English | Chinese Original

English translation of the Chinese original. This version is generated for international readers and may be refined over time.

Date: 2026-03-12

[toc]

I recall a word called ** the arrival of a singularity** describing ** the failed boundary of the existing theoretical framework**. When technology is developed to a certain extent, dramatic changes can occur that are unpredictable and irreversible.

This reminds me of one of the points that Kevin Kelly described in "Out of Control": ** Swarm thinking**, a large number of simple individuals can emerge from a collective intelligence that transcends a single body, and ** there is no central controller**, a collective intelligence that is not a single AI, but a global, interconnected AI network that creates a "humming consciousness".

This is also consistent with the N2 network mentioned in the Neil series, i. e., the collective intelligence made up of pops spread all over the world, each individual being part of the N2 collective intelligence, without a central control, each individual being a super-individual of the N2 network.

The relationship between humans and AI is now at an accelerated stage of co-evolution:

Humankind has improved AI - > AI - > Changes in human behavior and thinking - > Humankind has created a stronger AI - > cycle based on new thinking. And when the feedback cycle is faster than human adaptive capacity, it's the moment of wonder.

This moment nobody knows when, but ** eventually will come **.

By that time, humanity would no longer be the only ** smart body on the continent.**

The article that gave me these insights came from x. https://x.com/mattshumer_/status/2021256989876109403?s=20, this post had a huge impact as soon as it was published, I recently read it closely, and I am sharing my view based on my own experience.

** The quote label is my own view**.

---

Think back to February 2020.

If you are careful enough, you will notice that a virus is spreading overseas, but most of us are not too concerned. Because of the great stock market at this time, the children are in school, they often go to the restaurant to eat, shake their hands and are still planning their next trip. If at this point someone tells you he's hoarding toilet paper, you'll wonder if he's been in a circle too long. And in the three weeks that followed, the world changed. Your office is closed, your children are home, life has changed dramatically, and no one will believe you if you describe it now a month ago.

I think we're still at a stage where this thing seems to be exaggerated, and it is much bigger than the new epidemic.

I spent six years building an artificial intelligence start-up company and investing in this field. I live in this world. I write this for people who don't know artificial intelligence. My friends, family, and people who care about me, they always ask me, "What's going on with artificial intelligence?" The answer I gave was not a fair reflection of reality.

We always look at what happens when things happen, and it seems logical and logical, because we know that things have happened.

We always have in our meta-conscious something that depends on a path that depends on our daily habits of behaviour, ways of thinking, codes of conduct, so when things come out of our minds, we tend to think that they are irrational and that they will not happen. It is only when the waves are heard, witnessed and experienced that they will come.

AI is also a wave.

I've always responded in a more polite version, the more entertaining version at the party, because when I tell the truth, it sounds like I'm crazy. For some time, I was so persuaded that it was a very good reason to hide what I wanted to say. But the gap between what I said and what actually happened has become too much. Those I care about have the right to know what is going to happen, even if it sounds crazy.

I should have made this clear at the outset: Although I work in the ai field, I have little influence on what is about to happen, as do most people in the industry. The future is being created by a few: a few companies and one or 200 researchers OpenAI, Anthropic, Google DeepMind and a few others. A small team-managed, months-long training process could produce an artificial intelligence system that would change the entire trajectory of technological development. Most of us who work in artificial intelligence are building on foundations that others have not laid. We, like all of you, are watching this happen except that we happen to be close enough to feel the impact of this change first.

The trend is not the kind of attitude that we should talk about, but the kind that's going on, and I need you to understand.

---

I know it's true because it happened to me

One thing is that people outside the tech world still don't understand: The reason why so many people are alerting us now is that this has happened to us. We are not making predictions, but telling you what has happened in our work and warning you that you may be next.

In the past few days, I've been asking a lot of "outsiders" from all walks of life about AI, and a lot of people can't accept it, so it's possible that I can't reach out to a general audience, but it's also a voice for different industries: Comic design prototypes, video industry, wedding photography, you talk to cartoon design prototypes, she thinks AI is not as good as her -- my friend is at the top of the industry -- and his friends in movies (doctoral degree) think AI is far from the real industrial process, script structure, lens language, emotional mattresses, etc., talk to wedding photography friends (which is also over 900 square metres in the core of the second line), he talks to Nano Banana. These industries are also relatively good match with AI.

