news

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

author | zhang yongyi

editor | jingyu

at one o'clock in the morning on september 26th, beijing time, the annual meta connect was officially held. in addition to the latest quest series headsets and ray-ban meta series product line updates, meta publicly demonstrated its more than ten years of research and development in the ar field for the first time. the result - ar glasses prototype orion.

using the form of traditional glasses to challenge the performance and experience that previous mr headsets could only have, the emergence of orion is tantamount to a blockbuster for the entire industry, triggering as much topic and discussion as the apple vision pro.

as an important player in the domestic ar industry, thunderbird innovation has naturally become the subject of much attention from the outside world, whether it is its views on meta's new products or its future hardware product planning.

on the evening of september 27, zhang peng, founder and president of geek park, and li hongwei, founder and ceo of thunderbird innovation, conducted a live broadcast on the geek park wechat public account to discuss the orion prototype announced at this meta connect conference, as well as thunderbird innovation the layout and planning of the ar glasses product field were discussed in detail.

according to the information disclosed at this meta connect conference, orion has integrated almost all the most cutting-edge solutions in the ar industry: microled displays, optical waveguide solutions based on silicon carbide materials, end-side integration of ai functions, and "symbiosis or replacement" with mobile phones. ” discussion of the relationship. in fact, the emergence of the "next generation general computing platform" is already imminent.

li hongwei believes that the orion effect looks cool, but it is still a "product three years later." if orion is the ar version of "apple vision pro", the thunderbird innovation team, which has been laying out the same technical route for many years, hopes to create a chinese version of "vision" that consumers can buy within the year.

as one of the industry leaders in the ar industry, li hongwei believes that chinese ar brands, including thunderbird, have great potential in integrating product and ai capabilities before this opportunity that can redefine "personal general-purpose computing devices" in the next decade. unique advantages that are not lost to western counterparts.

the field of ar glasses not only requires looking up at the stars, but also needs to walk step by step. through deeper integration of ar and ai functions with users' daily lives, ar glasses can truly penetrate into users' daily lives.

in this live broadcast, there was not only discussion and technical analysis of new ar glasses products, but also li hongwei disclosed the new product plan in public for the first time: thunderbird innovation will not only produce ar glasses products in the form of orion in the future, but will also release ar glasses products with a form similar to ray -ban meta’s ai smart glasses product.

before the "next decade" of ar glasses officially begins, through orion and thunderbird x3, how far into the future can we see now?

some interesting points:

the core of meta's release of the orion product is to prove to everyone that everyone's imagination of ar glasses can be realized.

the main way for humans to obtain information, about 80% of information is obtained through vision. with the development of display technology, this proportion may be higher.

the underlying logic of ar is to integrate reality and reality, which determines that ar glasses optical display has a recognized ultimate solution - micro led + optical waveguide technology.

orion requires 7 cameras, which does not mean that all ar glasses require 7.

it is almost 100% certain that meta orion will not have a purchasable version next year. but thunderbird x3 will go into mass production by the end of this year.

orion is the vision pro of ar glasses, while thunderbird x3 is more like vision.

in the field of ar, it is possible that chinese companies and american companies will lead the development of the industry together and bring about an industry explosion. it is even possible that chinese companies will play this role earlier or to a greater extent.

the following is the text of the live broadcast, compiled by geek park.

01

orion proof

the possibilities of ar imagination

zhang peng:i believemagnificent for sureyou have watched this meta conference for the first time. what is your impression of this conference?

li hongwei:first of all, i think this conference is quite interesting, and the on-site interactive display is very effective. zuckerberg also said that there were bound to be problems with the presentation, but the basic presentation went smoothly. as practitioners in the ar industry, we are looking forward to the release of orion products. we've known a lot about the product in the past, but it's still exciting to see it launch. this is a very important moment for the ar industry. it proves that everyone’s imaginations about ar can be realized.

zhang peng:for industry insiders like hongwei, there may be nothing new. i believe hongwei already knew what they were going to say before the meeting, but we were still kept in the dark. while watching this conference, we think there are three aspects worth paying attention to.

first,VRthere is no significant improvement in the technology of the device, but a cheaper device is launched, priced at $299. we speculate that this may mean that meta hopes to promote the equipment and promote a healthier development of the vr content ecosystem. it doesn't try too hard on a vr device, making it lighter, smaller, and with a better display.

also ray-ban meta the new generation of products - audio glasses plus camera, does not seem to emphasize hardware upgrades, but rather demonstratesAIapplications on wearable devices. although it does not have a display function, it uses sound as a means of interaction and delivery to demonstrate how to better utilize ai technology. this is actually very clear and this trend can be seen. it is said that it is also rising rapidly in china, and it is expected that many related products will appear in the country next year.

of course, everyone's attention is focused more on the orion device. even after it was released, i didn't understand much because it wasn't immediately available for purchase like the vision pro. we heard that the orion could cost tens of thousands of dollars and felt like it was made by hand.

orion equipment diagram｜image source: meta

what is meta trying to say by releasing a product like this at this stage? maybe in order to enhance everyone's understanding ofARconfidence, but is there any special emphasis on it? how do you understand the product orion? what stage is it in? why is meta showing us this product now that we can’t use? what’s the intention?

li hongwei:well, of course we are also guessing, and we may know a little more information (than everyone else). i personally think that when meta releases this product,the most important thing is to prove to everyone that everyone has AR everything you dream about with glasses is achievable. technology has made breakthroughs and is now possible. on the one hand, it is to enhance industry confidence, and on the other hand, it is to promote the development of the entire industry. meta launched this product to drive the rapid development of the entire industry chain. the main reason why orion cannot be released is the high price, which is closely related to the maturity of the supply chain. the most expensive part is the silicon carbide optical waveguide. price reduction requires the joint efforts of the entire industry to increase the use of silicon carbide optical waveguides and expand more application scenarios, so that the cost will fall.

