"Apple Intelligence"
This is Apple’s highlight moment at WWDC 2024.
If it weren’t for Apple’s attempt to redefine AI, the headlines of the next morning would have turned into April Fool’s joke updates like "Apple parodies Android again" and "iPad finally has a calculator app."
Fortunately, this update is indeed expected.
Based on the built-in large model, it can integrate the information of a series of hardware products such as iPhone, iPad, Mac, etc., and provide appropriate assistance while perceiving user requests.
This highly integrated approach to the software and hardware ecosystem can only be said to be the Apple after "giving up cars and focusing on AI".
Of course, Apple is not alone in embracing the AI model.
In the past period of time, mobile phone manufacturers are setting off a wave of big models. Almost all the leading manufacturers such as Xiaomi, Huawei, Samsung, OV, and Honor have announced the news that they will be landing big models. In a trance, people dreamed of 5G, full-screen, folding screen and other concepts. When the mobile phone conference took off, AI seemed to have become a new battlefield for consumer manufacturers.
For others to see, it is called a trance.
And then, at some point, as you see this steady stream of AI devices, you start to ask yourself, how different are these AI devices from each other, or let me switch to a more straightforward way of asking the question…
Will there be any difference in function between an AI phone that costs only 100 yuan and an AI phone that sells for nearly 10,000 yuan?
With this doubt, I carefully selected in the market and found the two products with the largest price difference under the same brand. Is the "AI phone" a gimmick, and what kind of experience can it bring at different price points?
Ray will bring everyone together today.
Selection with different prices
You may not believe it.
But the most difficult part of this plan.
In fact, it is necessary to find the lowest priced AI mobile phone under the same brand. Which one is it?
The main reason for this problem is that various manufacturers will actually habitually use the emasculated version system for the 100-yuan machine. Compared with the normal version system, these systems will basically only change the version number, and there will be no substantive improvement in functionality.
For example, the Redmi 13C/13R 5G released this year is probably the cheapest HyperOS phone.
However, the biggest change in this product’s updated HyperOS is the update of the system’s bottom layer and security patches. Functions such as split-screen, negative-screen, and small floating windows have also been castrated, not to mention the various AI functions on top of it.
Tsk, I can only say.
I can understand the reasons why various manufacturers are doing this. After all, the performance of the 100-yuan machine is highly unlikely to bring these functions, but it has indeed caused me a lot of trouble in my work.
Fortunately, the brothers in the Xiaomi community are still reliable.
After searching and communicating for half a day, I finally found the cheapest "AI phone" under Xiaomi’s system.
Or "smart phones with AI capabilities", after all, Mr. Lu really doesn’t like the term AI phones.
Come, and I have carefully selectedRedmi Note 11T ProLet’s meet.
As a product released in May 22, Redmi Note 11T Pro is equipped with a Tianji 8100 processor, has the fifth generation of MediaTek APU, and the MIUI 13 system installed at the factory has also been officially updated to the HyperOS 1.0.2 version a few days ago.
The most important thing is that this machine is currently brand new 12 + 256G in the Pinduoduo for over a thousand yuan, and the low-end second-hand price is only around 500 yuan. It should be regarded as the cheapest product with HyperOS AI function at present.
Considering that the Tianji 8100 is still active in the new AI mobile phone after two skin changes, my expectations for it are not low.
As a comparison, we present the Xiaomi Mi 14 Ultra, which was released in February 24.
One side is the Redmi thousand-yuan machine released two years ago and just upgraded to HyperOS, and the other side is the factory-ready HyperOS system, equipped with the Snapdragon 8 Gen3 flagship processor, which theoretically has the top flagship of all Xiaomi’s AI capabilities.
This comparison really makes people look forward to it.
AI assistant: exactly the same
To say that the AI function that is most closely related to everyone’s life, I guess it is still the "AI assistant" that comes with each manufacturer.
After all, even when there is no AI ability, everyone will enjoy teasing their voice assistant.
And on Xiaomi, that’s empowered by the big model"Xiao Ai Classmate".
As the flagship phone of HyperOS, Xiaomi Mi 14 Ultra’s "Xiao Ai Classmate" is ready to use out of the box, while Redmi Note11T Pro can be upgraded to the latest version of "Xiao Ai Classmate" after updating HyperOS.
In terms of startup mode, there are many ways for Xiaomi/Redmi mobile phones to wake up "Xiao Ai Classmate", including Keyword Spotting, long-pressing the power supply for 1 second to let go and wake up, wire-controlled/Bluetooth headset wake-up, etc. After changing the system navigation mode to the classic navigation button, you can set long-pressing the Home button, long-pressing the menu button, long-pressing the return button, or long-pressing a custom key combination to wake up, or even tapping the back to wake up.
In terms of functional content, the functions of the two are basically the same, which means that the Redmi Note11T Pro can use all the capabilities provided in the large model skill.