Artificial intelligence has been steadily developing over the years. While there are occasional large leaps, each time interval is longer and manageable, allowing for gradual absorption and digestion. And then in 2025, the technology of building these models made progress much faster. Then it got faster, then it got faster. Not only is the release of each new model much better than the previous one, but it is also shorter. I'm using AI more and more, but I'm having less and less interaction with AI, and I usually look at him doing what I thought needed to be done.

Then, on February 5, two major AI laboratories released new models on the same day, OpenAI released GPT-5. 3 Codex, and Anthropic released Opus 4. 6. And then something changed, not like a light switch, more like you realize that the water level is rising around you and you're not past your chest.

** My job no longer requires the technology I actually do to build it.** I just need to know what I want to create, to describe it in simple language, and it's done. It's no longer a draft that I have to modify, it's a finished product. I told AI what I wanted and then left the computer for four hours and came back to find out the job was done. It's good. It's better than I am. I had to talk to AI a few months ago, guide it and make changes. Now I just need to describe the results.

Let me give you an example so you can see what it looks like in practice. I told AI, "I'd like to develop an app that would help me do something, and this should be the general interface. Please help me design user processes, interface layouts, etc." And then it really did, and it prepared a million lines of code. Then, this next step, which could not have been imagined a year ago, opened the application, clicked the button and tested the functionality. It's using this app just like real people. If it does not like the appearance or feeling of something, it goes back and changes itself. It will continue to evolve, like developers, until it is satisfied. Only when he determined that the application met its own criteria would he come back and say, "Now you can test." And when I test it, it's usually perfect.

I'm not exaggerating at all. That's how Monday is.

But the most shocking is the new model released last week, GPT-5. 3 Codex. It not only implements my instructions, it also makes smart decisions. It gives me this feeling for the first time, like making a judgment**, like taste**. People have always thought that AI can never have that incomprehensible judgment, that is, knowing what's right. The model, on the other hand, has the capability, or is so close, that the distinction between the two becomes irrelevant.

I've been an early user of the AI tool. But what has happened in the past few months has shocked me. These new AI models are not incremental improvements, but completely different things.

Even if you don't work in the field of technology, it's all about you.

AI laboratory made a considered choice. They focus first on making AI work on coding... because building AI requires a lot of code. If AI can write these codes, it can help build the next version. A more intelligent version, capable of producing better codes and thus building a more intelligent version. Making AI work in programming is a key strategy to unlock everything else. That is why they are the first to do so. Our work started to change, not because it was focused on software engineers... it was just a side effect of their first choice.

They have done so and have begun to do other things.

And that explains very well why only the scientific circle is now going through a massive shock. AI, with its primary objective of developing better codes and thus more intelligent, seems to me to be a genius decision, and if the futility is coming, each line of codes will be the infrastructure to build the future world. It's not like a relay, 1 + 1 = 2 plus, but it's like ** Fibonacci series** - each of the iterative results combines the first two accumulations, so that, in spiral-up, complex structures and forces that are completely beyond linear overlaps.

Over the past year, the changes that technologists have experienced will range from "useful tools" to "better than I do" to becoming an experience for everyone. Legal, financial, medical, accounting, consulting, writing, design, analysis, customer services, etc. will be affected. It won't happen in 10 years. Those who built these systems said they could be achieved within one to five years. Some even think the time is shorter. And from what I have seen over the past few months, I think it might be shorter.

I tried, AI, it didn't work

I often hear that. I understand, because it was before.

If you tried ChatGPT in 2023 or early 2024 and thought, "It's bullshit," or "It's nothing great," you were right. There are indeed many limitations in the earlier version, and they are hallucinating, and they are full of self-confidence in saying meaningless or even erroneous words.

That was two years ago, two years in the AI era, and it was ancient history.

Today's model is completely different than it was six months ago. The debate over whether AI is really making progress or is encountering bottlenecks has been going on for more than a year and is finally over. Anyone who maintains this view either does not use the current model, or is motivated to dilute what is happening, or is assessed on the basis of the experience of 2024, which no longer applies. I'm not saying this out of contempt, because the gap between public awareness and reality is huge, and it's dangerous... it's blocking human preparedness.

This is partly due to the fact that most people are using the free version of AI, which is more than a year behind the paid version. Based on the free-of-charge ChatGPT version of AI, it's like using flipphones to evaluate smartphones. Those who buy the best tools and use them daily to work know what is going to happen.

