By | Beyond the Layout, Author | Huahua
I had only planned to experience it for 5 minutes.
I was even thinking of finishing quickly to see the next booth.
But I ended up sitting in that chair for nearly 20 minutes.
Outside the exhibition hall, there were robots dancing, product launches, parameter displays on big screens, and wave after wave of people holding up phones, standing on tiptoes to take photos.
But here, it was so quiet it felt a bit unlike WAIC.
Wearing headphones, I watched the brainwave curve undulate slowly on the screen. An AI music therapy system developed by a team from the Shanghai Conservatory of Music was generating a piece of music unique to me at that moment, based on my brainwaves and emotional state.
No robots, no Agent demos, no product launches. The students on-site didn't mention a single model name to me the whole time.
There was only music, and myself gradually calming down.
At that moment, I suddenly felt that AI doesn't always have to amaze me; it can also bring me peace.
01
Leaving that booth, I walked back into the main WAIC hall.
Familiar liveliness returned before my eyes.
Huawei's Atlas 950 Super Node was surrounded by visitors, booths of large model vendors were blocked by queues, AI glasses were almost the hottest terminal this year, various robots shuttled through the halls dancing, shaking hands, running, and every few steps, someone would be holding up a phone chasing after them to film.
If last year's WAIC was about arguing whose model was stronger, this year, the keywords I heard most were:
Agent, computing power, terminals, workflows, embodied intelligence.
The entire industry is answering the same question: What else can AI do?
It can write code, create PPTs, generate videos, handle complex tasks, control robots, and increasingly intervene in our work and life.
This has almost become a consensus throughout the conference.
But precisely because of this, I started thinking about another question:
What is all this for, in the end?
The most obvious change at WAIC this year is that AI is moving from model competition to system competition; models themselves are becoming part of the infrastructure.
What really determines the AI experience is the complete system built on top of the models: computing power infrastructure, Agent frameworks, data loops, tool invocation, terminal entry points, and ultimately, the ability to land in real-world scenarios.
This is also why so many new terminal forms have appeared at this year's WAIC.
AI glasses aim to become new entry points for AI to perceive the world. Agent Phones aim to turn smartphones from passive tools into proactive assistants. Robots aim to give AI a physical body to truly enter the real world.
In the past, people asked: Can AI think? Today, people ask: Where will AI reside?
The answer is becoming increasingly clear: AI will enter every device closest to people.
But it was also amid this hustle and bustle that the quiet music therapy space stood out as incongruous.
Later, thinking back carefully, the reason it was memorable wasn't because the technology was the most advanced.
Quite the opposite.
It might have been one of the least flashy products in terms of technology at WAIC. No multimodal large model, no complex tool invocation chain. Just brainwave collection, plus a set of music generation algorithms.
In terms of technical specs, it could hardly rank among the most popular booths at this year's WAIC.
But in terms of emotional impact, it was the only place where I let my guard down and quietly stayed for 20 minutes.
This was the first contrast WAIC gave me this year.
AI is getting stronger, but what truly moves people is increasingly close to the human essence itself.
02
Looking around the entire exhibition hall, almost all cutting-edge technologies point towards the same goal:
Extreme efficiency improvement.
This is not wrong.
Efficiency has been the biggest story in AI entrepreneurship over the past three years and the direction most recognized by the capital market.
Whoever can help people accomplish more in less time secures an entry ticket to the next round of competition.
But over these two days shuttling between halls, a vague feeling emerged in me.
Efficiency is not equal to happiness.
In fact, as the efficiency revolution reaches today, it's giving rise to a new technological paradox.
The anxiety many people feel today is not because efficiency is too low.
Quite the opposite, it's because the entire world has been thoroughly accelerated by the logic of efficiency.
Too much information, work too fast, schedules packed full. Phones ring from morning till night, endless messages to handle, emails to reply to, decisions to make every day.
AI can indeed help you finish these faster.
And then?
The time saved is often immediately filled with new tasks. When all tools are improving efficiency, the system's expectations of people are also rising simultaneously.
In the past, technology demanded our physical strength; now, technology demands our attention.
This isn't actually a problem unique to the AI era.
The internet of the past 20 years has essentially been driving an efficiency revolution as well.
Search made information retrieval faster, e-commerce made shopping faster, food delivery made eating faster, mobile internet made connecting people and information faster.
Technology has constantly been saving us time. But one question has never been answered. What to do with the time saved?
03
Another very obvious change at WAIC this year is that AI is becoming more human-like.
Robots are starting to have expressions, shake hands, offer companionship. Agents are starting to remember your habits, understand context, proactively help arrange tasks. AI glasses are starting to observe the world for you. Phones are attempting to become always-on intelligent assistants.
The entire industry is striving to make AI more human-like.
But at the same time, I feel that content truly about humans is somewhat less.
People discuss Tokens, inference speed, context length, tool invocation, but rarely discuss human anxiety, fatigue, loneliness, and the real needs technology should ultimately address.
The more powerful technology becomes, the more human vulnerability seems to have nowhere to rest.
The music therapy space happened to address this sense of displacement.
It didn't save me a single minute, nor did it complete any work for me. But with 20 minutes of quiet, it briefly pulled me out of the entire hall's commotion.
Actually, it's not just music therapy.
If you observe carefully, you'll find another category of AI products quietly emerging within the industry.
AI psychological counseling, AI companionship, AI elderly care, AI emotional management.
They are not the most dazzling stars of WAIC today. Because they aren't as stunning as robots, nor as buzz-worthy as model launches.
But they respond to a different kind of need. In the past, the internet built information infrastructure; mobile internet built connection infrastructure.
And in the future, AI might be building a new kind of infrastructure: emotional infrastructure.
It solves not just how things get done, but focuses on how people, in an increasingly complex world, can be understood, accompanied, and cared for.
04
Over the past few years, the AI industry has been sprinting around one keyword: Intelligence. Who is smarter, who has stronger reasoning, who can complete more complex tasks.
This is the first stage of AI development. Without this stage, there would be no AI wave today.
However, as model capabilities inevitably move towards commodification, computing power, parameters, toolchains—things that money and resources can gradually level.
What is truly hard to replicate is always the ability to understand a person.
In the past, technology solved how to do things faster. In the future, technology might need to answer how to make people live better.
This is also why the term 'intelligent partner' is worth pondering. Tools focus on utility; partners focus on relationships.
Leaving the exhibition hall, robots were still dancing, product launches were still ongoing.
Today's WAIC showcased the new heights AI capabilities are reaching. Greater computing power, stronger models, more diverse terminals.
But that quiet music therapy space also reminded me of another thing: the value of AI depends not only on how many capabilities it possesses, but also on how it enters people's lives.
When model capabilities gradually become infrastructure, the truly difficult part might become understanding specific people, specific scenarios, and specific problems.
Last year's WAIC, I remembered models. This year's WAIC, I remember a piece of music.
The next stage of AI still requires stronger intelligence.
But it also needs deeper understanding.