After opening a continuous conversation in the "Xiao Ai Lab", even the Redmi Note11T Pro can achieve a continuous conversation effect with Xiao Ai after a wake-up call, enabling a more natural understanding of the context, and the communication experience with Xiao Ai’s voice has been significantly improved.
To try the basic cross-app operation first, I asked Xiao Ai to "help me record my itinerary to Haizhu Xintiandi at 9:30 this Saturday morning" on the calendar, and after automatically setting it up, I asked her to help me plan the route to Haizhu Xintiandi.
Just as before, there were no problems with application calls, but now voice commands can be expressed more fluently and with greater complexity.
Say it or not, it’s pretty cool to use.
As for the planning schedule.
Text generation, no problem.
In short, in the performance of Xiao Ai’s big model version.
At present, the Redmi Note11T Pro can indeed enjoy the same experience as Xiaomi’s top flagship.
Picture generation: 10,000 yuan flagship won
Just measuring text generation is definitely not enough.
Let’s just ask the big model directly and let it generate pictures for a try.
In short, if you ask him to draw pictures, then both products are fine. The Redmi Note11T Pro can also respond to programming instructions and output AI images with a resolution of 768 * 768, which is naturally the same as the Xiaomi Mi 14 Ultra in terms of quality.
Of course, even if both can generate images, it doesn’t mean that the functions of the two phones are the same.
Open the "Xiao Ai Voice" – "Xiao Ai Lab" – "Image Large Model" of Xiaomi Mi 14 Ultra, and you can click to download the 1.2G MiAI Text-to-Image model.
In this mode, users can use the Text-to-Image function unlimited times, which is essentially a lightweight Stable Diffusion model for local computing, but the generated content has to be reviewed in the cloud, so it cannot be used in disconnected mode.
In comparison, Redmi Note11T Pro, which cannot download MiAI, can only use the cloud Text-to-Image function. In fact, the generation speed is faster, but the number of uses per day is currently limited. I tried it, no more than 3 times.
Ah, that’s very reasonable, the positioning is suddenly pulled apart.
And some digressions.
Although Xiaomi may be the first batch of domestic manufacturers to implement the Text-to-Image function, the current related functions are a bit behind the development progress of other manufacturers.
For example, what kind of long-sentence picture generation, Xiao Bu can, Xiao Ai can’t.
Xiao Ai can only recognize a single phrase, no matter how long the prefix is, it must be in that phrase.
For example, what specific style of wallpaper is generated, Xiao V can do it, but Xiao Ai can’t.
Xiao Ai can only imitate the style of some well-known painters, such as Van Gogh and Monet, and her creative space is limited.
Of course, the current domestic mobile phone built-in image generation function, mainly to let consumers taste it.
In terms of freedom, generation effect and speed, these functions are weaker than third-party image generation tools, some of which even have a clear generation limit, hoping to achieve secondary payment, it is really a genius to generate revenue.
My assessment is…
If you really have that spare money, you might as well try using the local Stable Diffusion or GPT4.0 integrated DALL. E bar.
Image editing: 100 yuan machine small loss
Next, let’s take a look at two AI features that I consider to be the most valuable.
One, yesImage editing.
When many people take photos outside, they have a worry, that is, "random passers-by"; and we digital editors often encounter such troubles when taking samples or pictures of products.
Due to this, the AI elimination function that can eliminate passers-by and sundries will become the guest of honor at the conferences of various manufacturers, often occupying the C position in the introduction of AI functions.
Yes, apples are not immune.
And this feature, on the Redmi Note 11T Pro — doesn’t exist at all.
At present, the album editing application of this machine is still the old version, so only the most basic magic elimination, and there is no magic elimination Pro function on the flagship.
I tried a random picture, and the result was as follows:
Here, the disparity between the two products is revealed.
However, the Mi 14 Ultra also exists"Superfluous."The problem is that this is a common problem that AI is currently eliminating.
Of course, the same is true for other flagship-exclusive editing features.
For example, the very useful "AI Expansion Map" does not exist on Redmi Note11T, and even because of the existence of cloud audit, there is no way to bypass it –
Is that really the case?
Haha, not really.
In March this year, Xiao Ai launched a test of "multimodal" related functions, that is, it can implement functions such as picture Q & A or picture editing for pictures uploaded by users.
With this feature, Redmi Note11T Pro can also achieve advanced image editing capabilities such as AI image expansion and passer-by elimination, as shown in the following figure:
Of course, this function is currently incomplete.
For example, passer-by elimination does not support custom elimination subjects for the time being, and AI expansion currently only supports multiples between 1 and 2 times for expansion.
But at least, being able to use it is a good thing.
Smart summary: basically the same
The second is the intelligent summary.
Nowadays, busy modern people all have a "common problem".
Just after reading a webpage, I completely forgot what was written on it. I need to read the information website to find some information, but I have a headache when I see an article of more than 500 words.
At this moment, I really wish the whole world had become like Weibo.
