is relevant to virtual assets; and (c) whether the investor has previous experience trading in virtual assets. .

In practice, an investor will be considered to have sufficient knowledge of virtual assets if he or she has made five or more transactions in any virtual asset within the past three years. For those retail investors who are assessed as not having the relevant knowledge, the platform operator should provide them with the necessary training before opening an account for the investor.

III. Risk tolerance: Position limit assessment

Considering the high risk of virtual asset investment, the SFC requires platform operators to conduct a comprehensive assessment of retail investors' potential trading losses and risk tolerance level, and does not set an overall position limit for each customer in virtual assets to ensure that retail investors can reasonably bear the relevant risks and losses in the event of market volatility.

In trade execution, the trading platform generally has automated pre-trade monitoring measures in place so that if a client enters any trade order that would exceed the limit specified for each client, the trade order will be prompted and cannot be executed. Of course, this position limit is also subject to change, and the platform operator will regularly review the risk profile of retail investors and adjust the position limit as appropriate.

The assessment method of the position limit and risk tolerance level is generally designed by the platform operator itself, which can be determined by the platform operator after obtaining the customer's identification information, financial status, investment experience and investment objectives, etc., but the platform operator must explain the assessment method and the assessment results to the customer. If the platform operator uses a risk scoring questionnaire to collect information, the platform operator shall pay special attention to the questions and the appropriateness of the design of the relevant scoring mechanism to ensure that the data in the scoring questionnaire can accurately reflect the true personal situation of the customer.

IV. Risk Alert: Disclosure Obligations of Platform Parties

Although retail investors who can participate in the trading activities of virtual asset exchanges have been evaluated by professional knowledge and risk tolerance level, they still need more protection than professional investors, therefore, the Hong Kong Securities and Futures Commission has imposed the obligation of risk indication and disclosure on platform exchanges.

At the legal level, when trading with retail investors, platform operators should inform retail investors that virtual assets are not legal tender, are not guaranteed by the government and relevant authorities, and that there is uncertainty as to whether virtual assets can be considered "property" under current law, which may affect the nature and enforceability of clients' interests in such virtual assets. This uncertainty may affect the nature and enforceability of the client's interest in the virtual asset. Second, policies in the area of virtual assets are subject to frequent change, and legislative and regulatory changes may adversely affect the use, transfer, trading and value of virtual assets.

With respect to investment risks, platform operators should fully advise retail investors that the protection provided by the Investor Compensation Fund does not apply to transactions in virtual assets by design, regardless of the nature of the token, and that transactions in virtual assets are irreversible and losses resulting from fraudulent or accidental transactions may not be recoverable.

With respect to trading, any advice provided by a platform operator to its clients should be made after thorough analysis and should be fair to all clients and without conflict of interest. For example, if a platform operator highlights a particular virtual asset to a client for commission rebates or other benefits, such conduct may be in breach of the relevant SFC regulations. Where a complex product is involved, the terms, features and risks of which are not reasonably likely to be understood by retail investors, the platform operator should provide a warning statement to investors about the complex product before selling or advising to ensure that the customer is suitable to trade in the complex product.

At our organization, we are passionate about studying the science of the human voice, and one of our areas of focus is the power of deep voices. Through our research, we have discovered some intriguing insights into why a deep voice is so alluring and how it can affect others' perceptions of a person.

So, why are we so drawn to deep voices? Let's dive into the science behind it.

Firstly, it's important to understand that the sound of a person's voice is largely determined by the size of their vocal cords. The longer and thicker the vocal cords, the deeper the voice. As a result, men typically have deeper voices than women, and some men have deeper voices than others.

Research has shown that a deep voice is perceived as a sign of masculinity, strength, and dominance. These are qualities that have been historically valued in men, particularly in traditional gender roles where men were expected to be providers and protectors.

One study found that women rated men with deep voices as more attractive, particularly for short-term relationships. The researchers hypothesized that this may be because a deep voice is associated with high levels of testosterone, which is linked to qualities such as physical attractiveness and sexual prowess.