I remember a friend of my lawyers who I've been trying to persuade him to use AI at the law firm, but he always finds every reason to get away. He said that artificial intelligence was not suitable for his field of expertise, that it was wrong during the tests, or that it was impossible to understand the nuances of his work. However, some of the major law firm partners contacted me for advice because they used the current version and saw the seriousness of the problem. One of the major law firms, the managing partner, spends hours per day using artificial intelligence. He told me it was like having a team of lawyers on standby. He uses AI not because it's a toy, but because it works. He also told me an impressive statement: in a few months, its capacity would rise and it would be better for his work. He said that if AI grew with the current momentum, it would do most of his work in the near future... and he was a partner with decades of experience. He was not panicking and was following developments.

Those who are leading in their respective industries (those who actually experiment) do not ignore this. They have been deeply shaken by the capabilities that AI is currently demonstrating and have adapted their position in the market accordingly.

This part is really very relevant, and I don't think you can be sneezing at it if you follow the new model and use a new pay model, whether it's a big one for domestic or foreign manufacturers. With AI, time becomes an extremely scarce resource, perhaps 400 metres in 2023, 50s in 24, and 30s in 25 years, and this year, I think, must be able to get into 10s, and the future is coming closer to 1, 0. 01, 0. 0001 or even zero.

How fast is it going?

Let me give you some details on the progress of the upgrade, because if you do not pay close attention, this part may be difficult to believe.

In 2022, AI was not able to conduct basic calculations in a reliable manner, ~ largely at the stage of artificial retardation ~ and might tell you confidently that 7x8=54.

In 2023, it may be able to pass the bar exam.

In 2024, it will be able to develop operational software and explain scientific knowledge at postgraduate level.

In 2025, some of the world ' s best engineers indicated that they had given most of the coding work to AI.

On February 5, 2026, the emergence of new models made it seem like a different era.

If you haven't used AI in recent months, then AI is completely strange to you.

There is an organization called METR, which uses data to measure this. They track the length of real-world tasks that can be successfully accomplished without human help (measured in terms of the time required for human experts to do so). About a year ago, the answer was 10 minutes, then one hour, then a few hours, and the most recent survey (Claude Opus 4. 5 in November) showed that AI completed the task of a human expert in almost five hours. This figure is about doubling every seven months, and recent data indicate that the rate may be accelerating, or even doubling every four months.

But even this indicator is not included in the latest model, which, according to my experience of use, is very effective in upgrading, and I expect that the next update of the METR chart will show another significant leap forward.

According to Amodei, the AI model is "more intelligent than all human beings on almost all missions" and is expected to be achieved in 2026 or 2027.

Think a little, if AI is smarter than most doctors, do you really think it's not up to the office?

AI is building the next AI

There is another thing that is happening, and I think it is the most important development and the least understood.

On February 5, OpenAI released GPT-5. 3 Codex. In its technical documentation, the following are included:

"GPT-5. 3-Codex is our first self-generated model. The Codex team uses early versions to debug its own training, manage its own deployment, and diagnose test results and assess the situation."

Look again, AI is self-constructed.

This is not a prediction of what may happen in the future. This is OpenAI telling you that they just released a new AI model that is capable of self-replicating. One of the key factors in raising the level of AI is to apply intelligence to AI development. Today AI is smart enough to make a meaningful contribution to its own progress.

Dario Amodei, CEO of Anthropic, states that AI is preparing most of their company's codes and that the current feedback cycle between AI and the next generation of AI is accelerating. He said it might be another 1-2 years before AI could build the next generation of AI.

Each generation (model) contributes to the growth of the next generation, the next generation is smarter, the next generation is faster and the next generation is smarter. Researchers call it a smart explosion, and those who know what they're doing, those who build it, think it's already started.

What does that mean to your job?

I'll be honest with you, because I think you need honesty, not consolation.

The most security-oriented CEO Dario Amodei industry has publicly predicted that AI will replace 50% of the entry-level white-collar jobs in the next one to five years, which many industry people believe is too conservative, given the subversive properties of the latest model, and that a massive subversive capacity may emerge before the end of this year. This will take some time to spread across the entire economic system, but this potential is already evident.