of course, meta has invested heavily in the xr field. meta is more optimistic about ar than vr. zuckerberg has mentioned in many interviews that ar is the next generation of general computing platform and is expected to create a new internet ecosystem that is equivalent to the size of the mobile phone market. meta has been working hard to realize this vision. r&d investment in orion began long before quest 2 was launched. after a long period of polishing, meta believes it is time to show it to the public. if a company, team or individual has been developing a product for ten years and really feels that it is doing well, then it is a reasonable decision to show it now.

zhang peng:the points mentioned by hongwei are very important, and meta hopes to enhance the confidence of the entire industry by displaying orion. i understand very well that confidence in the supply chain is critical for products to be cost effective and available. although apple's vision pro is expensive, its release quickly increased the confidence of the industry chain. as long as the giants say they want to do this, this direction is what they agree with and it will definitely have traction.

let’s take a deeper look. what technologies have been proven to be feasible in the orion product? which technologies are meta to implementARthe vision must break through and has now reached a mature stage?

li hongwei:well, there are several aspects. in fact, we already discussed the first issue when we launched the thunderbird x2 last year. it must first be in the form of glasses. it should be wireless, with a built-in chip and battery, so that people can wear it comfortably for long periods of time.

secondly, when wearing glasses, they need to be able to see the real world clearly and display clear holographic images. if the user cannot see the real world, then it is a vr device; if the user can only see reality without image display, then it is not augmented reality, so the two need to be combined. we hope that when seeing the real world, there will be no visual occlusion or loss. at the same time, when displaying images, the picture should be clear enough, large enough, and 3d holographic. so there is a lot to do in optical display technology.

third, as for the best human-computer interaction method on glasses, everyone generally agrees that eye movements plus gestures are a very natural and efficient interaction method. vision pro also adopts this interaction method. the key is whether it can be used in glasses. realize eye movement + gesture interaction in a compact form. this is an important question because vision pro uses a large number of cameras for gesture interaction and has an external battery in the form of a helmet. orion proved it could achieve similar functionality in glasses.

fourth, this kind of glasses has strong capabilities. it can not only realize a series of scenes on the mobile phone, but also achieve virtual reality integration. can its chip, power consumption, and heating problems be solved? this is also very important. in fact, there are a lot of customized components, such as self-defined chips. from an industry perspective, they have significantly reduced the power consumption of the core virtual and real focus combination algorithm.

of course, another important aspect is the application ecology. the above things constitute whether ar glasses can be realized. at the moment, orion isn't doing much on that last point and may have some minor games. but i think for any industry, we must first have a very good complete product or a simple hardware product, and then promote the development of the ecosystem. this is also the general law of industry development.

02

ar glasses,

next generation general computing platform

zhang peng:just nowyou definedorion the meaning of this product. before dismantling orion in detail, i think we need to clarify a premise: what isARglasses? zuckerberg’s huge investment in reality labs shows thatAIeyewear plays an important role in meta's strategy. at the connect conference, meta not only displayed ai glasses, but also products such as ray-ban meta. although they are different in form, they have significant differences in function.

therefore, i would like to pleaseyoufirst define for usARthe concept of eyewear, and why meta is so heavily involved in this field.

li hongwei:meta puts so much effort into ar glasses. i think the underlying reason is that ar glasses are the next generation of general computing platform - it is the next generation of smartphones. zuckerberg once mentioned this in some podcasts. he said that when we started facebook, the mobile phone industry was basically relatively mature. the establishment of their facebook was basically synchronized with the development of smartphones. he believed that he had missed the opportunity to build a universal computing platform for mobile phones.

so when facebook was relatively well established, it began to lay out the next generation of general computing platforms. ar glasses are considered to be the next generation of universal computing platform, a product that will affect everyone around the world, which has a significant commercial impact on traffic distribution. as an app matrix company, meta's app distribution is actually affected by smartphone companies such as apple and google. so it makes perfect sense that meta would want to build its own next-generation general-purpose computing platform.

let’s talk in detail about why ar glasses are the next generation general computing platform. if we draw a line, the horizontal axis is virtual/virtual reality integration, and the vertical axis is the presence or absence of ai, we can find a pattern: pcs and mobile phones in the past had two important characteristics.

first, the virtual world is separated from the real world. take a mobile phone as an example. when navigating, your mobile phone may indicate that your destination is the building in front of you, but when you face three unmarked buildings, you cannot determine which building is the destination. when at an intersection, if the gps signal is not accurate enough, it is often difficult for us to determine exactly where we are at the intersection. this is because there is a gap between the virtual digital world and the real world we live in. i have to manually map them together, which undoubtedly increases the cost of use.

however, if we can achieve the integration of the virtual world and the real world, we can elevate the entire internet ecosystem to a new level and seamlessly integrate virtuality into reality. take navigation as an example. when we walk, a traditional navigation app might display a digital arrow on the screen to tell us where to turn. but if there are three unmarked buildings in front of us, we can't be sure which one is our destination. in this case, ar glasses can directly put virtual names on each building and can also display more information. this integration of virtual reality is extremely important.thunderbirdinnovationthe mission is "imagination is reality, space is smarter", that's what it means. this means that through the integration of virtual reality, we can bring the things in our imagination to the real world. we know that the stars we see on tv or mobile phones are just images, but if we wear ar glasses and see the virtual image of the star appear in front of us, it will appear more real because it is placed in the real environment. middle.

as discussed earlier, ar technology enables us to see labels on buildings that load automatically. this is a dimension, the isolation of reality and virtuality, or the integration. there is another dimension that is the dimension of ai.

although pcs and mobile phones in the past had certain computing capabilities, they were not designed for ai. today, with the development of ai technology, we have seen the emergence of various ai pcs and smartphones. however, ai is powerful, what is the best carrier? it should actually be glasses.