Another study found that men with deep voices were perceived as more dominant and more likely to succeed in a competitive environment. This may be because a deep voice conveys a sense of authority and confidence, which can be advantageous in situations where leadership and assertiveness are valued.

But it's not just women who are drawn to deep voices. Research has shown that men with deep voices are also perceived as more attractive by other men, particularly in professional settings. In fact, a study of CEOs found that those with deeper voices tended to lead more profitable companies.

So, how does a deep voice actually affect our perceptions of a person? One theory is that it activates certain brain regions that are associated with reward and pleasure. When we hear a deep voice, our brains release dopamine, a neurotransmitter that is associated with feelings of pleasure and reward.

This may explain why a deep voice can be so appealing, even on a subconscious level. It's also worth noting that the sound of a person's voice can affect how we perceive their personality traits, such as their intelligence, trustworthiness, and likability.

In one study, participants listened to recordings of different voices and rated the speakers on various personality traits. The results showed that people with deeper voices were perceived as more trustworthy, while those with higher-pitched voices were perceived as more intelligent.

So, what does all of this mean for you? If you have a naturally deep voice, it may be an advantage in certain situations. However, it's worth noting that the sound of your voice is just one factor that can influence how others perceive you.

At our organization, we believe in exploring all aspects of the human voice, from its physical properties to its emotional and cultural significance. By understanding the science behind our voices, we can gain a deeper appreciation for the power they hold.

If you're interested in learning more about our research and insights into the human voice, we encourage you to subscribe to our account. As a thank you, we're offering a free NFT to all new subscribers.

https://opensea.io/assets/ethereum/0xD9439050CDBEee191FF14aB6ae2b33D46083daD1/0

In conclusion, the power of deep voices is undeniable. Whether we're attracted to them because of their association with masculinity and dominance or because of the pleasure they bring us, there's no denying their allure. By continuing to study the science behind the human voice, we can gain a greater understanding

On Tuesday, March 14, EST, OpenAI announced the launch of GPT-4, a large multimodal model that is "more creative and collaborative than ever before" by taking in image and text input and then outputting text, and "because it has broader knowledge and problem-solving capabilities, it can solve puzzles more accurately." OpenAI said that because GPT-4 has the ability to parse both text and images, it can interpret more complex input.

In addition OpenAI said it will open up the API as soon as possible this time, keeping working with several companies to allow them to integrate GPT-4 into their own products. Microsoft also said within the first hour of the announcement, "If you've used the new version of Bing at any time in the last six weeks, you've gotten an early look at the power of OpenAI's latest model." Even from across the screen and across the Pacific Ocean, I can feel Microsoft's face is full of "complacency".

The Past Life of GPT Series

GPT is called GenerativePre-trainedTransformers, which is a natural language generation model derived from the "converter architecture". The term "pre-training" in the GPT name refers to the initial training process on a large text corpus, where the model learns to predict the next word in a text. This provides a solid foundation for the model to perform well on a limited number of tasks with specific data downstream.

As mentioned earlier, the GPT family is derived from the "converter architecture", which is essentially a deep learning model that uses an "attention mechanism" to assign different weights according to the importance of the components of the incoming data, and is mainly used in natural language processing (NLP) and computer vision. processing (NLP) and computer vision (CV). The architecture was first proposed by Google in 2017 and, like the Recurrent Neural Network, which has been used for more than three decades, is designed to process "sequential input data" such as natural language (in this case, language specifically refers to written language). Unlike recurrent neural networks, which can only process one word at a time, the attention mechanism used in the "converter architecture" provides context for any position in the input sequence, allowing for more parallel computation and less training time, which is a significant This is a significant improvement.

In addition to the now-hot ChatGPT, there are also BioGPT and ProtGPT2, the former is Microsoft's self-developed GPT model dedicated to biomedical field, and the latter is a GPT model dedicated to protein research. The main reason for their obscurity is that they are not as widely used as ChatGPT, so naturally they do not attract many fans.