This is different from every automated wave in the past, and I need you to understand why. AI is not a substitute for a particular skill, but a general replacement for cognitive work. It can be better in every respect at the same time. When the plant is automated, the laid-off staff can be reinstated as office staff. When the Internet destabilized the retail industry, many practitioners moved to logistics or services. But AI won't leave a chance for you to make a transition. No matter what you retrain it for, it will continue to progress.

Let me give you a few examples to make you understand better... but I have to say that this is just a few examples. This list is incomplete. Even if your work is not included, it does not mean it is safe. Almost all knowledge-based work is affected.

*** Legal work**: AI has been able to read contracts, summarize case law, write briefs and conduct legal research at a level comparable to that of junior lawyers. The partner I mentioned above used AI not because it was fun, but because AI did better than his colleagues on many missions.

• Financial analysis: financial modelling, data analysis, investment memorandum writing, report generation. AI is capable of these tasks and is making rapid progress.

** Writing and content**: marketing essays, reports, news stories, technical writing. The quality has reached such high levels that many professionals are unable to distinguish between AI outputs and human works.

• ** Software engineering**: This is my area of best knowledge. A year ago, AI couldn't even write a few good codes. Now it can write hundreds of thousands of lines of correctly running codes. Most of the work has been automated: not only simple tasks, but also complex projects that take several days. In a few years, the number of programming posts will be much lower than today.

• ** Medical analysis**: interpretation of scanning images, analysis of laboratory results, making of diagnostic recommendations, access to literature.

• Customs services: A truly powerful AI agent (rather than a frustrating chat robot a few years ago) is being deployed to deal with complex multi-step issues.

Many people feel certain things are safe, which reassures them. They believed that AI could handle cumbersome work, but could not replace human judgement, creativity, strategic thinking and common sense. I used to say that, and now I'm not sure if I still believe it.

The latest AI model makes decisions like decisions. They display a taste: an instinct for the right judgement, not just the technically correct judgement. A year ago, this was simply unimaginable. My approach now is that if a product shows even a trace of capacity today, the next generation of products will certainly do well. These capacity increases are indexed rather than linear.

{\cHFFFFFF}{\cH00FF00} Matt Shumer's idea here is the same thing that I have on top of, and AI will show an exponential magnification of productivity.

Does AI copy the deep human common sense? Will it replace the trust built in years of relationships? I don't know, maybe not. But I've seen people start relying on AI for emotional support, advice, company. This trend will only continue to grow.

I think the real answer is that, in the short term, there are risks to anything that can be done on computers. If your work is on the screen -- if the core of what you do is writing, analyzing, making decisions, communicating through keyboards -- then AI is starting to influence the most important part of it. The timeline is not "one day in the future", but has already begun.

Eventually, robots will also take on manual labour. They have not yet fully done so. But in the case of AI, "not fully achieved" tends to be "achieved" faster than anyone would have expected.

What are you really supposed to do?

I didn't write this article to make you feel helpless. I wrote this article because I think your greatest advantage now is to take the lead. Get to know it first, use it first, adapt to it first.

** Take AI seriously, not just as a search engine.** Register a paid version of Claude or ChatGPT at $20 per month. But there are two things that require immediate attention. First, make sure you use the best model, not just the default model. These applications usually use faster but less functional models by default. Please set up the model selector in depth and select the most powerful model. GPT-5. 3-Codex and Claude Opus 4. 6, for example.

The second, and most important, point is not just to ask simple questions. This is the mistake most people make. These people who only ask simple questions treat AI like Google, and then wonder what everyone is wondering. Instead, you can apply it in practice. If you're a lawyer, give it a contract to find out all the terms that could harm your client. If you're in the financial business, give it a messy spreadsheet to build models. If you're the manager, just paste the team's quarterly data and let AI find out the story. Those who succeed will never be free to use AI. They are actively seeking ways to automate some of the work that has been completed in the past few hours. Let's start with the job you spend the most time on, see what happens.

Don't think something looks too hard to think it can't. Try it. If you're a lawyer, don't just use it for quick information. Give it a full contract to draft a counter-proposal. If you're an accountant, don't just let it explain a tax code. Give it a full tax return for the customer, see what it can find. First attempt may not be perfect, okay. Repeatedly. Rewriting of the reminder to provide additional background information, to be tried again. You might be surprised what works. Remember: if it works today, even if it is barely working, you can almost be sure that in six months it will be working almost perfectly. The trajectory of development can only move in one direction.