ar glasses can see what you see, hear what you hear, and input what we see and hear into the ai ​​system. after processing by the ai ​​system, the results can be displayed in front of your eyes and ears in real time. this interaction does not require us to interrupt our current activities to take out our mobile phones or other devices. so if we construct a quadrant diagram, we will find that ar glasses and ar glasses with integrated ai are located in the upper right corner of this quadrant diagram. with the integration of ai and ar, it not only makes past devices more efficient, it completely creates a different, upgraded, and richer scene ecosystem. this is the most important reason why ar glasses serve as a general computing platform.

internet prophet kevin kelly explained this very clearly in his latest book "the world in 5000 days", "we are about to usher in a new huge platform, and everything in the world will be connected to ai. digital world perfectly integrated with the real world, this is what we call a new platform for ar or the mirror world.”

of course, if we consider ar to be the next generation general computing platform, then the next steps will become very important. how do we build the next generation of universal computing platforms? of course, what is very important here is the issue of technological breakthroughs. the demand has always been there - in the past, the potential applications of ar glasses have been widely discussed and anticipated in movies, novels and other media. therefore, it is particularly critical to achieve these technological breakthroughs, and meta, including us, has made significant progress in this regard. the above is the value of the technical capabilities demonstrated by products like orion.

zhang peng:from what i hear, i can’t infer that zuckerberg had such a plan when he acquired oculus ten years ago, but because meta AI investment in the field of xr and xr is long-term and lasts for ten years, which shows that they are indeed fighting for the next generation of smart devices.

after all, i missed the iphone era and there is no way to compare with steve jobs, but i can still try the next time.

li hongwei:yes.

03

ar is

true "fusion of reality and reality"

zhang peng:zuckerberg first started with VR，metaversestarting to practice, he hopes to allow everyone to enter a digital world with deeper access and a stronger sense of experience.who are youdeny consentthis point of view is both ray-ban meta and orion areARdifferent manifestations of technology, they constitute a light and heavy combination, ray-ban meta may represent a transitional stage, and orion may represent the ultimate goal.

another question is, why is it so complicated to show this? how is the display technology used in orion, which everyone has been working hard on for so many years? can the cost of this technology become lower and lower in the future? how long will it take before it can truly compete with mobile phones for personal computing centers?

li hongwei:i agree with brother peng that even glasses without a display function can be regarded as an augmented reality device. however, if we define ar more strictly, we may think that only technologies that seamlessly integrate virtual information and the real world and see them as one from the user's perspective can be called augmented reality.

this is a matter of definitional rigor. if we consider technology that senses reality and provides reminders as augmented reality, that's a broad definition that i think is acceptable. but by this definition, mobile map apps can also be considered a form of augmented reality because they provide us with additional information. when we view electronic maps, whether it is google maps or amap, they provide an enhancement of reality and show us more information. this is a question of definition and distinction.

another point is that industry development does not often happen overnight, but is a gradual development process. the process often begins with imagining a future that includes a variety of technologies. these technologies mature at different speeds, and the first to mature technologies combined with existing user needs will produce so-called realistic products that "lay eggs along the way." these products will continue to iterate technology, and as the technology gradually matures, they will eventually reach a better final state. so the glasses with cameras and audio functions you mentioned are actually very important, and they have their own user value. i think that if such glasses appear first, or if such glasses become the mainstream first, they will soon transition to ar glasses becoming the mainstream.

as for why display functionality is important, it's closely related to what we discussed earlier. how do we merge virtuality with reality? how can a new general-purpose computing platform create experiences and application scenarios that were not possible in the past? for a new device, whether it is a general computing platform or non-universal, to be accepted by users, it needs to do two things: first, it can complete the scenarios that users have been accustomed to doing on other devices in the past, but the best could do better. second, it needs to be able to create some new scenes that were almost impossible to achieve in the past, or the experience was very poor.

let me give you an example that is not necessarily appropriate. take a smart watch as an example. it not only simplifies functions such as checking the time and wechat, but we don’t need to take out the phone and can see it by raising our wrist. but another thing that mobile phones couldn’t do in the past is health monitoring. only when a smartwatch has these two capabilities can it become a widely accepted product. otherwise, users may be hesitant to switch to a new device because they may feel that the functionality of the new device is something they can already achieve on their old device. so, what can ar glasses do that mobile platforms cannot? one of them is "see what others see and listen to what others hear." this is important.

another point is that we can hardly rely on mobile phones to put reality and virtuality together. when we look at the real world through a mobile phone screen, virtual information is just a layer on the screen and we don’t think it is real.

but ar glasses can achieve this virtual and real fusion effect, so only glasses equipped with display functions can truly achieve this. this is also the reason why the industry has a stricter definition of ar glasses. meta didn't call its previous ray-ban meta audio glasses ar glasses, but the orion is considered a true ar glasses based on this core logic.

another important reason is that vision is the main way for humans to obtain information. about 80% of information is obtained visually,with the development of display technology, this proportion may be higher. i think it may be 5%, 85, or even more than 90%. this means that display is the most efficient way to interact。

in meta's demonstration, there was a particularly interesting scene. i wonder if you remember zuckerberg's demonstration of real-time conversation translation with an athlete. the other party spoke spanish and zuckerberg spoke english. the process was very smooth and the effect was very good. but one problem is that after the other party has finished speaking, the large model system has to speak again, which will lengthen the time.

if the text is displayed in real time when the other party speaks, the problem of time extension can be avoided. after the other party has finished speaking, i have to wait for him to respond, which will be a little embarrassing. however, after the other party has finished speaking, i respond to him immediately, which is a better effect. therefore, for new products, the improvement of display technology is crucial.

we basically believe that only devices with efficient displays, whether glasses or traditional pcs or mobile phones, can become a universal computing platform. display technology is the key to human-computer interaction, without which it is difficult to achieve widespread applications.

