Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? No matter how strong AI is, it cannot replace the judgment ability that comes from hands-on practice.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? I came across a report on Tsinghua University's "First Lesson of the New Semester" given by the university president a few days ago when the new semester started, and I found it quite interesting, so I'd like to discuss it with you today.
Starting with Tsinghua University President's "First Lesson of the New Semester": Large models can program, so why do students still need to learn themselves? To begin with, this sounds like a speech for graduate students at Tsinghua University, but the final point of this speech, I think, is relevant to each and every one of us. Especially if you're using AI in your work every day, there's quite a bit of inspiration to be found.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? On August 27, after the opening ceremony of the 2026 graduate class at Tsinghua University, President Li Qiming gave a lecture titled "Standing at the Forefront, Open and Inclusive, Cultivating Solid Abilities through Unremitting Efforts" at the Comprehensive Gymnasium.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn how to code themselves? The structure of this lesson itself is what I want to talk about today.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? Resources are being maximized, but the requirements are focusing on basic skills.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? He began by laying out resources, discussing Tsinghua University's forward-looking and comprehensive layout in the field of artificial intelligence, and the latest developments in how AI empowers teaching, scientific research, and discipline construction.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? A freshman commented after class that the school has already made giving play to the enabling role of artificial intelligence a key task in its "15th Five-Year Plan".
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? The resources provided are extremely abundant, almost as if the entire university is laying out all its assets to tell you: you've caught the best time.
Starting with Tsinghua University president's "first lesson of the semester": large models can now program, so why do students still need to learn themselves? However, I have some reservations about this. A few years ago, I attended an EMBA program at Tsinghua and found that many internal information systems were actually not very user-friendly. But as a university, it's probably unrealistic to expect too much in this regard.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? After laying out his expectations, he summarized the requirements for students into four key points: enhance problem awareness, enhance evidence-based capabilities, enhance judgment capabilities, and enhance sense of responsibility. Additionally, he emphasized the importance of adhering to long-termism and working hard in a down-to-earth manner.
Starting from Tsinghua University President's "First Lesson of the New Semester": Large models can now program, so why do students still need to learn themselves? The discrepancy between the resources provided by the school and the expectations placed on students is the most striking aspect to me. The school provides top-notch resources, but the requirements for students are pushed back to the most basic level.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? The impression I was left with from these requirements can be summed up in four words: work diligently and honestly. Of course, I want to emphasize that this is my own condensed expression, not the president's original words.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? But think about it, isn't it just about honestly speaking with facts and data, and being able to take responsibility?
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn how to code?
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? However, just this one sharing is still just a gesture. What really caught my attention was another speech given by Li Luoming on June 27, at the undergraduate graduation ceremony. There is a full transcript of that speech, which revolves around AI throughout.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? He said that since this year, the school has held a series of seminars on talent cultivation in the intelligent era, and teachers have reached a high degree of consensus on one issue: although the programming capabilities of large language models have become very powerful, Tsinghua's science and engineering students still need to learn programming.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? This is not just a casual remark by the president, but a consensus reached by teachers at a seminar.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn how to code themselves? I was quite surprised when I read this. Because in this era where AI has become a dominant field, it actually seems more in line with the trend to not learn programming, or to just muddle through with a few vague phrases.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? However, teachers unanimously believe that even with AI becoming increasingly powerful, there is still a need to learn programming.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn how to code themselves? And this is not just a figure of speech. Official speeches are cost-free, you can say something like "stay passionate, look up at the stars" on stage, and everyone will applaud, but they will forget it all after a good night's sleep. However, behind this statement, a group of AI experts have held a meeting to discuss the issue.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? The attendees of the meeting were those who understand large models the best in China, are closest to the forefront, and have the most resources at their disposal.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? It's worth noting that, whether it's data, algorithms, or computing power, from hardware to software, Tsinghua's technological accumulation in this regard is absolutely formidable. Consider that the country's top large models, such as Zhiyun, actually originated from Tsinghua's laboratories in their early stages.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? They are fully aware of the level of code that current models can produce, and they also know how much more time it takes for students to continue writing code by hand. However, experts are highly unanimous: students should still learn.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? If you ask someone from an AI company about this, they might think it's not cost-effective. Machines can already do the job faster and better than you, so what's the point of having a twenty-year-old relearn everything from scratch?
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? But I understand that what the school is after is for students to have actually gone through that process themselves.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? AI can provide answers, but it can't take that journey.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? Li Luming said it more directly at the graduation ceremony. He said that completely outsourcing thinking to large models may seem convenient and efficient, but in reality, it may lead to weakened cognitive abilities.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? He also quoted Mr. Yao Qizhi, saying that in the context of artificial intelligence replacing more and more cognitive labor, many abilities still need to be formed through hands-on practice and taking the difficult path.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? Understanding must come after experiencing chaos.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? In May 2025, I came across a study on brain neurons that mentioned the principle of brain intuition. It said that the reason you can remember something is not because you saw the answer, but because your brain had previously experienced being stuck.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? You have to start with a mess, and then that mess can be reorganized. At the moment of reorganization, the original neural activation pattern in the brain will undergo a structural switch, as if switching to a different channel, and that's when knowledge is engraved into long-term memory.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? So, if you directly jump to getting answers from AI, what you're skipping is not just the process of "thinking", but possibly the entire pathway for your brain to form long-term memories.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? AI responses may create fake epiphanies.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? I once noted a sentence in my notes, but to be honest, I don't quite recall whether I wrote it myself or saw it somewhere else. It mentioned that AI responses have three characteristics: logical consistency, fluent wording, and clear structure.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? These three things combined are likely to be an effective engine for manufacturing pseudo-epiphanies.
