Zuckerberg probably didn't expect to miss a decades-old windfall when he was so focused on the metaverse.

According to recent reports in foreign media, Meta lost at least one-third of its research staff last year. The displaced employees were drained by their work and lost confidence in the company's strategy.

There is no doubt that Meta has fallen behind in this generative AI race. It is trying to catch up with the front-runners with some messy footsteps, but finds itself overwhelmed.

Meta had a strong interest in AI, but like many companies, it failed to find a direction for commercialization. meta turned to the meta-universe, which looked more promising, and threw AI to lab scientists.

In February, Zuckerberg said he had assembled a "top-tier" AI team, and in an internal letter sent to employees in March, Zuckerberg said that the company's next biggest investment would be in the development of a new AI system. In a March internal letter to employees, Zuckerberg said that the company's biggest investments would be in AI.

The Meta advertising business is underperforming, and internal layoffs are unsettling. Being in the vortex of internal and external troubles, Meta needed AI to inject fresh blood, and it could not and did not dare to be absent. However, when Zuckerberg turns to run wildly towards AI, where is he going to put the Meta universe?

A misstep spanning a decade

In 2013, Facebook tapped Yann LeCun, the "godfather" of AI, from New York University, who had been a pioneer in the field of AI since the 1980s, to work on convolutional network models.

The addition of Yann LeCun did give Meta a reputation in the academic community and led to some important papers. However, Meta faced the common problem of AI companies at that time, which was the difficulty in finding grounded scenarios to commercialize the technology.

After all, Lichun Yang was from the academic world and was not sensitive to the business world. When the big model started to heat up in 2020, Lichun Yang flatly refused to devote himself to this new direction, citing the reason that the big model was too "popular" and had no "scientific value".

Yang's position is actually very common in the academic world, where academics usually put profit after research value. Until OpenAI exploded onto the global scene, Lichun Yang was still not a big fan. At an online meeting of small media and tech executives in early 2023, Lichun Yang commented, "In terms of the underlying technology, ChatGPT is not particularly innovative or revolutionary. Many research labs are using the same technology and doing the same kind of work." After being taken down by the OpenAI CEO, Lichun Yang was more direct: ChatGPT has a very superficial grasp of reality.

For hard technologies, getting out of the lab and into large-scale applications is an insurmountable threshold. No matter how superficial or not revolutionary Yang Lichun thinks ChatGPT is academically, at least commercially, ChatGPT has been an unparalleled success and has found a landing point for AI technology after many years.

It can be said that when Meta hesitated whether to invest for the big model, they have missed the key opportunity to seize the first opportunity.

It was only after the big model became a buzzword that Meta caught up with hindsight. Once again, Zuckerberg has his core business firmly in hand. The new generative AI team reports directly to CPO (Chief Product Officer) Chris Cox, Zuckerberg's highly trusted right-hand man and one of Meta's first 15 engineers who has been with the company for 18 years.

Meta's technical strength is not bad, and it quickly came up with its own open source model, LLaMA, but for a huge company with 80,000 people, a drastic turnaround and reorientation is bound to cause chaos; Meta just started to tackle the big model not long after it started on the metaverse, and the infrastructure is not ready yet. The result of the plucked power is a massive departure of the research team. 6 out of 14 authors of LLaMa-related papers have indicated that they are leaving. Compared to the flexible and energetic startups, the clunky and sluggish Meta is losing its appeal to AI professionals.

Metaverse and AI either/or?

In this competition, Meta is already behind at the starting line. Meta was not invited to the AI summit held at the White House in May, which implies that Meta is no longer the company that holds the cutting-edge technology in the AI field at the moment.

Of course, it is unfair to blame Yang Lichun alone for Meta's lagging behind. The company was already in a period of adjustment, finally recovering slightly in 2023 after experiencing a plummeting stock price and major layoffs in 2022. Zuckerberg described 2023 as the "Year of Efficiency" - shutting down projects and downsizing staff.

