Subscribe to shotx
<100 subscribers
Share Dialog
Share Dialog
Earlier this year in February, Tidu CEO Xia Yiping announced that Tidu automotive robots will integrate the full capabilities of Baidu Wenxin Yiyin.
Only half a year has passed, the wind of the big model will be swept to a large number of car companies with the momentum of a prairie fire. According to incomplete statistics, the current Azure, Xiaopeng, ideal, Great Wall, Geely, Chery, etc., have applied for GPT related trademarks, even the main a perma-built car Skyworth car, in June is also close to the current events to play the GPT this card.
The gale sweeping, do not engage in a big model of the car companies, the painting style has become strange instead.
In fact, the big model is not new, but under the "butterfly effect" of ChatGPT, the whole population on the model will be the key to open the second half of the new car-making intelligence, or is it another AI process "wolf" story?
This article will focus on the following three core points to make further analysis:
First: the primary market is still more in the wait-and-see, the fire of the big model is mainly reflected in the secondary market level, but the concept value shaping period is declining, and the technical background is gradually becoming the protagonist.
Second: the big model on the car is expected to change the valuation logic of the new automobile stocks, reshaping the host plant investment value.
Third: for the big model, the current relationship between technology companies and OEMs is somewhat like auto-pilot, the fusion of the fuzzy boundaries of the game, clarify the difference between "landing" and "value landing" will be the key to creating barriers in the process of the big model on the car.
Capital never sleeps?
From PC to mobile Internet, from internal combustion engine to new energy era, the transformation of which is the high light of human beings at the beginning of the twenty-first century, and it is also based on this technological change, which has led to the achievement of countless legends in the investment world.
But if we observe the time of the seam in the two transformations, AI can be called the old backstabber.
It is not difficult to find that every time after a big wave of change fades, the wave of AI will hit the beach again to make up for this intermittent silence. From smart speakers to VR, to blockchain, meta-universe, AI landing this "wolf" story told over and over again, more and more outrageous, but let a person more and more can not stop.
Why is it so? Liu Wen (a pseudonym), a senior investor, summarized to us in five words: "capital never sleeps".
After all, when there is nothing on the market to speculate, someone has to come up with something for everyone to speculate, a no one knows when it will become inevitable inevitable, in fact, is the charm of AI.
With the big model of the wind rolled into the automotive industry has also been a period of time, Liu Wen further to talk about the Prime said AI to share their own observations, "around the really booming, but the primary market is still very cautious about this wave. Whether it's domestic autonomous driving or the so-called burst of big models, it's actually not easy for AI startups to get financing today."
"Especially after a few waves of cold winter down in previous years, return prospects objective there, but more and more investors have understood that AI one is a big investment risk is extremely high, the process of cutting-edge projects is far slower than once expected, and secondly, the new concepts and tend to come too fast, research and study of the comprehensive threshold is also very high."
But why is the fire of the AI model still able to spread rapidly between car companies in a very short time? From the capital point of view, Liu Wen believes that this can not be separated from the support of the two core points, on the one hand, the social media, on the other hand, the secondary market, "I personally see the (big model) fire instead is reflected in these two places."
As he said, like from Musk's "Twitter rule division" to the beginning of the year ChatGPT burst red, today for the capital market, the power of social media has been not to be underestimated.
Recently reversed to the hand cramps "room temperature superconductivity" event, in fact, is a social media to lead the secondary market quality case.
For example, the A-share related concept plate, not even involved in the "room temperature superconductivity" related business, and never carried out related research and development and investment in the Farsheng, directly in this wave of chopping 6 days 4 boards, the entire superconductivity concept plate is also in just a few days time to ride a seemingly quite scribbling roller coaster.
Behind this naturally no lack of investors with short-term speculative thinking, but for AI big model of the soft technology track, there is also a layer of fundamentally different, that is, hard science and technology is often revolutionary, the opening is to have or not, but AI is in the soft track, is often a concept to start, and finally landed in the marathon.
Although not as room temperature superconductivity such as non-0 that is 1, but in the view of the Talking Engine said AI, the big model is divided into three phases to speak, one is the concept of value shaping period, followed by R & D show muscle, and ultimately return to the application of the landing competition.
