Real and effortless? AI workers are coming, and their work ability breaks human cognition

Mondo Workplace Updated on 2024-02-06

There's a famous saying in the internet world: "You never know if it's a person or a dog on the other side of the network." ”

Now, with the gradual maturity of AI technology, it is time to change this sentence to: "You never know if the other side is AI or a human".

Recently, iFLYTEK's AI learning machine has attracted the attention of many netizens, which allows users to have conversations with AI virtual humans such as "Li Bai", "Confucius" and "Einstein" who have traveled through time and space.

Each avatar can chat with the child on a daily basis and "read minds". For example, "Confucius" has a lot of knowledge, not only can you play idiom solitaire with children, but you can also guess your child's inner thoughts by chatting with your child.

Many netizens said that if such artificial intelligence can be applied to the Internet, they can become "contemporary Confucius" and "a generation of writers" in seconds.

In fact, artificial intelligence has long been close to us, but most ordinary people are not aware of it.

For example, the popular mobile game Glory of Kings, there is an enlightened AI play, this AI has a deep learning algorithm, which can simulate the player's game behavior, and can continuously optimize its own strategy, the level is comparable to that of professional-level e-sports players, many high-level players have been beaten by AI, and there are countless strategies on how to defeat the enlightened AI on the Internet, **Breaking one million.

There are also anchors on major platforms, many people look at the delicate makeup, glamorous and bright, rich expressions and fluent sales, but if you look closely, you will find that many anchors are not real people, digital people have already poured into the live broadcast room, AI customer service has a smooth dialogue with you in the dialog box, AI anchors are also becoming the mainstream force in the live broadcast room, maybe you have brushed them up in countless late nights, but you have never noticed.

To this extent, thanks to the natural language processing and computer vision capabilities of AI large models, through the learning and research of the content in the database, AI is more intelligent and closer to people's hearts in the process of continuous optimization, so it can "confuse the real with the false".

Even the elimination function of mobile phones, with the help of AI, is more professional, like millions of retouchers.

For example, the OPPO Find X7 series, which shined on the night of Weibo not long ago, as the industry's first mobile phone to deploy a large model with 7 billion parameters on the end side, the Find X7 series has more intelligent interaction capabilities, so it can support the recognition and segmentation of 120 types of subjects such as passers-by and objects, and realize hair-level segmentation, up to 6 multi-subject separation, and the filling and natural generation of ultra-large area images through a large amount of AI computing power.

This ability ensures that the AI elimination function of the Find X7 series is completely different from other mobile phone experiences, not only can it quickly check and identify the elements that the user needs to remove, but also "calculate" the background and complete it through the surrounding environmental elements, and the effect is immediate.

For example, the comparison of the images of Xin Zhilei on the Weibo night shared by the blogger, while the erroneous arm completely disappeared, the floral elements in the background were also completed, and the overall effect was like "clearing", real and natural. It's hard to imagine that this is the effect of the AI elimination function, even if it is said that it has been processed by professional PS, some people believe it.

Recently, when the first monthly solicitation activity of the OPPO 2024 Super Image Contest "Year of the Dragon" is underway, you can use the AI elimination ability of your mobile phone to take a picture of the "Year of the Dragon atmosphere" in your heart and have a happy New Year.

In addition, the Find X7 series also launched the first AI large model voice summary, which can accurately understand the content of the call, just click the AI summary button in the call interface, the call content will be automatically saved to the note, open the note, you can easily view the call time, highlights and other summaries.

For workers who often need to communicate with Party A, this function undoubtedly greatly improves work efficiency and saves a lot of time and energy.

Through the upgrade of the large model, this time OPPO also brought a new Xiaobu assistant function. With the empowerment of AndesGPT, Xiaobu is not only a voice assistant, but also a personal assistant that understands users better.

For example, it can support a variety of functions such as text generation**, text generation, AI article summarization, etc., and the room for imagination is huge. With the help of the assistant, you can also write resumes, draws, make travel guides, etc., what used to take a day to complete, with the help of the assistant, it may only take a few minutes.

In addition, the Find X7 series also performs very well in other aspects, equipped with OPPO's self-developed software and hardware chip fusion technology solution - Tidal Architecture, which further breaks through the boundaries of the performance optimization link and penetrates into the system-level cache and L3 cache that can only be optimized by chip companies.

The standard version of the Dimensity 9300 has also been pushed to the upper limit of performance, allowing the phone to have more powerful performance, higher energy efficiency, and better stability and battery life.

In tests that simulated a user launching an app, the Tidal Architecture Find X7 achieved a standard deviation of only 11ms, which is the first among all flagships, and for the first time, an Android flagship has a startup performance experience that surpasses Apple's.

It is worth mentioning that the Find X7 series has also achieved comprehensive and meticulous business experience in terms of business experience: the launch of "anti-leakage" anti-acoustic field call privacy protection technology, AI environmental sound cancellation technology, international getaway service that can be easily enjoyed by transnational data, and the four major business functions of the VIP mode, which has been widely praised and can bring privacy and security to business people, have strengthened the security, privacy and convenience of people's use of the machine in life and work, it can be said that this mobile phone is top in many aspects.

Although AI is developing rapidly, the current AIPC is still in the stage of AI ready, that is, the stage of upgrading the hardware to support large language models. The reality is that the AI ON stage of better product AI experience and AI application has not yet fully arrived.

For example, iFLYTEK's AI virtual human stays more at the level of language communication; There is also a mobile phone assistant, although it can play memes, but it has not evolved to be able to fully "understand" people.

Fortunately, the AI model is still being polished, and there are already many products that let us see the prototype of the AI-style future.

Anura MagicMirror, which recently won this year's CES Innovation Award, can identify and analyze people's facial blood flow through its emotional technology Deepaffex, translate it into various health indicators including blood pressure, and ** the probability of disease.

There's also Cappella's mobile app, which can "translate" your baby's cry based on a large model to tell you if they're hungry, unwell, tired or in need of a diaper change. According to Cappella, the technology that uses artificial intelligence and machine learning to interpret a baby's needs is about 95 percent accurate, compared to about 30 percent for humans trying to guess a baby's needs.

Therefore, with the help of huge datasets and multi-level pre-training, there are more application logic and scenarios for AI large models to be realized, and it is expected that more enterprises will research AI in the future and integrate it more deeply into our lives.

Related Pages