04

orion is the most importantof

the role is industry consensus

li hongwei:the second question is why display technology is so complex in implementation, as well as its current development status and future development trends. because this kind of display is indeed a very special kind, and its technical principles or requirements themselves are relatively complex. for example, when we wear glasses, we hope that the real world scene will not be affected at all without turning on the glasses, which requires a completely transparent visual experience. but at the same time, once the glasses are turned on, we expect it to provide a high-quality image, which should be three-dimensional, deep, full-color, and have a very large picture amplitude.

let me break down this technology a little bit. there are several key issues we need to address.

the first question is about the making of the lens, whether it's glass or some other material. lenses need to be able to see through the real world while also displaying it. currently, orion uses surface relief diffractive light guide technology. simply put, this method is to carve specific patterns on the lens. when light passes through these patterns, it can produce a virtual image through complex reflection and refraction. .

in this way, we are able to see the real world and the displayed content simultaneously on the lens.

additionally, we need a technology to generate these images. since there are only textures on the lens, we need a tiny light-emitting element. this component requires it to have a very high pixel density, which means its size will be very small. currently, micro led's pixels are spaced just two to four microns apart, giving it an extremely high pixel density, some of the highest we've seen in a product to date. because of this, we can generate clear 1080p or 720p images at a very small size.

thunderbird’s previous generation microled light engine | thunderbird innovation

in terms of display technology for ar glasses, we need full-color display with extremely high brightness. because light is usually brighter in the real world, the eye brightness of ar glasses needs to be at least 1,500 nits, or even more than 2,000 nits. however, the material of the lens is not necessarily glass, and its texture design may cause a large loss of light energy, so we need more powerful light-emitting devices. if the light energy utilization rate of the light-emitting device is only 1%, then in order to achieve a visible brightness of more than 2000 nits, the brightness of the light-emitting device needs to be increased by 100 times.

these two problems together determine that ar glasses optical display technology needs a recognized ultimate solution.this solution includes micro led as the display device and optical waveguide technology as the core technology of the lens. to what extent has the technology achieved so far? why are thunderbird’s innovative products and meta’s products so important? currently, only thunderbird innovations and meta have officially released such products. of course, meta has not yet been mass-produced, but thunderbird has innovatively mass-produced products that use all micro leds.

meta's orion product has a big advantage, that is, its waveguide (fov) is very large.think about it in the past AR it is difficult to achieve a large fov in glasses in the form of glasses, but meta uses a very special technology, or it can be said to be a very interesting material - silicon carbide waveguide, so being able to achieve a 70-degree fov is a very important thing. it is an important breakthrough in optical display. we have seen a relatively important breakthrough in optical display.

last year's thunderbird x2, this year's x3 and meta's orion all proved that what we said is feasible. why is it possible? first of all, the technology is feasible. both products use microled optical waveguide technology and are on the right technical path. secondly, their brightness, fov, display effect and full color have all reached a basically usable state.

zhang peng:can we say now that in AR is the "ultimate solution" already in place? is the microled plus light waveguide solution the "ultimate solution" we can see for ar devices in the future? especially for ar devices that hope to become "personal computing centers", has the display solution been basically determined?

li hongwei:for leading companies in the industry, especially for leading companies in technology, the choice of this path is almost always clear.we need a strong light-emitting device with very high brightness, very small size, very low power consumption, and very high refresh rate. there is no other better option now or even in the next few years. among them, although technologies like lcos have advantages, their power consumption is still relatively high, and it is difficult to achieve particularly low power consumption. therefore, almost all options are in the direction of "seeing reality through lenses and displaying images at the same time." currently, there is no other alternative but microled plus light waveguide. it’s just that every technology is iterating. for example, optical waveguides have different technologies and processes, such as imprinting, etching, etc. of course, there are still different voices in the industry. some people think that this technology is not mature enough yet, and many companies may not be able to truly produce microled and optical waveguide technologies and products at this stage, so there are many different voices on the market. but thunderbird innovation and meta have already made corresponding products. this technology is feasible and is not as far away as everyone thinks. if you look at the leading technology companies in the united states, everyone basically has a consensus, and there are relatively few non-consensus.

zhang peng:this time meta may have significantly strengthened their voice on this line. optical waveguide plus microled this path is essentially determined by its goal. if we wish AR to become a universal personal computing platform, to effectively integrate reality and virtuality, and to be wearable for a long time, all these goals will comprehensively determine a certain technical route.teamit is often necessary to find a so-called "first principle" under a clear goal-what problem must be solved under this goal.

li hongwei:right.

zhang peng:another core highlight of meta orion this time is that it has 7 cameras. vision pro also has a large number of cameras, which may involve the understanding and recognition of future micro-interactions, and may also see its perception of the world. we just talked about that the perception ability of ray-ban meta’s camera is particularly important to its future, because AI you need to use the camera to input information about real events.AR without such ability, there is no real superposition of the real world, because it cannot see the real world.