With half a year past, the big model concept in the secondary market fire has gradually weakened, so car companies want to continue to rely on the concept of lifting the valuation of the operable space has been little. After the conceptualization period, strategic layout and in-depth research and development will be the new priority.
The core here is that, on the one hand, the current market related business performance as a whole is very difficult to compliment, on the other hand, the concept of the big model related stocks have generally been to a higher valuation, relying solely on the concept of shaping, it has been difficult to continue to frame this wave.
So all in all, from the AI big model to the recent room temperature superconductivity, can see the secondary market for a new wave of technological innovation how urgent the demand. However, the concept will eventually with the time constraints, just like capital never sleep, but also will not die a store bouncing, late inevitably still have to be supported by the final value of the landing.
Today for car companies, the wave of big models after the past, what exactly can be left on the beach? Will be the new stage is urgent to find the answer.
Through the conceptualization period, what is the value of big models?
In the view of AI, the value of big models for OEMs can be viewed from three directions: capital side, product side and production side, the latter two corresponding to the value of upgrading product power and reducing costs and increasing efficiency in the manufacturing side, respectively.
The capital side has been elaborated in detail in the previous article, and in this paragraph we might as well focus on the product and production side that needs to be attached to the capital value assessment.
First talk about the most important product, in the view of Talking Engine AI, car companies have to put the big model "forced" on the car, in addition to the capital on the wind, there is also a big driving factor is the big model in the value of the landing, can give the industry's valuation reshaping opportunities.
Specifically, the new energy era host plant profit model can be separated with two points to say, one is hard profit, the second is soft profit.
Hard profit is based on the manufacturing industry, look at the car sales, look at the car profit. New energy vehicles have actually been walking a smooth slow line, although there is no plunge, but even Tesla, BYD, etc., growth has been stretched.
The essence behind this is actually from the internal combustion engine to the three electric power of this "hard profit" growth curve, has no elasticity.
Based on this, we can well understand why the OEMs are piling up to start drumming up big models today, because this is an opportunity to renew soft earnings.
Soft profitability of this matter industry ow a lot of years, let's say the early analogous to cell phones and other consumer electronics and the birth of the car software payment ideas, or intelligent driving subscription, and even car companies based on the data value of cash UBI car insurance, can be described as a variety of things, but really pull out this line of soft profitability beautiful, is really very few corners of the earth.
Why is it so? To put it bluntly is that the soft product power has been insufficient to shake the market, like today's intelligent cockpit, although it has reshaped the traditional car, but compared to consumer electronics such as cell phones, can give consumers incremental experience is still very limited, feeding the hard product power, but it is difficult to provide incremental profit model.
Smart driving products are the same, the main problem is still demand, technology can be to L4 will naturally occur qualitative change, but before that, even Tesla, today FSD compared to the ability to sell the car's ability to absorb money is also completely incomparable.
Therefore, in the view of AI, if based on the big model, the future of the car end products can really break through the "worth" this paid subscription threshold, then the host plant valuation logic will be remodeled accordingly, which is also the key to winning the next big reshuffle of the automobile stock after the energy transition.
So what value can AI big model bring to soft profit? The main thing is smart cabin, smart driving and smart manufacturing.
For the intelligent cabin, the current mainstream car enterprises are similar to ChatGPT's voice interaction field, in addition, it can be extended to more other scenes of intelligent interaction, so that the 2C ascension product power.
For intelligent driving, it is expected to evolve from the traditional sub-modular processing to a large model processing that can be more generalized.
Let's say that Huawei, milli end Zhixing and so on have been trying to apply the data labeling, theoretically can be in the cloud link to reduce costs and increase efficiency, or through the virtual scene of deep excavation of long-tail data, and gradually overcome the ultimate bottleneck of the long tail.
Theoretically, if you look at the production side, you can also optimize the car building process based on AI models, like Tesla's integrated die-casting car building process innovation. Although this may seem a bit insignificant compared to the intelligent cabin intelligent driving, but just like the big model to the intelligent driving automatic data labeling possibilities, its cost reduction and efficiency potential should not be underestimated.
Sounds good, may wish to look at the current domestic car companies mainstream large model products, although most are still in the research and development stage, but there is a phenomenon is difficult to ignore, that is, a swarm of large models on the car, the product has not come out of the homogenization of the serious, basically the shadow of ChatGPT.