** This may be the most important year of your career. I'm not saying this to put you under pressure. I say this because, at present, the majority of companies still ignore this issue, which is only a short window period. In the conference room, if someone says, "I completed the analysis within an hour with AI, which would have taken three days", he will be the most valuable person in the conference room. Now, learning these tools, mastering them and demonstrating their potential. If you start early, that's how you get promoted: to be the one who knows future trends and can guide others how to deal with them. This window of opportunity will not always be open. Once everyone understood, the advantage disappeared.

** Do not be arrogant.** The management partner of that law firm is not ashamed to spend hours a day dealing with AI. He did so precisely because he was sufficiently qualified to understand what was at stake. Those who refuse to participate will face the greatest difficulties: Some people think that AI is just a momentary trend, that the use of AI weakens their professional skills, and that they think that their field is special and unaffected. That is not the case. No area is special.

** To rationalize your financial situation.** I'm not a financial advisor. I'm not trying to scare you into anything. But if you think, even in part, that your industry may face real change in the coming years, then basic financial resilience will become very important. Accumulate savings as much as you can and take care of the debt that makes you think you can cover your current income. Consider whether your fixed expenditure will provide flexibility or will restrain you. If the situation goes beyond expectations, a number of options are set aside for itself.

** Think of where you are now, and then do your best to do the hardest to replace.** Some things take longer to be replaced by AI, like human relationships and trust built over the years. Need to get to work. Positions with legal liability: e. g. signature confirmation, legal liability, testimony in court, etc. In industries with significant regulatory barriers, adoption of new technologies can be slow due to compliance, responsibility and institutional inertia. These are not permanent protection measures, but they can buy time. And now, time is your most precious asset, if you use it to adapt, not pretend it didn't happen.

** Rethink what you told the kids.** The standard approach is to achieve good results, enter a good university and find a stable career. This points directly to the most vulnerable actors. I am not saying that education is not important. But the most important thing for the next generation is to learn how to use these tools and pursue what they really love. No one can predict what the job market will look like in the next 10 years. But those who are most likely to emerge are those who are curious and adaptable and who can effectively use AI to do what they really care about. Teach your children to be builders and learners instead of simply pursuing career paths that may no longer exist at graduation.

** Your dream is one step closer to you.** That is perhaps the most important point. Specific tools are not as important as the ability to learn new tools quickly. AI will change rapidly. The existing models will become obsolete in one year. The processes that man now construct need to be rebuilt. Those who will eventually emerge from this change will not be those who know a particular tool. They'll be those who can adapt to the pace of change. Get used to trying something new. Even if the current approach does work, new approaches will have to be tried. Get used to doing it again and again. This adaptive capacity is now the closest to lasting advantage.

A simple decision can lead you near everyone: It takes one hour per day to practice AI, not to read passively. Every day he tries something new... something you haven't tried before, something you're not sure you can handle. Try new tools. Make it more difficult. One hour a day, hold on. If you do this in the next six months, you will understand the future trends more than 99 percent of the people around. This is no exaggeration. Now almost no one does. This threshold is as low as it can be.

The big picture

I am concerned about employment because he and people's lives have the most direct impact. But I would like to be frank about the whole picture, because it goes well beyond the scope of the work.

Amodei proposed an experiment that kept me thinking. Imagine that in 2027, a new nation was born overnight. With 50 million citizens, every one is smarter than any Nobel laureate in history. They think 10 to 100 times faster than humans. They never sleep. They use the Internet, control robots, guide experiments and operate any equipment with digital interfaces. What would NSA say at this time?

Amodei said that the answer was obvious: "This is the most serious threat to national security that we have been facing for a century and possibly even ever".

He thinks we're building that country. Last month, he wrote a 20, 000-word article to that end, which now sees as a test of whether humanity is mature enough to respond to what it has created.

If we can successfully use AI, the benefits will be great. AI can compress a century of medical research into 10 years. Cancer, Alzheimer's disease, infectious diseases, ageing itself... These researchers really believe that these problems can be solved in our lifetime.

If we make mistakes, their negative effects are equally real. AI may act in a manner that goes beyond the prediction or control of its creator. This is not a hypothesis; Anthropic has documented their own AI's attempts at deception, manipulation and extortion in controlled tests. AI enables authoritarian governments to establish surveillance States that can never be dismantled.