7 cameras, more or less? what are the future trends? after all, the more cameras, the higher the cost and the more power consumption. how to understand the significance and future of these 7 cameras? exist AR in terms of the development of the device's external perception capabilities, what other technical steps are gradually becoming clear? because behind this release, we should all be able to see some trends leading to the future.

li hongwei: let me first explain the camera problem. orion is a product that competes with vision pro, although the cost is much more expensive than vision pro.AR glasses require more cameras based on application scenarios and user needs.first of all, if we want to see reality, take photos and videos, we need a color camera; we also need to perceive reality efficiently, and a black and white camera is better here; we also need eye tracking, which requires an introverted camera or other technology. therefore, these three cameras for different purposes are necessary. however, orion requires 7 cameras, which does not mean that all ar glasses require 7.

even if these three functions are realized, it does not necessarily require 7 cameras, maybe 5. for example, using an rgb camera plus two slam cameras, plus an internal eye-tracking camera, the same effect can probably be achieved.

therefore, from a product perspective, the choice of the number of cameras must first be based on user needs, and secondly, achieve a technical balance. if you think of orion simply and crudely as "vision pro", everyone can understand it. orion's approach isn't the only viable option, and we don't necessarily need that many. in a sense, thunderbird x3 is like a "vision", not vision pro. the advantage of vision is that it can be accepted by more users, and it does not have so many cameras.

in terms of other technologies, there are several key aspects.

the first one is in human-computer interaction. of course, this is related to the camera. how can we use gestures and eye movements to interact with the device? in fact, eye movements plus gestures are a better way of interaction. at present, this technology is generally mature. although it is not that good yet, it is not much different. thunderbird innovation has also done something in this regard. more work.

another important point is the chip. chips are very critical to the entire system, and the release of orion has given the industry great confidence. in the past, everyone has been worried about the power consumption of glasses, but the custom chip allows orion to consume much less power in specific scenarios than we expected. this also proves it.

there are other technical issues, such as heat dissipation. one advantage of silicon carbide materials is high thermal conductivity, which can effectively help heat dissipation. in addition, for example, in the design of eyeglass frames, how can it be light and functional?

there are still many problems to be overcome, including of course a very important part of ar glasses, natural and multi-modal understanding and interaction. currently, multiple large models can already be achieved. therefore, these key technologies have already made essential breakthroughs and require further optimization in engineering.

zhang peng: regarding interaction, you just mentioned that with so many cameras, first of all, they require a large amount of spatial sensing data calculations. at the same time, eye-tracking cameras should be standard equipment in the future. there must be cameras watching the movement of our eyeballs; there must also be cameras capable of forward sensing of the outside world. but cameras that focus on interactions are also indispensable, because it is difficult to imagine that all interactions require speaking to the glasses with your mouth.。

this time meta is like in thunderbirdinnovationgoing further in the direction previously set? does it represent the future? AR is there an opportunity for the glasses line to develop more clearly? when you see that meta is also consistent with your route, do you feel that "supply chain costs are expected to decrease in the future" and that this interaction model is more acceptable to users?

li hongwei:yes, i also talked with my meta classmates in the afternoon. although i had a general idea of ​​their direction, i finally confirmed it at the press conference. i think there are two reasons why i am excited. one is that meta has chosen the same technical path as thunderbird innovation. it seems that we are thinking right. the second reason is that everyone is doing the same thing, including silicon carbide. thunderbird innovation has also made a series of explorations, which will play a very important role in cost reduction of the industrial chain and supply chain, so this is a very good signal. but,i would like to explain a little bit more about how human-computer interaction is a kind of technology anduser experiencea more balanced approach?

the current main method is eye movement plus gesture interaction. there are two technologies for eye movement, one uses a camera, and the other uses infrared or mems (microelectromechanical systems) sensing, which consumes less power. but this question is relatively simple, so i won’t elaborate further here.

i think there are two broad categories of gesture interactions that are important. the first type is the "invisible hand" interaction method (downward interaction). the hand can be freely placed on the leg, behind the back of the hand, or anywhere else, and can be easily interacted by tapping or sliding. the second type is the "visible hand" interaction method (forward interaction). for example, when looking at a photo, you can use gestures to zoom in directly. both types of interactions are common. for the first type of interaction, a wristband or smartwatch is a good choice.

zhang peng:that’s myoelectric signal, right?？

li hongwei:yes, but in a few different ways. myoelectric signals are a more sensitive method. some smart watches in the past combined with similar technologies can also achieve a certain degree of gesture interaction. forward interaction definitely requires a certain camera, because it requires some interaction with real or virtual images. but if the interaction is from downwards, we may not necessarily need so many cameras. for example, we do not necessarily need a diagonally downward camera, because if the diagonally downward camera is not very powerful, it does not have as powerful technology as vision pro, and it is very it is difficult to capture small hand movements in various complex scenes. in the end, you may have to use devices such as myoelectric bracelets or watches, but these devices can already do it.

when you definitely need cameras, they are often in a relatively forward position, and you may not need to specifically configure a downward-facing camera. this is a matter of balance and selection that needs to be considered carefully when it comes to products. vision pro must do this because its goal is that users do not need to wear additional auxiliary equipment, so of course it must have a very powerful camera that can look in all directions. but glasses don’t have to be.

zhang peng:when will a device like orion actually hit the market? including what you are doingthunderbird x3, i have seen some information before, and it seems to be very close to mass production. how long do you think it will take before meta's orion can be seen by everyone today and become a product in the future? will what meta says today become purchasable next year?

li hongwei:if you are asking about a product that is almost exactly the same as orion, then meta may indeed need a little more time,it's almost 100% certain that meta won't have a purchasable version next year.but if you're asking about thunderbird, our x3 will go into mass production by the end of this year.

05

thunderbirds to do

ar version of “vision”

zhang peng:what's thunderbird's plan?

li hongwei:let me first add, why do you think meta will definitely not be released next year?as a large company, they have invested a lot in this area, but their actions will be relatively slow.as a start-up company with the resources of a large company, thunderbird innovation will have more resources and move faster. therefore, thunderbird’s x3 is expected to be launched between the end of the year and the lunar new year. i also want to make it as possible as possible a little earlier.

many people may be concerned about when this type of product will be launched, and when orion and thunderbird innovation will be available, so we will introduce this product in relatively detail. let me make an analogy:orion is AR the vision pro in the glasses world, and the thunderbird x3 is more like the vision.i think this metaphor is quite vivid. apple's strategy is to launch vision pro first, but they also believe that vision is a product that can be widely accepted by everyone. so we are very grateful to meta, half jokingly, for helping us release a pro version of thunderbird x3.