Like the set degree behind the Wenxin big model, ideal based on the big model algorithm self-research Mind GPT, or Geely AI big model function release, the most core is actually voice and other interaction scenes, which also gives the industry a short-term future will be a homogenization of the hidden worry of the fight.
There is also a noteworthy data, web analytics company Similarweb statistics show that in the first five months of 2023, ChatGPT global visits increased by 131.6%, 62.5%, 55.8%, 12.6%, 2.8%, respectively, the growth rate fell significantly, in June ChatGPT visits fell 9.7%, for the first time since its launch for the first time since.
The big brother ChatGPT is starting to smell bad, so how much more energy can be released by the followers who are coming forward? Therefore, in the view of AI, how to realize the "value landing" of the big model after the bubble is over will be the key to win the battle of car enterprises.
Landing ≠ value landing
Next, we might as well talk about the issue of "landing", compared to the "patchy" scramble to get on the car without barriers, how to take advantage of the AI model to really make the value of intelligence in the car to be reshaped? It seems to be a more worthwhile question.
Or continue to look at the status quo of the industry, as mentioned earlier, the current car model is in a R & D show muscle stage, which also contributes to a more or less self-talking style of painting.
For example, Geely "chaos of flowers" type of attack, in April of this year on the 2023 Ali Cloud Summit, Geely Automobile said, will be with Ali Cloud in the big model related scenes to start technical cooperation exploration and co-creation, first a table of progressive posture.
By the end of July, KAM family read in the internal management work conference revealed that Geely will be released in the second half of the year, "the industry's first full-stack self-research of the whole scene of the AI model", the relevant technology will be carried in the Geely Galaxy L6 application, and then show technical confidence.
On August 2, Baidu Apollo announced that Yigatong Technology, Geely's "pro son", has become one of the first batch of its exploration partners for the application of Wenxin large model smart cabin, some of the functions are expected to be the first to land on the Link, smart and other mass production models, all brothers under the five lakes and all over the world.
From Geely a foot to explore co-creation, a foot full-stack self-research repeated horizontal jump, we can interpret it as a harmonious co-creation, but also seems to be able to understand where it wants to mix a foot of impatience.
Or Baidu's AI products, previously Apollo in the landing process repeatedly halted, although today by the wind of the big model, Wenxin Yiyin partner ecology seems to flourish, but based on Wenxin Yiyin partner attributes, talk about the Prime said AI that it seems to also exist in the late stage of the landing of the difficult to robust hidden worries.
Specifically, in addition to Geely, a partner who has repeatedly jumped across the board, the Wenxin big model has also recently become the Great Wall Motor's exploratory partner, but this cooperation is worth an in-depth look.
On the one hand, Wenshin Big Model is undoubtedly going to empower Great Wall's Smart Cabin, but on the hardware level, Great Wall has developed its own V3.5 high-computing power cockpit platform, and on the software level, it is mainly co-created with Jiayu Intelligence and Xianlin Intelligence. Great Wall can be said to be one of the few domestic wisdom cabin technology full stack software and hardware self-research car enterprises.
Tianye check APP shareholder information shows that Jiayu Intelligence, Xianlin Intelligence two enterprises are the roots of the Great Wall's own son, especially Xianlin Intelligence, also in the field of deep plowing AI voice interaction.
It is not difficult to find, Baidu from Apollo to Wenxin Yiyin, it is difficult to say that its partner ecology are not concerned about the "body soul theory" of the Lord. So with a self-research faction to engage in co-creation, Baidu AI landing is to eat fat meat or drink soup? It seems to have to put a question mark.
On the other hand, cooperation often has progressive value, the new energy era, there is no lack of multi-faceted technology companies, like Mobileye in the early years of some of the smart driving program cooperation, their own chips will also be arranged by hand.
Baidu Wisdom Drive Wisdom module, although all, but and the Great Wall such self-developed school cooperation, late in the Wisdom Drive level of opportunity seems to be very slim, after all, the big model is also in full swing millimeters of this pro son, Baidu want to shake its position is obviously very difficult.
Overall, the relationship between car companies and technology companies today, compared to a pure co-creation of utopia, more like saving together to play mahjong. The surface of the happy, but in fact, each said each, the host plant emphasizes self-research, technology companies emphasize the partner ecology, both to all brothers under the five lakes, but also to be afraid of the soul of the cover is not in their own hands.