The GPT concept was first proposed on June 11, 2018, when OpenAI published a paper titled "Improving Language Comprehension through Generative Pre-training". At the time, the best performing natural language generation models relied mainly on supervised learning models, which had many large objective limitations and many difficulties in practical applications (e.g., translation and interpretation) due to the lack of textual material in which a corpus could be built up for many rare languages. In addition, the time and capital costs required to train very large models are also very high. In view of this, GPT proposed a "semi-supervised" (later commonly referred to as "self-supervised") approach: a pre-trained model is trained on unlabeled data, and then a discriminable fine-tuned version of the model is trained on a small amount of labeled data.

Before GPT-4, GPT-1, GPT-2, GPT-3, GPT-3.5, the most obvious changes are the number of internal parameters and the increasing volume. The first generation has about 120 million (4.5GB), the second generation about 1.5 billion (40GB), and the third generation about 175 billion (570GB). However, it has been dismissed by OpenAI CEO Altman as "complete nonsense", but he didn't disclose the exact number either.

The real face of GPT-4

From the information available so far, GPT-4 has been developed in the following aspects: powerful image recognition; text input limit raised to 25,000 words; significant improvement in response accuracy; ability to generate lyrics and creative text; and implementation of style changes. We hope to bring it to everyone soon." U.S. Rep. Don Beyer also confirmed to the New York Times that Alterman demonstrated GPT-4 during a visit to Congress in January, specifically showing the improved "security controls" compared to other AI models.

OpenAI has ambitions, and although the paper showing GPT-4 was written more like a technical report, that won't stop the progress. In addition, OpenAI has organized a number of experiments that show that GPT-4 performs at a similar level to humans on various professional tests and academic benchmarks. For example, it passed the mock bar exam and scored in the top 10% or so of test takers; in contrast, the previous GPT-3.5 score was not at all complimentary.

In addition to word processing, GPT-4 can also handle image input, and the new capability goes hand-in-hand with plain text, allowing the user to specify any visual or verbal task. Specifically, it generates the corresponding textual output (natural language or code, etc.) given a scattered graphical and textual mix of inputs by the user. GPT-4 demonstrates similar functionality to text-only input on a range of domain-specific domains, such as documents with text and photos, diagrams, or screenshots.

Unlike the previous GPT-3.5's fixed lengthy, calm tone style personality, GPT-4's ability to balance text and graphics really shines through, even for foreign stems and diagrams, I just don't know how it will be in the face of domestic ones. Now developers as well as ChatGPT users can specify the language style of their AI and the tasks it will handle by describing these directions in system messages. And this time OpenAI is the first to open up API access, so users can customize to achieve different user experiences within a certain range. Obviously, the official knows exactly what users want to do with ChatGPT.

The last trial run is scheduled for August 6-12 on the Goerli test network. If all goes well, the merger will take place in September.

Despite the upcoming Merge™️, there are still too many misconceptions about this unprecedented network upgrade on Ether.

Why is it a big deal?

How does it affect the ETH token economy? Will it affect users' gas fees?

There are many questions that need to be answered.

This article provides much-needed FAQs on everything related to the merger.

1. What is a "Merge"?

"The Merge is the name of the event when the Ethernet blockchain changes from using Proof of Work (PoW) to Proof of Stake (PoS).

It is called a "merge" because it is a merger of two separate blockchains currently running in parallel. The main ethereum blockchain is being "merged" with a special-purpose blockchain called the Beacon Chain.

The beacon chain launches on December 1, 2020. The purpose of Beacon Chain is to do one thing and one thing only: to be a proof-of-stake (POS) blockchain.

There are no transactions on the beacon chain. There are no tokens or DeFi apps. It is an empty blockchain, used only to run a proof-of-stake consensus mechanism.

Because the beacon chain is an "empty chain," it can be merged with the Ether blockchain, replacing Ether's PoW mechanism, without worrying about any other variables.

Once the two chains are merged, Ether's PoW authentication will be replaced by a new PoS consensus mechanism.