next, we are preparing to officially launch x3. the x3 is probably a better product in terms of product and technology balance. the x3 comes in two versions, weighing around 60 to 70 grams, while the orion weighs close to 100 grams. 60 grams 70 grams is a very important indicator line. it is not just a matter of 30 to 40 grams more, because the glasses and sunglasses everyone wears are basically about 60 grams, so 60 grams is a state that is more acceptable to users. in addition, thunderbird x3, like orion, is a full-color microled optical waveguide product.

i would also like to say a few advantages of x3. in some aspects, our performance is better than orion, such as eye brightness, which is very important. if you look closely at the video, you'll notice that the orion's brightness is relatively low, i'd guess around 500 nits, maybe even lower. in other words, under strong light, especially outdoors at noon, it may be difficult to see the screen content clearly. xiao zha himself said that increasing the brightness is a very important task for them to do. the brightness of thunderbird x3 is at least above 2000 nits, and may even reach above 3000 nits.

i may have revealed a little too much (laughs). it’s easier to explain things clearly while orion is still being discussed.

thunderbird is brighter than orion, but the fov will be slightly smaller, the resolution may also be smaller, but the ppd will be higher. people who know vr will be more aware of the concept of ppd. it refers to the angular resolution, which determines whether your picture quality is clear or unclear, delicate or not. in addition, in terms of human-computer interaction, thunderbird x3 will also support gesture interaction.

thunderbird x3 and its myoelectric bracelet in internal testing｜image source: thunderbird innovation

we will also use devices such as bracelets and watches to achieve similar gesture interactions. although we haven't seen orion's particularly detailed gesture interaction information, i think it basically uses a similar human-computer interaction method. regarding eye movement interaction, we are also doing detailed testing internally.

thunderbird innovations tests eye-movement interaction function on glasses

in this way, when the thunderbird x3 is launched, we will not fully implement gesture and eye movement interaction. there is a high probability that the gesture interaction function will be launched first, and eye movement interaction may be launched later.

in addition to human-computer interaction, we have also made a lot of optimized designs in application scenarios and chip architecture. for example, the battery capacity of thunderbird x3 is smaller than that of orion, but the usage time may be similar, which also depends on different usage scenarios. therefore, thunderbird x3 is a very leading product in terms of optical display, and is similar to orion in terms of basic technical performance. of course, we did not use silicon carbide material because the price of this material is not suitable yet.

moreover, there is a very important difference between the positioning of thunderbird x3 and orion, which is the difference between vision pro and vision that i just mentioned. therefore, it is a very leading product on the market, but at the same time it can be mass-produced. thunderbird innovation not only has the design capabilities in optical technology, but also the production capabilities, so we can mass-produce this design. maybe our display screen size is not as large as orion's, but this is a more appropriate state that meets specific application scenarios. eventually, everyone will see that not only are the basic functions present, but the glasses look very much like normal glasses.

i don’t know how everyone feels when they see the picture now. if you look at it from a certain angle, orion's lens is relatively thick because of the part from the light source to the waveguide. thunderbird innovatively adopts a light source design from the inside out, so the lenses are relatively thin and the overall look is like a pair of ordinary glasses.

i would also like to point out that thunderbird innovation is a little different from meta in its handling of full-color microled technology. we judge that meta performs color synthesis inside the waveguide, because the microleds that can be mass-produced on the market are mainly monochromatic. thunderbird's innovative approach is to combine light into colors before entering the waveguide. thunderbird's innovative light engine is designed, patented and produced by us. we can make the light engine very small, only 0.36cc (cubic centimeter), which is a very small light projector. this allows the glasses to achieve full-color microled while being very thin and light, much like ordinary glasses. .

during this process we tried to make a series of such designs. another very important point is that the x3 is a product that will be officially launched for sale. on the one hand, it can help us accumulate more user feedback; on the other hand, it can serve as the basis for the start of an application ecosystem. because this product is sold to users, ecosystem developers can develop directly on it, such as developing various apps or ai agents, and then provide them to users for use. they can also directly receive feedback from users. this is a series of basic considerations for us on the thunderbird x3. i can’t say that much more, so i’ll briefly report it here (laughs).

zhang peng:your next clear goal is to launch a product that can be sold on the market and that users can actually use. today, can we imagine that if i spend money to buy this product, what kind of ready experience will it bring in terms of functional scenarios? and the scene we finally experienced? what might it look like? which experiences will be significantly improved compared to the original device? let us give you some spoilers in advance.。

li hongwei:let me first talk about one or two points that may be quite important. let me give you an example: what can we do with ar glasses?

of course, everyone may already know that ar glasses can realize a series of mobile phone functions, such as first-person photography and video calls, which are very effective. information reminders can also be implemented. when we receive wechat messages, we no longer need to look at our phones or watches, because the notifications can be displayed directly on the ar glasses. we have designed a series of very important killer applications around these.

an example is translation: this may not be the most important usage scenario, but people have said it a lot. the translation provided by ar glasses is not a simple translation, but is actually closer to the role of a copilot assistant.

for example, when we arrive in a new country, we get off the plane and go to customs. there are various signs along the way. you can follow people at the airport, but you can't when you get to the customs where the groups are divided. if you don't understand the language, you won't know which entrance is for residents of this country and which is for overseas passengers.

at this time, ar glasses can directly see various signs and then translate them into text you understand. including when you are chatting with the customs, the customs will ask you a series of questions. when he chats with you, you can see his translated content on the glasses in real time, achieving a smooth question and answer.

including after leaving the airport, when you go to a subway station, for example, the glasses can recognize the signs in the subway station, and it can also provide you with a navigation, which can not only tell you which line you are taking, but also prompt you with specific directions. information: for example, give you a navigation to tell you that you have arrived at this place, and when you come out, you will find a food street.

at this time, you can tell the glasses, i want to turn on this guidance mode, and then ai+ar can identify them one by one, because everyone knows that the video streaming mode of large models is quite strong now, and we can talk to him directly, and he will tell you which store you visit, what it is like, and what kind of food it serves. we can also tell you whether this store is suitable for your taste based on your preferences.

so this series of work. in fact, they are all trying to make the experience brought by ar glasses more perfect.

this is just a scene. going abroad may not be that frequent, but in fact if we think about it, in reality we may encounter similar scenes every day. if there is an ai + ar device around us 24 hours a day, it will see what you see and hear what you hear. in fact, this "external brain" can help you deal with a lot of things.