This can not help but think of the automatic driving from the "arms race" to the open integration stage of the course.
In the past two years, the traditional host plant full-stack self-research is difficult to run through, technology enterprises landing high demand, but even in this context step by step towards co-creation, today there will still be such as the host plant "forced" suppliers of white-box delivery of such games.
This is obviously not conducive to the development of the industry, such as the white box problem, often leading to technical enterprises to compromise on the low side, the OEMs are also very difficult in the white box requirements, so that the supplier to come up with the advantages of technology in good faith.
We might as well use this as an analogy to the late big models in the industry to land a battle, how to avoid this kind of similar "internal conflict" game in the future?
In the view of AI, the core is to clarify the difference between "landing" and "value landing".
Just like Huawei into the "car", early years seem more secure supplier model is difficult to run through, but the recent burst of control of the smart car mode is more and more popular. To put it bluntly, the smart car mode is not a pale landing, but based on the whole family bucket type intelligent program, in the landing to the target audience in the industry's scarce value.
So based on the AI big model, let's say the intelligent cockpit, if the future of each family is chasing ChatGPT, then even more novel technology to a homogenized market, will be like today's smart driving suppliers, the value of the infinite dilution.
So talk about the Prime said AI think, to the big model on the car to build landing value anchor point, either have to have technology scarcity, or to explore the scene scarcity. Technical scarcity is hard to come by, but might as well talk about scene scarcity.
Let's say today's new energy vehicle black technology endless, but many in the use of the link instead of chicken ribs, the problem here is to a certain extent in the technology to build the bottom at the same time, ignoring the adaptive value of the scene.
The car machine to do strong performance, for most people just need or navigation + music two big things, car companies have been equipped with voice interaction value in the end where? Seems to be more of a person have I also have to have the obsession, the so-called barriers, that is, in such a homogenization of the following layer of annihilation, to the involution.
Therefore, compared to the technical arms race, based on the advantages of large model-enabled interaction opportunities, the future to focus on the car scene to explore, derived from a high degree of adaptability of the scene of the product, is not lost as a line of thought.
Earlier this year in February, Tidu CEO Xia Yiping announced that Tidu automotive robots will integrate the full capabilities of Baidu Wenxin Yiyin.
Only half a year has passed, the wind of the big model will be swept to a large number of car companies with the momentum of a prairie fire. According to incomplete statistics, the current Azure, Xiaopeng, ideal, Great Wall, Geely, Chery, etc., have applied for GPT related trademarks, even the main a perma-built car Skyworth car, in June is also close to the current events to play the GPT this card.
The gale sweeping, do not engage in a big model of the car companies, the painting style has become strange instead.
In fact, the big model is not new, but under the "butterfly effect" of ChatGPT, the whole population on the model will be the key to open the second half of the new car-making intelligence, or is it another AI process "wolf" story?
This article will focus on the following three core points to make further analysis:
First: the primary market is still more in the wait-and-see, the fire of the big model is mainly reflected in the secondary market level, but the concept value shaping period is declining, and the technical background is gradually becoming the protagonist.
Second: the big model on the car is expected to change the valuation logic of the new automobile stocks, reshaping the host plant investment value.
Third: for the big model, the current relationship between technology companies and OEMs is somewhat like auto-pilot, the fusion of the fuzzy boundaries of the game, clarify the difference between "landing" and "value landing" will be the key to creating barriers in the process of the big model on the car.
Capital never sleeps?
From PC to mobile Internet, from internal combustion engine to new energy era, the transformation of which is the high light of human beings at the beginning of the twenty-first century, and it is also based on this technological change, which has led to the achievement of countless legends in the investment world.
But if we observe the time of the seam in the two transformations, AI can be called the old backstabber.
It is not difficult to find that every time after a big wave of change fades, the wave of AI will hit the beach again to make up for this intermittent silence. From smart speakers to VR, to blockchain, meta-universe, AI landing this "wolf" story told over and over again, more and more outrageous, but let a person more and more can not stop.
Why is it so? Liu Wen (a pseudonym), a senior investor, summarized to us in five words: "capital never sleeps".