1. Why is the "merger" getting so much hype?

This merger is considered one of the biggest events in crypto history since the birth of Bitcoin.

There are several recognized reasons for this.

No blockchain in the history of cryptocurrencies has ever undergone such a significant change. Blockchains don't often change such a critical part of their operation, and none have made changes with such a large and robust economy built on top of them as Ether.

The market cap of ETH assets today is $203 billion (previously $550 billion 😭), with billions more based on the value on top of the network. Ether is by far the largest and most powerful economic ecosystem in crypto, and the security of all this economic activity will shift from a PoW-based economy to a PoS-based economy.

Of course, there are many risks associated with errors, which is one of the main reasons why the merger is taking so long.

There is still a very large amount of testing and refinement to be done.

1. What is the impact of the merger on ETH?

The merger has had a huge impact on the economy of ETH. For investors, this aspect of the merger is the most important.

The merger has completely changed the economics of ETH in two ways: reducing ETH issuance and treating ETH as a native earning asset.

Reducing ETH issuance The merger will reduce annual ETH issuance from 4.3% to 0.43%.

This is due to the fundamental efficiency gains from the PoS consensus mechanism, which is designed to provide the highest level of blockchain security at the lowest cost, and to redirect those savings to ETH by reducing the amount of ETH that needs to be issued to pay for security.

PoW is costly and requires significant resource overhead to compensate security providers (miners) for their services.

In contrast, the cost of PoS security is simply the opportunity cost of money and does not represent any real-world commodity or tangible cost. Unlike PoW, PoS does not require the issuance of large amounts of currency to cover the cost of security. Therefore, these lower security costs make the PoS consensus mechanism more efficient.

Due to the reduced need to pay PoW miners, Ether was able to reduce its annual ETH issuance from 4.3% to 0.43%. The reduction in new ETH issuance is usually considered ...... very bullish.

This is because PoW miners sell most of the rewards they receive immediately, and over time, they sell close to more than 90% of their total mining.

With PoS, issuance is reduced by more than 90%, while the administrative costs of being a PoS verifier are essentially reduced to zero.

PoW proponents argue that the high costs of the PoW blockchain are a feature, not a bug. They argue that these costs create volatility in assets, which hinders centralization because they force sellers to pay the cost of PoW miners. While PoW may indeed create asset decentralization, it also creates security centralization, which is a secondary effect.

👨💻 For further reading of my previous Bankless article, please click here for it.

1. Why the deflation of the merged ETH?

Almost exactly one year ago, on August 5, 2021, Etherpad launched EIP-1559, an upgrade that changed the way Etherpad transaction fees are managed.

Rather than simply paying miners the full transaction fee, the majority of the transaction fee is destroyed.

There are many reasons for doing so. To dive deeper into this rabbit hole, listen to our podcast with crypto-economics researcher Hasu 🎧.

After the merger, ETH issuance dropped by over 90%, increasing the proportional magnitude of how much ETH each block burns.

When Ether's gas cost is 7 gwei or higher, the rate of ETH destruction is higher than the rate of ETH issuance, leaving a reduced supply of ETH.

At the peak of the bull market, gas prices stayed at 200 gwei or higher for months on end, making the 7 gwei threshold a very low one. This is true even now in a bear market.

👉 See historical Gas fees here.

👉 Simulate a merger scenario for ETH issuance here.

1. Will the merger lower the gas fee for Ether?

No, it will not. This is a misconception that comes from the confusion between "Ether 2.0" and "merge".

"ETH 2.0" is the name of the future state of Ether that is no longer used in the Ether community. "ETH 2.0" refers to the future version of Ether that will enable PoS and sharding.

For a while in the history of Ether, the two updates, PoS and Sharding, were thought to be coming at the same time. As development progressed, developers realized that they could divide these updates and release them separately.

Unfortunately, the "ETH 2.0" nomenclature persists.

Partitioning will reduce Ether's gas costs on L1, but the real gas cost minimization and elimination for end users will ultimately occur on L2s such as Optimism, Arbitrum, Polygon, StarkNet, zkSync, or other L2s.