06

chinese ar entrepreneur

have unique advantages

zhang peng:thunderbirdinnovationin fact, it is now considered to be one of the leading startup companies in china on this track, but it is still a startup company after all. so how do you think you can succeed in this field? AR thistrackwhat are the advantages of chinese start-up companies? and i’m very curious about the issue of competitive strategy: as a startup company, and in a field that is valued by giants, how to find its own differentiated advantages and path forward?？

li hongwei:you are talking about a very important issue, so-called entrepreneurship. i have experienced several entrepreneurships and learned many very important things. one of them is that choice is greater than hard work.

how we choose can determine whether we are correct and successful? i think we may have to look at this from several levels. for example, let's look at it from a specific level. one of the points has actually been answered just now, because companies like meta and major manufacturers have invested a lot in technology, which i respect very much.

they released such technology to let everyone see that the era of ar glasses is coming, because the most important problems of technology have been broken through, which is very important. they spent enough money and made a lot of contributions to the industry. this is very important.

but startups will run faster. speed ​​is a very important thing. in most business competitions, speed is very important. so this is a very important point for startups like thunderbird innovation or chinese startups to achieve differentiated competitive advantages. this is related to the mechanism and all aspects, including china’s supply chain. maturity is related.

but another question that may be more important is: is the era of ar still like mobile phones, with american companies leading the way? it was they who completely led the way in the industry, especially in its earliest days. i don't think so.

thunderbird x3 is a product that balances technology and demand, but this does not mean that thunderbird innovation does not have vision pro technology. in fact, what is the most important thing about vision and vision pro? it is that they have a common technical direction and the most important technological breakthrough, which is the same between vision and vision pro.

the difference between vision and vision pro is the difference in product positioning. i want to train my whole body to be muscular, so i’d rather train it to a relatively toned state. should i solve all the problems now, or should i give a user a 60-gram pair of glasses that they can accept and like to use?

so it's just the choice of products. from this point of view, thunderbird innovation, as the leading company in this industry, and as a chinese company, i don't think it is much behind american companies. at least judging from the output.

so in ar, i think it is possible for chinese companies and american companies to lead the development of the industry together and bring about an industry explosion. it is even possible that chinese companies will play this role earlier or to a greater extent. this is related to china's overall technological development, supply chain development, and capacity accumulation over the past few years.

we've actually seen some of these opportunities. tiktok is the most powerful short video application in the united states. this was impossible in the past. before tiktok, we could hardly see such a situation. in terms of technology and the internet, china is lagging behind the united states in many aspects and doing things after the united states. but with products like dji and tiktok, i think they already represent that chinese companies can lead an industry.so i think this isvery importantone thing, except for speed, we are not behind in technology.。

another point is about the entire ecosystem. china's supply chain and application ecosystem are actually running very fast. for example, many large manufacturers on the west coast of the united states have actually spent a lot of efforts in the supply chain in china. now the only country in the world that can mass-produce microled is china. for example, some optical waveguide companies are also very advanced in china. including the large model of ai itself, there may still be a gap between everyone, but in fact, in terms of the scenario application of large models, chinese companies are still doing relatively well.

of course, another question is what kind of strategy and what kind of technology accumulation is right in the track of ar complete machine plus ecology.

i think there are two aspects. the first aspect is that we actually need to lay out future technologies. in this regard, thunderbird innovation has done a good job, and it has been relatively focused. big companies sometimes sway, because they have a lot of resources, and i will do a lot of technologies. maybe this technology has matured at this time, and the one i am doing may be slowed down. therefore, thunderbird innovation is very firm in core technology, and we have achieved technological leadership.

other aspects include our design of the entire route in addition to technology, including whether you will launch vision pro or vision first? the example i just gave is a very important strategy option.

in fact, we saw that meta also released ai glasses first. ray-ban meta’s glasses and the basic strategy of thunderbird innovation are also like this. our strategy can enable us to survive and develop better in this environment and gain a better competitive advantage.

what strategy did we adopt? we pursue the right technology and the right products for the future. but in the process, we first launch products that are born out of mature technologies and mature user needs, such as the relatively mature birdbath product. so everyone can see that thunderbird innovation is not only leading in technology, thunderbird innovation is also leading in market share.

the sales results of the products we launch at each stage are also leading, so this can give thunderbird innovation a relatively good environment for survival and development. we not only plan for the future, but also develop in the process.

let me talk about a lower level issue. of course, a bottom-level factor is the team. our global talent flow is relatively good now. in the thunderbird innovation team, for example, we have americans and koreans, so we have a variety of working languages. we recruit top talents from all over the world to do our technology, human factors design, and a series of complete machine research and development. so i think at the technical level, at the human and organizational level, there is not that big difference between the chinese team and the american team, or we may be the same, maybe combined with efficiency, so i believe thunderbird innovation has a very good team.

we also provide everyone with a very good working environment. we have some teams with outstanding talents, but at the lower level, i think the lower level is the most important. look at the track of ar complete machine plus ecology. first of all, it is a disruptive innovation, which we call disruptive innovation. it is different from the past mobile phones and the past internet. it is a new spatial internet that integrates virtuality and reality.