After all, when there is nothing on the market to speculate, someone has to come up with something for everyone to speculate, a no one knows when it will become inevitable inevitable, in fact, is the charm of AI.
With the big model of the wind rolled into the automotive industry has also been a period of time, Liu Wen further to talk about the Prime said AI to share their own observations, "around the really booming, but the primary market is still very cautious about this wave. Whether it's domestic autonomous driving or the so-called burst of big models, it's actually not easy for AI startups to get financing today."
"Especially after a few waves of cold winter down in previous years, return prospects objective there, but more and more investors have understood that AI one is a big investment risk is extremely high, the process of cutting-edge projects is far slower than once expected, and secondly, the new concepts and tend to come too fast, research and study of the comprehensive threshold is also very high."
But why is the fire of the AI model still able to spread rapidly between car companies in a very short time? From the capital point of view, Liu Wen believes that this can not be separated from the support of the two core points, on the one hand, the social media, on the other hand, the secondary market, "I personally see the (big model) fire instead is reflected in these two places."
As he said, like from Musk's "Twitter rule division" to the beginning of the year ChatGPT burst red, today for the capital market, the power of social media has been not to be underestimated.
Recently reversed to the hand cramps "room temperature superconductivity" event, in fact, is a social media to lead the secondary market quality case.
For example, the A-share related concept plate, not even involved in the "room temperature superconductivity" related business, and never carried out related research and development and investment in the Farsheng, directly in this wave of chopping 6 days 4 boards, the entire superconductivity concept plate is also in just a few days time to ride a seemingly quite scribbling roller coaster.
Behind this naturally no lack of investors with short-term speculative thinking, but for AI big model of the soft technology track, there is also a layer of fundamentally different, that is, hard science and technology is often revolutionary, the opening is to have or not, but AI is in the soft track, is often a concept to start, and finally landed in the marathon.
Although not as room temperature superconductivity such as non-0 that is 1, but in the view of the Talking Engine said AI, the big model is divided into three phases to speak, one is the concept of value shaping period, followed by R & D show muscle, and ultimately return to the application of the landing competition.
With half a year past, the big model concept in the secondary market fire has gradually weakened, so car companies want to continue to rely on the concept of lifting the valuation of the operable space has been little. After the conceptualization period, strategic layout and in-depth research and development will be the new priority.
The core here is that, on the one hand, the current market related business performance as a whole is very difficult to compliment, on the other hand, the concept of the big model related stocks have generally been to a higher valuation, relying solely on the concept of shaping, it has been difficult to continue to frame this wave.
So all in all, from the AI big model to the recent room temperature superconductivity, can see the secondary market for a new wave of technological innovation how urgent the demand. However, the concept will eventually with the time constraints, just like capital never sleep, but also will not die a store bouncing, late inevitably still have to be supported by the final value of the landing.
Today for car companies, the wave of big models after the past, what exactly can be left on the beach? Will be the new stage is urgent to find the answer.
Through the conceptualization period, what is the value of big models?
In the view of AI, the value of big models for OEMs can be viewed from three directions: capital side, product side and production side, the latter two corresponding to the value of upgrading product power and reducing costs and increasing efficiency in the manufacturing side, respectively.
The capital side has been elaborated in detail in the previous article, and in this paragraph we might as well focus on the product and production side that needs to be attached to the capital value assessment.
First talk about the most important product, in the view of Talking Engine AI, car companies have to put the big model "forced" on the car, in addition to the capital on the wind, there is also a big driving factor is the big model in the value of the landing, can give the industry's valuation reshaping opportunities.
Specifically, the new energy era host plant profit model can be separated with two points to say, one is hard profit, the second is soft profit.
Hard profit is based on the manufacturing industry, look at the car sales, look at the car profit. New energy vehicles have actually been walking a smooth slow line, although there is no plunge, but even Tesla, BYD, etc., growth has been stretched.
The essence behind this is actually from the internal combustion engine to the three electric power of this "hard profit" growth curve, has no elasticity.
Based on this, we can well understand why the OEMs are piling up to start drumming up big models today, because this is an opportunity to renew soft earnings.
Soft profitability of this matter industry ow a lot of years, let's say the early analogous to cell phones and other consumer electronics and the birth of the car software payment ideas, or intelligent driving subscription, and even car companies based on the data value of cash UBI car insurance, can be described as a variety of things, but really pull out this line of soft profitability beautiful, is really very few corners of the earth.