Meta's bet on the meta-universe has been bleeding money. According to Meta's financial report, its metaverse unit Reality Labs lost $13.7 billion in 2022, and the losses are still growing.

And on the revenue side, because of Apple's change in privacy policy and other reasons, Meta's revenue fell 1% in 2022, and its net profit fell as much as 41%.

Unlike Google, Microsoft and Android, intelligent cloud services, Meta revenue structure is very single, mainly relying on social platform advertising. After Apple changed its privacy policy, Meta could not find much solution, and its advertising business will probably continue to decline in the future.

And 2023 is the year when generative AI will explode. Meta, which is in the process of reducing costs and increasing efficiency, will have to spend more energy to deploy its limited funds in the face of AI projects that will cost money and people.

The bad thing is that Meta has not figured out which way it should go either.

Zuckerberg has no intention of giving up on the metaverse. in the first quarter of 2023, the metaverse division lost $4 billion, a 35% expansion in losses year over year, and Meta also expects Reality Labs losses to continue to grow in 2023.

But at the same time, Zuckerberg said again to invest more in AI, the amount invested will also set a record in the company's history. (single largest).

Meta's AI model is open source, and they are counting on this format to bring in huge revenues in the future, as Android has. But clearly, at the moment LLaMa will not bring in much revenue.

In June, Meta finally released its own AI music model, Music Gen. While ChatGPT and Midjourney were evolving like wildfire, Music Gen looked less impressive. Since then, Meta has released Voicebox, a speech generation system that generates audio response messages. Zuckerberg also announced that he would launch Meta's own chatbot.

With social platforms like Instagram, Facebook and WhatsApp under Meta, there is actually no shortage of AI landing scenarios.

But Meta cannot be greedy. No matter AI or Meta universe, it is a black hole that sucks money. While the main business does not see an upturn, Meta wants to grab two bottomless pits at the same time. Unlike the well-funded Microsoft and Google, Meta actually does not have much opportunity for trial and error. In the end, Meta will have to choose between the priorities of AI and meta-universe .

Zuckerberg probably didn't expect to miss a decades-old windfall when he was so focused on the metaverse.

According to recent reports in foreign media, Meta lost at least one-third of its research staff last year. The displaced employees were drained by their work and lost confidence in the company's strategy.

There is no doubt that Meta has fallen behind in this generative AI race. It is trying to catch up with the front-runners with some messy footsteps, but finds itself overwhelmed.

Meta had a strong interest in AI, but like many companies, it failed to find a direction for commercialization. meta turned to the meta-universe, which looked more promising, and threw AI to lab scientists.

In February, Zuckerberg said he had assembled a "top-tier" AI team, and in an internal letter sent to employees in March, Zuckerberg said that the company's next biggest investment would be in the development of a new AI system. In a March internal letter to employees, Zuckerberg said that the company's biggest investments would be in AI.

The Meta advertising business is underperforming, and internal layoffs are unsettling. Being in the vortex of internal and external troubles, Meta needed AI to inject fresh blood, and it could not and did not dare to be absent. However, when Zuckerberg turns to run wildly towards AI, where is he going to put the Meta universe?

A misstep spanning a decade

In 2013, Facebook tapped Yann LeCun, the "godfather" of AI, from New York University, who had been a pioneer in the field of AI since the 1980s, to work on convolutional network models.

The addition of Yann LeCun did give Meta a reputation in the academic community and led to some important papers. However, Meta faced the common problem of AI companies at that time, which was the difficulty in finding grounded scenarios to commercialize the technology.

After all, Lichun Yang was from the academic world and was not sensitive to the business world. When the big model started to heat up in 2020, Lichun Yang flatly refused to devote himself to this new direction, citing the reason that the big model was too "popular" and had no "scientific value".

Yang's position is actually very common in the academic world, where academics usually put profit after research value. Until OpenAI exploded onto the global scene, Lichun Yang was still not a big fan. At an online meeting of small media and tech executives in early 2023, Lichun Yang commented, "In terms of the underlying technology, ChatGPT is not particularly innovative or revolutionary. Many research labs are using the same technology and doing the same kind of work." After being taken down by the OpenAI CEO, Lichun Yang was more direct: ChatGPT has a very superficial grasp of reality.