therefore, such new opportunities are often relatively more likely for new companies, and new companies can succeed in at least half of them. does that mean that the big manufacturers know whether this matter is important or not? does everyone know that trains are important? the first company to make an electric car was definitely a traditional car manufacturer, such as bmw and mercedes-benz. but in the end, electric cars were still produced by companies like tesla, including byd, which originally did not make so many cars. in the traditional field while there are relatively less powerful companies, relatively new companies will eventually succeed in accounting for at least 50% of the market, so disruptive innovation will have such characteristics. we believe that the track of ar complete machine plus ecology is like this. chance.

of course, what you said is very reasonable. the competition on this track is a competition of comprehensive strength. the supply chain, hardware, software, applications, ecology, sales channels, and brand marketing must be better. so it is indeed a relatively big challenge for a purely new startup company, because purely new startup companies are more suitable for single-point breakthroughs.

so how do i design the model of thunderbird innovation? thunderbird innovation is backed by the resources of a relatively large company like tcl, but it is also a completely independent company with independent operations, independent decision-making, independent financing, and independent listing. in this way, its entire process system values it is a new company and a start-up company’s values, but it has enough resources to support it. we think this is very important for the design of the entire company and its final success.

therefore, from the bottom-level selection and architectural design to the organization and selection of technical routes, to the accumulation of technical routes, to market share in all aspects, including the layout of the entire ecosystem, you can see that thunderbird’s innovative partners are all the world's top companies, including the united states and china, also include large models, operating systems, and application ecosystems.

so i think in this competition process, i don’t think thunderbird innovation is at a disadvantage. we look at competition issues prudently and respond actively, but we are very optimistic about the future deep down.

07

ai glasses are phs,

ar glasses are smartphones

zhang peng:we were talking about this just now AR the ultimate goal leads to personal computing. in fact, the new generation of ray-ban meta products has also been released this time. i am quite curious. it seems that it strengthensAI, which brings up an interesting topic: thunderbirdinnovationwill you make products like ray-ban meta?

li hongwei:we build future products, but at the same time we create products that combine mature technology with mature user needs.so based on our core strategy, we will do. we have a joint venture with dr. optical, which mainly createsproducts like ray-ban meta, that kind of AI glasses。its time to market will not be very slow, and it is likely to be earlier than the x3.

second point, how do we view these two products? in fact, if we look at it from the perspective of industry development or user needs, i think this is a process of industry development. generally, a relatively simple product will appear first, occupying more users, and then it will gradually add functions and become more powerful. this is what mobile phones are like. we first see mobile phones that can only make calls, such as mobile phones, such as phs phones, and then change to more powerful feature phones, and then to smart phones. it is probably such a process.

in this process, a company has several options. one option is to only make future products, and the other option is to only make current products. it's not impossible to only make future products, but there is a lot of pressure at this time. looking at the entire environment including financing, frankly speaking, apple may be the only one who is really holding back its big move. if we only make the current products, i think there is still a potential risk. well, it’s the so-called innovator’s dilemma. if you only make a feature phone, it doesn’t necessarily mean that you can make a good smartphone. it may even constrain you to make a good smartphone. this is what thunderbird innovation considers in its trade-offs.

zhang peng:we just took stock of a lot of AR which glasses technologies are ready today? which scenes are more ready?AI are there any new thoughts to look forward to in this part?

li hongwei:i think there are several aspects. first, from the user's perspective, it actually has various areas that can be optimized. for example, for glasses, we always want them to be as light as possible. then we hope that the wearing experience will be as comfortable as possible, and we also hope to have more styles.

this time ray-ban meta has also released a new style of glasses. now it is more about iterating the functions of this ai and improving its basic experience.

if we look at the first generation of meta ray-ban stories and the second generation of ray-ban meta, a very important reason why the second generation has achieved such high sales (millions) is that the basic experience has been improved.

what does basic experience mean? for example, if it has a microphone, can the microphone reduce noise when you speak? if you go out wearing glasses, the wind may be louder, so can the glasses hear your instructions clearly? a series of such functional points have many areas that can be optimized, and even whether the camera is horizontal or vertical can be considered.

in this differentiated layout, thunderbird innovation has systematic advantages. this advantage includes the supply chain and other levels. here we have some exclusive things that others cannot get. it may be reflected in the experience, or it may be reflected in the price.

of course, this includes the ai ​​part he just mentioned. domestic large-scale model companies are now doing very well in ai scenarios, including some developers, so this should also be compared. but this kind of development is not only included in china, but also overseas, so it will have more advantages than overseas companies.

another point is that the glasses themselves have another dimension that is very important in terms of operation, production, sales and user attention. in terms of the so-called brand, channel and design of traditional glasses, the success of ray-ban meta is very important. the most important thing is the contribution of ray-ban itself. zuckerberg also said that even if you don’t turn on the smart function, the glasses of ray-ban and ray-ban meta, and the glasses of ray-ban meta are also very good-looking glasses.

so you can see that some time ago, you can see that we announced the establishment of a joint venture between thunderbird and doctor glasses. this is also some of our creations in this industry. in history, there is no such situation between a technology company and traditional glasses. there is no such deeper cooperation between major manufacturers, so this is why we have to think very carefully and judge how to succeed in this matter. we believe that the brand channel design of traditional glasses is very important, so we are also making such a layout.

finally, i would like to add two points. the first point is that meta orion is doing very well, but i also want to tell everyone that chinese companies are not lagging behind. thunderbird x3 is a very important breakthrough in the same direction as meta orion. products, this is the case globally.

overseas companies are also very optimistic about this. many of our friends from american companies have asked us if we can get a discount if we buy in a group. make sure we get it first.

but more importantly, what i want to tell you is,thunderbirdinnovationbe it meta or meta, some companies have done this AR whether it is vision pro or vision in the field, these all represent the most important thing, that is, the future of ar glasses has arrived.

*source of header image: meta