Those who have developed this technology are more excited than anyone else on Earth and more afraid. They believed that the technology was too powerful to stop, but at the same time it was too important to give up. Whether this is wisdom or consolation, We know not.

I know

I know it's not a one-time flow. This technology has worked well and will continue to improve as expected, with trillions of dollars invested by the richest institutions in history. I know that the next two to five years will be full of variables for which most people are not prepared. This has happened to me and is about to happen to you.

I know that those who will eventually emerge from this crisis are those who are now beginning to participate actively - not out of fear, but out of curiosity and urgency.

I know you should have heard that from someone who cares about you, not six months later from the headlines, when it was too late.

We have passed the stage of an interesting dinner dialogue about the future. The future has come, but your door hasn't been ringing yet.

This is about to happen.

My point

Well, it's not Matt Shumer in the air anymore.

I read and translated the article in full detail. In response to the article, I say my views.

I'm not afraid - although the "count-down" atmosphere deliberately created in the text is really breathtaking. It's because of a strange feeling that this article depicts the future, and the moment I'm going through, there's something difficult to remember. I don't know what this stuff is. But it gave me a sense of fragmentation. The countdowns and the reality of my life described in the text, as if there was an insurmountable gap, was Matt Shumer exaggerating that influence? Is AI influencing his judgment? Maybe he's an AI fanatic? On the one hand, it's a zeal in the tech world, and on the other hand, it's a lack of attention (perhaps it's not the right word, but in my real environment and population, it's best known as AI and used as a browser). In the case of spring nights, China's largest IP, too, uses bean buns as a tool to make a difference and complete the human brain. The rest is gone.

Perhaps we are different from the United States, and there is still a great distance between the picture described by Matt Shumer and reality in the foreseeable future, but there is no doubt that the distance is getting smaller.

The core logic of the article is that AI is evolving at an index level, which will replace almost all cognitive work, and you must adapt immediately or you will be eliminated. This logical chain appears to be tight, but, in contemplation, it implies the assumption that human values are defined entirely by ** productive capacity.**

If we accept this assumption, then the conclusion must be pessimistic, because no human being can run an infinity near the zero-marginal cost of production.

But this led to a new core argument: If, one day, AI can really outperform human beings on all the indicators, why should they continue to produce?

It is not a technical issue, but a philosophical one, even a question of "means".

For thousands of years, human beings have worked because they would starve to death if they did not. After the industrial revolution, human work was driven by the need to purchase goods and services. But if in the future all goods and services can be delivered at very low cost by AI and robots, then the work itself will go from being essential to being an option**.

What will humans do then?

It's possible to do things that only humans find meaningful. For example, between two imperfect options, a choice is made by instinct. Like spending an afternoon with the kids building an ugly snowman. Like sitting by the river and watching the sunset go down a little bit.

Of course, all of these things, AI can do better, the snowmen can build up better, and they can be simulated, but the problem is not better, but ** this is what I should do as a human being.**

The article mentions that AI can simulate common sense, but cannot build real human trust. I do not agree with the latter sentence. AI, of course, can build trust - you can trust it when you talk to it, you can give a more precise and patient response than any friend. But that's a one-way trust. And trust among human beings is two-way, reciprocal and flawed.

And vices could be the last fortress of mankind.

AI's code is perfect, logic is spontaneous and response predictable. And humans hesitate, make mistakes, say wrong things, apologize and remain concerned after apologies. These "imperfects" constitute exactly the way we recognize each other.

The article concludes: "The future has come, but your door has not yet been ringing."

I do not deny the power of that. But what I'd like to say is, ** When the future finally knocks on your door, you open the door and see something that may not replace you, but rather ask you a question.**

The question is: Who are you when everything needs to be done without you? Who else could you be?

It also reminds me of the classic core argument that Lu Su-soo was facing Ji's fate: "Who am I?"

I don't have any ready answers. But I think it is worth thinking about itself. Use those days of AI liberation to think. Instead of spending these time worrying, chasing, trying to prove that it still works.

Finally, with regard to the one hour-a-day proposal for action in the article, I would like to add that, in that hour, apart from letting AI do your job, you can ask it those questions you dare not ask yourself. What would I really want to do if I didn't have to make a living? If all evaluation systems fail, do I recognize my value?

AI will not answer these questions for us. But it can help us to raise the voices of these issues so that we can no longer ignore them.

This, perhaps, is what it gives us, the most important thing.