According to the test, Xiao Ai students of both products support the "article summary" function. As long as they call Xiao Ai and say "summarize this article" on the browser page or the information software page, the function can be triggered smoothly.
If you want to summarize a Chinese article, then Xiao Ai will automatically generate a summary containing "one-sentence description", "article top 5 points", "golden insight" and "article inspiration".
If you want to summarize an English article, then Xiao Ai will automatically translate and generate a Chinese abstract for easy reading.
However, Xiaomi, which has always been dominant in AI calls, did not follow up the "AI call summary" function in time this time, which is indeed puzzling.
To achieve a similar function, you must first record the call on Xiaomi, then convert the call recording into text, and finally let Xiao Ai summarize the text.
Does it work?
It sort of works.
The problem is that this type of AI summary function focuses on being able to intelligently identify the content of the call and generate a summary of key information, without having to go back and repeat the questions.
If the operation is so cumbersome, how many people will do it?
summary
It’s not hard to see from our testing this round.
There are certain functional differences between AI phones at different price points.
The question ****** this disparity caused by hardware?
Well, the answer is fifty-fifty.
How can I put it? The OPPO Reno12 based on the Tianji 8250 Star Speed Edition has a series of AI functions including AI matting, AI elimination, and AI summarization. These functions have good practicality in the fields of work and study, image processing, and the fundamental reason for all this is that it is a new product priced at 2699 yuan.
At the same time, even if I try to install the latest "Xiaomi Album-Edit" for the Redmi Note11T Pro, when using functions such as smart image expansion or magic elimination of Pro, it will still fail because of the inability to authenticate through the cloud, which is obviously not a limitation of hardware.
In other words, the functional differences are due in part to the manufacturer’s positioning of the product.
Of course, there are indeed some functional gaps due to hardware.
For example, the Xiaomi Mi 14 Ultra was heavily promoted at the time of its launchXiaomi AISPFunction, which enhances image capabilities through algorithms, is basically meaningless on the Redmi Note11T Pro with only a 64MP prime camera.
However, since the AI functions of mobile phones are still basically in the cloud or end-cloud collaboration state, these AI functions can be replaced by third-party software, and sometimes, third-party software may be more convenient to use.
When the AI functions that manufacturers have emphasized can be realized by downloading a software, consumers will inevitably think –
Why do you want to buy a new cell phone?
Therefore, I say that today’s mobile phones are still some way from the era of true AI mobile phones.
)
As a high-performance model of Levante, the new car is equipped with a 3.8T V8 twin-turbo engine, and the acceleration time from 0 to 100 km/h is only 3.9s.
If you want to have an SUV with super performance, then Mercedes-Benz AMG GLC 63 S is absolutely no choice.
The 15th Anniversary Edition of Dongfeng Honda xrv15 inherits the excellent quality of suv made by Dongfeng Honda, and integrates the latest design concept of Honda in the world, taking into account the value and strength.
GAC Mitsubishi announced its new compact SUV Eclipse Cross, which will go on sale in November.
"online celebrity Car" which highly restores the appearance of the concept car and has a sense of design at the same time.
Guangzhou Automobile Chuanqi’s first SUV model also fills the gap between GS4 and FS7.
Dongfeng Fengxing T5 is the first product in the era of Dongfeng Fengxing 3.0. The biggest highlight is that it is equipped with FutrueLink3.0 car networking system developed in cooperation with Tencent and equipped with ADAS intelligent driving assistance system.
The tough route chosen in the design makes the new car more rugged in visual effect.
The new car will continue to be equipped with a 1.6T engine from PSA, and the new appearance is more youthful and fashionable.
As a super-cow plug-in mainstream SUV, the new car still maintains the strength genes such as high face value, large space and low fuel consumption.
The red version of Bordeaux’s new car has made many fans feel like it’s advanced.
Subaru’s new generation forester (fifth generation) is newly built based on Subaru SGP global modular platform. The handling performance, passive safety and power system compatibility of SGP platform have been upgraded.
Chery’s high-end SUV is also quite fashionable in interior design. The 1.6T engine powered by Chery is the first TGDI direct-injection turbocharged engine officially mass-produced in ten years.
In body design, the arc of the roof starts to press down from the C-pillar, which makes the side lines move a lot, while in the interior, it is simple and has a strong scientific and technological atmosphere.
The body size of 4830/1868/1775mm allows the new car to provide 5-seat, 6-seat and 7-seat layout designs to fully meet the needs of different users.
This "hormone SUV" is loved by more young people who are the main consumers. Its domineering appearance, 10.1-inch independent touch screen, safety configuration and entertainment configuration are among the best in its class.
In appearance, sylphy pure electricity can be regarded as a combination of listening to the wind and sylphy. The overall shape conforms to the moderate aesthetic of Chinese people and has some sense of science and technology.
The handling performance and safety protection ability of the new car will be further improved, and at the same time, it will be equipped with a brand-new intelligent network connection system to meet the requirements of young consumers for intelligent interconnection.