When Ethernet L1 gas fees are reduced, L2 fees will be reduced by an order of magnitude. Ethernet's Gas fee solution was never about lowering them on L1, it was about migrating users to L2 and letting them enjoy the fast and cheap trading experience there.

1. will the merger increase the transaction speed of Ether?

This is actually the same question as above (will merging lower the gas cost of Ether), just phrased differently.

Transaction volume and transaction fees provide the same supply and demand dynamics in a cryptocurrency network.

The merger does make Ether's block-out time (how often a block is added to the Ether blockchain) slightly faster, from an average block time of 13.6 seconds to 12 seconds.

This represents a 12% increase in transaction capacity, and therefore a 12% reduction in gas cost.

However, this is a negligible amount and should not be considered a "reduction in gas costs". 7.

1. Has the merger reduced the energy consumption of ethereum?

Yes. This is a great moment and one of the main results of the merger and PoS.

After the merger, Ether will consume about 99.95% less energy than it currently uses.

PoS uses capital, not energy, to protect the security of the blockchain. As a result, the remaining energy required to maintain Ether is comparable to basic computer usage, such as what you are doing right now, such as reading this article, sending a tweet, downloading a movie to your hard drive, etc.

With PoS enabled, the energy cost of Ether is just to run one node - about 2.6 MWh per year. That's about 1,300 times less than what the entire U.S. gaming industry consumes.

Ether will actually become the most environmentally friendly financial system the world has ever seen.

Will ETH pledgers sell off their ETH after the merger?

No! They can't.

ETH withdrawals from ETH pledged nodes will not be activated immediately after the merger. This is due to making things as simple as possible while Ether is going through the biggest and most complex upgrade the industry has ever seen.

Withdrawals are expected to be unlocked 6-12 months after the merger.

So, will pledgers simply dump their ETH once the withdrawal feature is enabled?

Probably, but there are still limitations in that regard. There is a withdrawal/deposit queue that limits the speed at which people can pledge and unpledge. This is another mechanism to keep the chain stable and not let the rapid fluctuations in the Ether application layer affect the security of the chain.

The bottleneck limit for deposits/withdrawals is X/ETH per day, where X is equal to.

#validators / 65536, rounded to the nearest integer. The number 65536 is derived from the 16th power of 2. For some reason, ethereum developers like these square numbers. It is apparently called the "Churn Limit Quotient" and you can read about it here.

Currently, there are 433916 Ether verifiers on the beacon chain. To find how many validators there are per epoch, divide it by 65,536 and round down.

433,916 (total number of validators) / 65,536 ≈ 6 validators per epoch So the number of activations/deactivations is 6 validators per epoch. An Ether epoch is 6.4 minutes, and there are 225 epochs in 24 hours.

225 epochs * 6 validators per epoch = 1350 validators per day Therefore, the current validator activation/deactivation rate is 1350 validators per day.

Each verifier has 32 ETH, so a maximum of 43,200 ETH can be unlocked per day.

32 ETH * 1350 validators = 43,200 ETH In addition, APY as an ETH pledgee will increase after the merger, as ETH pledges also charge transaction fees.

This is expected to increase the ETH yield from 4.2% to over 5%, but it will be higher during periods of heavy gas consumption.

1. why 32 ETH?

Why does it take 32 ETH to run a node and not 31, 33 or any other number?

The answer is that the more nodes there are, the greater the total amount of information passed between nodes. If the amount of ETH is low, more nodes can come online. While this is good for decentralization (🙂 ), it is a limit to scale (☹️).

32 ETH was chosen as the best balance between these two ends of the spectrum, and also because it is a square number: 2^5.

Since node messaging is exponential, reducing the ETH validator requirement from 32 to 16 would increase the messaging volume of all nodes by a factor of 4. 32 was chosen as the minimum number of ETHs that could produce a "final result" in "768 seconds or 2 epochs". The minimum number of pledges that can produce a "final result" within "768 seconds or 2 epochs".