Why is it so? To put it bluntly is that the soft product power has been insufficient to shake the market, like today's intelligent cockpit, although it has reshaped the traditional car, but compared to consumer electronics such as cell phones, can give consumers incremental experience is still very limited, feeding the hard product power, but it is difficult to provide incremental profit model.
Smart driving products are the same, the main problem is still demand, technology can be to L4 will naturally occur qualitative change, but before that, even Tesla, today FSD compared to the ability to sell the car's ability to absorb money is also completely incomparable.
Therefore, in the view of AI, if based on the big model, the future of the car end products can really break through the "worth" this paid subscription threshold, then the host plant valuation logic will be remodeled accordingly, which is also the key to winning the next big reshuffle of the automobile stock after the energy transition.
So what value can AI big model bring to soft profit? The main thing is smart cabin, smart driving and smart manufacturing.
For the intelligent cabin, the current mainstream car enterprises are similar to ChatGPT's voice interaction field, in addition, it can be extended to more other scenes of intelligent interaction, so that the 2C ascension product power.
For intelligent driving, it is expected to evolve from the traditional sub-modular processing to a large model processing that can be more generalized.
Let's say that Huawei, milli end Zhixing and so on have been trying to apply the data labeling, theoretically can be in the cloud link to reduce costs and increase efficiency, or through the virtual scene of deep excavation of long-tail data, and gradually overcome the ultimate bottleneck of the long tail.
Theoretically, if you look at the production side, you can also optimize the car building process based on AI models, like Tesla's integrated die-casting car building process innovation. Although this may seem a bit insignificant compared to the intelligent cabin intelligent driving, but just like the big model to the intelligent driving automatic data labeling possibilities, its cost reduction and efficiency potential should not be underestimated.
Sounds good, may wish to look at the current domestic car companies mainstream large model products, although most are still in the research and development stage, but there is a phenomenon is difficult to ignore, that is, a swarm of large models on the car, the product has not come out of the homogenization of the serious, basically the shadow of ChatGPT.
Like the set degree behind the Wenxin big model, ideal based on the big model algorithm self-research Mind GPT, or Geely AI big model function release, the most core is actually voice and other interaction scenes, which also gives the industry a short-term future will be a homogenization of the hidden worry of the fight.
There is also a noteworthy data, web analytics company Similarweb statistics show that in the first five months of 2023, ChatGPT global visits increased by 131.6%, 62.5%, 55.8%, 12.6%, 2.8%, respectively, the growth rate fell significantly, in June ChatGPT visits fell 9.7%, for the first time since its launch for the first time since.
The big brother ChatGPT is starting to smell bad, so how much more energy can be released by the followers who are coming forward? Therefore, in the view of AI, how to realize the "value landing" of the big model after the bubble is over will be the key to win the battle of car enterprises.
Landing ≠ value landing
Next, we might as well talk about the issue of "landing", compared to the "patchy" scramble to get on the car without barriers, how to take advantage of the AI model to really make the value of intelligence in the car to be reshaped? It seems to be a more worthwhile question.
Or continue to look at the status quo of the industry, as mentioned earlier, the current car model is in a R & D show muscle stage, which also contributes to a more or less self-talking style of painting.
For example, Geely "chaos of flowers" type of attack, in April of this year on the 2023 Ali Cloud Summit, Geely Automobile said, will be with Ali Cloud in the big model related scenes to start technical cooperation exploration and co-creation, first a table of progressive posture.
By the end of July, KAM family read in the internal management work conference revealed that Geely will be released in the second half of the year, "the industry's first full-stack self-research of the whole scene of the AI model", the relevant technology will be carried in the Geely Galaxy L6 application, and then show technical confidence.
On August 2, Baidu Apollo announced that Yigatong Technology, Geely's "pro son", has become one of the first batch of its exploration partners for the application of Wenxin large model smart cabin, some of the functions are expected to be the first to land on the Link, smart and other mass production models, all brothers under the five lakes and all over the world.
From Geely a foot to explore co-creation, a foot full-stack self-research repeated horizontal jump, we can interpret it as a harmonious co-creation, but also seems to be able to understand where it wants to mix a foot of impatience.