For hard technologies, getting out of the lab and into large-scale applications is an insurmountable threshold. No matter how superficial or not revolutionary Yang Lichun thinks ChatGPT is academically, at least commercially, ChatGPT has been an unparalleled success and has found a landing point for AI technology after many years.

It can be said that when Meta hesitated whether to invest for the big model, they have missed the key opportunity to seize the first opportunity.

It was only after the big model became a buzzword that Meta caught up with hindsight. Once again, Zuckerberg has his core business firmly in hand. The new generative AI team reports directly to CPO (Chief Product Officer) Chris Cox, Zuckerberg's highly trusted right-hand man and one of Meta's first 15 engineers who has been with the company for 18 years.

Meta's technical strength is not bad, and it quickly came up with its own open source model, LLaMA, but for a huge company with 80,000 people, a drastic turnaround and reorientation is bound to cause chaos; Meta just started to tackle the big model not long after it started on the metaverse, and the infrastructure is not ready yet. The result of the plucked power is a massive departure of the research team. 6 out of 14 authors of LLaMa-related papers have indicated that they are leaving. Compared to the flexible and energetic startups, the clunky and sluggish Meta is losing its appeal to AI professionals.

Metaverse and AI either/or?

In this competition, Meta is already behind at the starting line. Meta was not invited to the AI summit held at the White House in May, which implies that Meta is no longer the company that holds the cutting-edge technology in the AI field at the moment.

Of course, it is unfair to blame Yang Lichun alone for Meta's lagging behind. The company was already in a period of adjustment, finally recovering slightly in 2023 after experiencing a plummeting stock price and major layoffs in 2022. Zuckerberg described 2023 as the "Year of Efficiency" - shutting down projects and downsizing staff.

Meta's bet on the meta-universe has been bleeding money. According to Meta's financial report, its metaverse unit Reality Labs lost $13.7 billion in 2022, and the losses are still growing.

And on the revenue side, because of Apple's change in privacy policy and other reasons, Meta's revenue fell 1% in 2022, and its net profit fell as much as 41%.

Unlike Google, Microsoft and Android, intelligent cloud services, Meta revenue structure is very single, mainly relying on social platform advertising. After Apple changed its privacy policy, Meta could not find much solution, and its advertising business will probably continue to decline in the future.

And 2023 is the year when generative AI will explode. Meta, which is in the process of reducing costs and increasing efficiency, will have to spend more energy to deploy its limited funds in the face of AI projects that will cost money and people.

The bad thing is that Meta has not figured out which way it should go either.

Zuckerberg has no intention of giving up on the metaverse. in the first quarter of 2023, the metaverse division lost $4 billion, a 35% expansion in losses year over year, and Meta also expects Reality Labs losses to continue to grow in 2023.

But at the same time, Zuckerberg said again to invest more in AI, the amount invested will also set a record in the company's history. (single largest).

Meta's AI model is open source, and they are counting on this format to bring in huge revenues in the future, as Android has. But clearly, at the moment LLaMa will not bring in much revenue.

In June, Meta finally released its own AI music model, Music Gen. While ChatGPT and Midjourney were evolving like wildfire, Music Gen looked less impressive. Since then, Meta has released Voicebox, a speech generation system that generates audio response messages. Zuckerberg also announced that he would launch Meta's own chatbot.

With social platforms like Instagram, Facebook and WhatsApp under Meta, there is actually no shortage of AI landing scenarios.

But Meta cannot be greedy. No matter AI or Meta universe, it is a black hole that sucks money. While the main business does not see an upturn, Meta wants to grab two bottomless pits at the same time. Unlike the well-funded Microsoft and Google, Meta actually does not have much opportunity for trial and error. In the end, Meta will have to choose between the priorities of AI and meta-universe .