Is 32 ETH permanent?

Not necessarily! With improvements in consumer hardware, message compression and better signature aggregation, it can certainly be revised to 16 or lower.

1. Proof of Benefit is not on-chain governance

A common criticism (mainly from the Bitcoin camp) is that PoS is equivalent to the "legal system" we are trying to get rid of.

What they mean is that whoever holds the capital, holds the power.

This is a frustratingly misleading position, and one that many of the frequent proponents of this view are probably spreading in bad faith to discredit any consensus mechanism outside of Bitcoin.

The role of an ETH verifier is exactly the same as that of a PoW miner. This is a 1:1 comparison.

ETH holders have no governing power over Ether. Just like in Bitcoin, this power is held by non-validating node operators (aka "communities").

Ironically, these camps promote PoS as a "rich get richer" scheme. And it is an undisputed fact that PoW mining facilities generate a greater return on investment for richer and more capitalized entities than smaller, less capitalized entities.

Tl;DR

PoS Ether with native ETH returns is the most democratized consensus mechanism because it offers the same proportional returns as 32, 320, 3200 or 32,000 ETH. They all get the same ~5% interest rate.

This is in stark contrast to PoW miners, where a $100 million investment in a mining facility will yield far more than 10x the arithmetic power of a $10 million investment due to all the economies of scale that come with PoW hardware and energy costs.

The term "metaverse" first appeared in science fiction. American science fiction writer Neal Stephenson first coined the term "metaverse" in his 1992 science fiction novel Snow Crash. In this novel, the term metaverse refers to a state in which the real world is mixed with a virtual world created through augmented reality and virtual reality technologies. Although this work was not as influential in the science fiction world as another science fiction master, William Gibson's similarly titled Neuromancer, published in 1984, the term "metaverse" has since come to rest in the limelight with the rapid development of the online world, and has since been borrowed from time to time by various works have been borrowed from time to time.

The 2018 science fiction movie Ready Player One is considered by many to be the most visual explanation of the metaverse concept. With the help of virtual devices, people can move freely between the real world and the virtual world. Needless to ask, the more unrestrained virtual world can bring people the surprise and joy are far more than the real world.

If the concept of metaverse, which was borrowed from science fiction movies, was confined to science fiction fans at that time, then from October 2021, the concept officially "broke the circle" and started to be discussed by people who knew nothing about it before.

In October 2021, Mark Zuckerberg, the founder of the 17-year-old American social networking giant Facebook, published a Founder's Letter (2021) announcing the company's name change to Meta, demonstrating his commitment to the meta-universe A big event in the Internet world happened. Although the creator of the term metaverse, Neal Stephenson, then took to his social media accounts to make it clear that the metaverse concept he had created had nothing to do with the newly renamed Meta, the sci-fi Internet concept was already sweeping the world at a breakneck pace. Then, in a video presentation featuring Zuckerberg, the word "metaverse" appears more than 80 times in nearly 90 minutes to explain to viewers what a metaverse really is.

After watching this nearly 90-minute video, I felt more confused than ever, except for the etymology of the word "metaverse" in English. According to Zuckerberg, the metaverse is a merger of the root word "meta" and the word "universe". The Greek word "meta" has the meaning of "beyond" or "above ......" (beyond). For example, the Chinese word metaphysics is a merger of meta and physics, and was once translated as meta-physics.

What about the rest? In this video, Zuckerberg and some of the researchers at Meta try to show the magic that people can experience living in a meta-universe. It seems that people are no longer bound to their physical bodies, that the limits of space-time are no longer an issue, that they can move at the speed of light, that they can choose their appearance at will, that they can show themselves as they are, or that they can choose an Avatar to meet their friends and family. ...... The question is, what elements of the metaverse life shown in the video have been The question is, which elements of the metaverse life shown in the video have become reality and which are still a distant fantasy? What kind of efforts do people need to make to enter the metaverse, and what kind of changes will people's real life undergo? These questions, like the face of Zuckerberg in the video, make it difficult to distinguish which is his real life, and which is computer effects?