Or Baidu's AI products, previously Apollo in the landing process repeatedly halted, although today by the wind of the big model, Wenxin Yiyin partner ecology seems to flourish, but based on Wenxin Yiyin partner attributes, talk about the Prime said AI that it seems to also exist in the late stage of the landing of the difficult to robust hidden worries.
Specifically, in addition to Geely, a partner who has repeatedly jumped across the board, the Wenxin big model has also recently become the Great Wall Motor's exploratory partner, but this cooperation is worth an in-depth look.
On the one hand, Wenshin Big Model is undoubtedly going to empower Great Wall's Smart Cabin, but on the hardware level, Great Wall has developed its own V3.5 high-computing power cockpit platform, and on the software level, it is mainly co-created with Jiayu Intelligence and Xianlin Intelligence. Great Wall can be said to be one of the few domestic wisdom cabin technology full stack software and hardware self-research car enterprises.
Tianye check APP shareholder information shows that Jiayu Intelligence, Xianlin Intelligence two enterprises are the roots of the Great Wall's own son, especially Xianlin Intelligence, also in the field of deep plowing AI voice interaction.
It is not difficult to find, Baidu from Apollo to Wenxin Yiyin, it is difficult to say that its partner ecology are not concerned about the "body soul theory" of the Lord. So with a self-research faction to engage in co-creation, Baidu AI landing is to eat fat meat or drink soup? It seems to have to put a question mark.
On the other hand, cooperation often has progressive value, the new energy era, there is no lack of multi-faceted technology companies, like Mobileye in the early years of some of the smart driving program cooperation, their own chips will also be arranged by hand.
Baidu Wisdom Drive Wisdom module, although all, but and the Great Wall such self-developed school cooperation, late in the Wisdom Drive level of opportunity seems to be very slim, after all, the big model is also in full swing millimeters of this pro son, Baidu want to shake its position is obviously very difficult.
Overall, the relationship between car companies and technology companies today, compared to a pure co-creation of utopia, more like saving together to play mahjong. The surface of the happy, but in fact, each said each, the host plant emphasizes self-research, technology companies emphasize the partner ecology, both to all brothers under the five lakes, but also to be afraid of the soul of the cover is not in their own hands.
This can not help but think of the automatic driving from the "arms race" to the open integration stage of the course.
In the past two years, the traditional host plant full-stack self-research is difficult to run through, technology enterprises landing high demand, but even in this context step by step towards co-creation, today there will still be such as the host plant "forced" suppliers of white-box delivery of such games.
This is obviously not conducive to the development of the industry, such as the white box problem, often leading to technical enterprises to compromise on the low side, the OEMs are also very difficult in the white box requirements, so that the supplier to come up with the advantages of technology in good faith.
We might as well use this as an analogy to the late big models in the industry to land a battle, how to avoid this kind of similar "internal conflict" game in the future?
In the view of AI, the core is to clarify the difference between "landing" and "value landing".
Just like Huawei into the "car", early years seem more secure supplier model is difficult to run through, but the recent burst of control of the smart car mode is more and more popular. To put it bluntly, the smart car mode is not a pale landing, but based on the whole family bucket type intelligent program, in the landing to the target audience in the industry's scarce value.
So based on the AI big model, let's say the intelligent cockpit, if the future of each family is chasing ChatGPT, then even more novel technology to a homogenized market, will be like today's smart driving suppliers, the value of the infinite dilution.
So talk about the Prime said AI think, to the big model on the car to build landing value anchor point, either have to have technology scarcity, or to explore the scene scarcity. Technical scarcity is hard to come by, but might as well talk about scene scarcity.
Let's say today's new energy vehicle black technology endless, but many in the use of the link instead of chicken ribs, the problem here is to a certain extent in the technology to build the bottom at the same time, ignoring the adaptive value of the scene.
The car machine to do strong performance, for most people just need or navigation + music two big things, car companies have been equipped with voice interaction value in the end where? Seems to be more of a person have I also have to have the obsession, the so-called barriers, that is, in such a homogenization of the following layer of annihilation, to the involution.
Therefore, compared to the technical arms race, based on the advantages of large model-enabled interaction opportunities, the future to focus on the car scene to explore, derived from a high degree of adaptability of the scene of the product, is not lost as a line of thought.
No comments yet