Articles and videos about the metaverse abound today, but it is still difficult to give a precise definition of the metaverse. What is certain is that the birth and development of the metaverse, and some of the concepts associated with it, are mostly related to science fiction and computer games. Not only is the metaverse itself derived from science fiction, but the term "avatar" as used in the metaverse, while fundamentally Hindu in origin, is actually more closely related to computer games - game designer Richard Garriott first used the term avatar" to describe the image of the game player in the game.

It can be said that each person has a meta-universe in his own mind that he defines. For example, many people believe that the metaverse is the next form after the evolution from the Internet to the mobile Internet, but it is difficult to define the core characteristics of the new generation of Internet forms. In people's descriptions of the metaverse, many technologies are integrated: what exactly will be changed when blockchain, virtual and augmented reality, sensors, cameras, 5G networks, and other old and new technologies are overlaid together? Many people are still confused.

In January 2020, venture capitalist Matthew Ball published a long article, "The Metaverse: What It is, Where to Find it, and Who Will Build It," in which he discusses some of the metaverse's The Metaverse: What It is, Where to Find It, and Who Will Build It, which discusses some of the characteristics of the metaverse, such as the fact that it expands the real and virtual worlds, contains a fully fledged economic form, offers unprecedented interoperability, etc. But when we look at the Internet that people are currently using, it has most of these characteristics, or is being refined toward these goals. Some people also conclude that the so-called metaverse, which is everything people can do in real life, can be done in the virtual digital world - but what is the point of such a metaverse?

Zuckerberg defines the metaverse as a new form of human-centered Internet. In a metaverse, people no longer need to stare at the screen of their computers or phones (or rather, there are screens everywhere in the metaverse), but become part of the content of the metaverse itself. In the metaverse, people can communicate and interact with their friends and family (or their avatars) without barriers, read each other's eyes and movements, and escape from the real world and the physical body to the greatest extent possible, which is perhaps the core value of the metaverse described by the newly renamed Meta.

It can be said that the metaverse concept not only bears the distinctive mark of the generation that grew up in the world of the Internet and games, but also Zuckerberg's own beliefs as a technological progressive and futurist. He believes that a new generation of technology is the key to solving most societal problems, and that every 10 to 15 years, a new disruptive technology or machine will emerge - so the metaverse is just in time. It is with this belief that Facebook has been able to transform itself rapidly under his leadership.

The question is, how far away from reality is this unfettered virtual world? To what extent can people actually get rid of the constraints of real life, so as to maximize the freedom of mind? If everything can be virtual, then what is the meaning of real life? What is the relationship between virtual and reality, and how can people tell the difference between the two?

"The Past and Present of "Virtual Reality It is a question of how far people can trust their senses and how far they can deceive their senses. It is another question to what extent we can rely on our senses to perceive "reality". Both Chinese and foreign thinkers have struggled with the subtle distinction between reality and dreams. Zhuangzi therefore proposed the idea that we may not have clear criteria to distinguish which is closer to reality between reality and dreams. As the saying goes, "In the old days, Zhuang Zhou dreamed of a butterfly, and it was a vivid butterfly. But when he suddenly realized it, he was also a superstitious Zhou. I don't know whether Zhou's dream was about Hu Die or Hu Die's dream was about Zhou. If the dream of Zhou and the dream of the butterfly is Zhou, then there must be a division. This is called materialization"-whether Zhuang Zhou dreamed and became a butterfly or Zhuang Zhou appeared in the butterfly's dream, it is really difficult to distinguish between the two.

The French thinker Descartes was also wary of the similarities between dreams and reality, so he believed that the reality perceived through the senses was only an indirect reality. Where then does the most direct reality come from? From a rationalist point of view, sensory experience is not trustworthy, but only thought is. Therefore, there is the theory of "I think, therefore I am".

Others deny the reality of the real world itself. The Pythagoreans believed that "everything is number" and that the world is merely a reflection of mathematics; the ancient Greeks believed that the world people perceive may only be a projection of a more fundamental reality in the world, just as prisoners in prison could only see the images of the outside world on the walls, which is another level of doubt about reality.

Reality may not be credible, but the senses are easily deceived - this is what people have long learned from their psychological and life experiences. People often complain about the constraints of reality, and the inconvenience of the physical body brings them mental pain; on the other hand, they hope to achieve the separation of the mental and physical bodies, so that the mind can travel without obstacles. As Shakespeare's poem describes the scene.

If the dull substance of my flesh were thought.

(Injurious ditance should not stop my way.)

For then despite of space I would be brought.

From limits far remote where thou art. (From limits far remote where thou dost stay.)

No matter then although my foot did stand, (No matter then although my foot did stand,)

No matter then although my foot did stand, What hindrance is there to me at the farthest end of the earth? (Upon the farthest earth removed from thee?)

For nimble thought can jump both sea and land, wherever it wants to reach.

For nimble thought can jump both sea and land. (As soon as think the place where he would be.)

One is surprised to discover that, by cleverly tricking the senses, Shakespeare's scenarios may actually be realized with the aid of modern technology - if one is willing to enter a virtual world made up of digital technology, or as it is more fashionably called: the metaverse.

On this level, both the so-called metaverse and the so-called virtual reality are ultimately an art of creating illusions by tricking one's senses with one's knowledge and consent. To the greatest extent possible, the separation of one's spirit from one's body, as if one were in an illusory world, can even be said to be a "basic need" of human beings at the spiritual level, as well as a maximum degree of "freedom". Perhaps it is because of this realization that the word "teleport" is mentioned several times in Zuckerberg's video preaching. In life, if you want your physical body to switch locations and scenes, not only by the laws of physics, but also by the constraints of the social environment - but if you just want to achieve spiritual travel, even with the current technical conditions, you only need to wear a virtual reality device can be achieved.

In retrospect, I have had two brief experiences with the metaverse myself. Once in Hong Kong Ocean Park, before riding a roller coaster, each passenger put on special glasses - with the start of the roller coaster, glasses also began to appear in a variety of three-dimensional images, there are a variety of magical animals, dragons flying around, I was in the blue sky, and then dive into the bottom of the sea ... ... ... I must admit, in that short minute or so, although I know that what I saw are only illusory three-dimensional images, but the feeling of excitement and excitement still makes me remember.

The next time I entered the metaverse was when I visited the Casa Battló in Barcelona, Spain, designed by the genius architect Gaudi. I put on a pair of glasses and walked into this apartment inspired and themed by marine life. As I walked into each room of the apartment, I could see various images of sea life appearing around me, with giant whales swimming past me from time to time, the illusory images blending perfectly with the surroundings.

Of course, my experience was not really into the metaverse, at least the device I was wearing did not have access to the Internet, and I could not communicate with other people through the glasses. Nevertheless, the feeling of being out of reality and entering an alien world instantly was still fresh. By metaverse standards, my two experiences, one called virtual reality and the other augmented reality, both relied on head-mounted devices to make them possible.

The idea of building a virtual sensory world is a long-standing one, and in fact such technologies were first used in special scenarios to train people to react in the real world. in the 1920s, American engineer Edwin Link invented the world's first flight simulator to train pilots. Later, there were theaters with seats that vibrated and rooms that emitted special smells to simulate movie scenes. As computer technology advanced, by the 1980s, computer scientist Jaron Lanier coined (or at least popularized) the term "virtual reality. Lanier set up a company in the 1980s to develop virtual reality eye wear and gloves, and the first hype for virtual reality technology began in the United States. But it was soon discovered that virtual reality devices were of little use to most people, and the hype quickly subsided.

When another round of hype about virtual reality technology emerged in the 21st century, some engineers preferred to use the more realistic term "virtual environment" in order to reduce the science fiction aspect of the hype; by the 1920s, even the term "virtual reality" could not satisfy the hype, and the more science fictional "metaverse" emerged. The hype was not useless.