Starting from the topic of Tsinghua University president's "first lesson of the semester": Now that large models can program, why do students still need to learn it themselves? You may feel suddenly enlightened after listening to it, but that feeling is likely to be false. You've just been persuaded by AI, not truly understood it. Because you haven't gone through the process of piecing together the puzzle from chaos to clarity.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? This is actually the reason why elementary and middle school education has always placed great emphasis on correcting wrong answers over the years. Because only by doing it wrong first can you find the problems in your thinking; and only by finding the problems can you solve specific problems.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? Being persuaded and truly understanding are fundamentally not the same thing in the brain. One is information that passes by and is gone. The other is memory that is etched in.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? Looking back at Tsinghua's decision, it's different. Letting students continue to write code by hand is not about retaining the code itself, but about retaining the mental struggle that occurs during the coding process.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? Code can be generated by models, but the journey cannot be taken by models on your behalf.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? AI is a magnifying glass, not a balancer.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? Li Luoming also quoted a student's words during the graduation ceremony, saying that AI is a magnifying glass, which can amplify abilities, but also amplify laziness. The student made this comment, and Li thought it was well said, so he brought it up on stage.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn how to code themselves? Judgment will also be greatly amplified.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? I particularly like the phrase "magnifying glass". AI is not an equalizer in many cases, it won't simply bring insiders and outsiders to the same starting line.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? If you have judgment, it helps you apply that judgment to ten projects. If you don't have judgment, it helps you apply the lack of judgment to ten projects as well.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? This is no longer just a matter for schools.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? I previously came across a case where Deloitte Australia was commissioned by a government department to conduct a compliance review report, with a contract price of 440,000 Australian dollars. After submitting the report, it was discovered that it contained fabricated court judgments and several non-existent academic papers, resulting in a partial refund of the payment.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? This report was originally tasked with evaluating whether a system was compliant. However, a report on compliance inspection itself failed to pass AI inspection.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? Domestic government departments have also openly written to remind people of the same thing, saying that AI has increased the speed and quantity of materials, making the data appear detailed, but in reality, it cannot withstand scrutiny.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? You see, the people writing materials are not lacking in tools, but rather in someone who can read through it before submission and spot the problems at a glance.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? The ability to evidence-based reasoning is the most important.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? Therefore, in the "First Lesson of the New Semester", among the four words mentioned by Li Qiming, from my own perspective, the most important one is actually evidence-based ability, or in other words, whether you can trace back and verify the source of a statement.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? Everyone is talking about having a questioning mindset, and judgment ability sounds like a slogan, but the two words "evidence-based" are very specific: it means tracing back to the point where it stands up to scrutiny.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? This ability is not about knowledge, but rather a habit, and it can only be developed by practicing it personally.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? When you see the perfect results given by AI, you should also ask it to provide the data or case sources, and then spend some time to verify whether these sources are reliable.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? However, if these words are only applied to students, it would be a waste.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn how to code themselves? If you remove the identity of being a student, the situation is the same as ours: tools have become tens of times more powerful than in the past, yet the work we submit is more easily seen through at a glance.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? Only by doing it personally can one have judgment and a sense of touch.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn how to code themselves? This also reminds me of a small incident that happened in my life over the weekend. My partner and I went to a nearby building materials market, where we actually just needed to replace a small cabinet.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? We had AI produce a very detailed specification document in advance, with various indicators written in great detail. For example, the mirror on the surface, AI told me that I must ask my boss to provide a super-white, copper-free, and silver-free mirror.
Starting from Tsinghua University President's "First Lesson of the Semester": Now that large models can program, why do students still need to learn themselves? I told my boss what was required and had him write it into the order. He looked at me with a smile and said, "Was it Dou Bao who told you this?" Then he pulled me to a wall where many mirrors were hung.
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? He pointed and said, do you know which one is the mirror you want? Should I give you a ultra-white one or a regular white one, how will you judge?
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn themselves? I was stuck right there. I had the parameter report given by AI, but I couldn't judge it at all.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn how to code themselves? I brought this up because it's actually similar to today's topic. Why do we still need students to learn programming in the AI era?
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? The key point is not to make students suffer unnecessarily, but rather, it's only by doing it yourself that you can develop a different kind of judgment and tactile sense about the results.
Starting with Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? Of course, it's not just programming - many scenarios are similar. For example, no matter how advanced AI-generated art is, it can't replace some designers' hand-drawn work, because what's being cultivated in the process is a sense of beauty.
Starting with Tsinghua's president's "first lesson of the semester": If large models can already code, why must students still learn to program themselves? Related link
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? 1. University President Li Luoming delivered the "First Lesson of the New Semester" to Tsinghua University's 2026 graduate students
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? https://www.tsinghua.edu.cn/info/1177/127921.htm
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn it themselves? 2. President Li Qiming's speech at Tsinghua University's 2026 undergraduate commencement ceremony
Starting from Tsinghua University President's "First Lesson of the New Semester": Now that large models can program, why do students still need to learn programming themselves? https://www.tsinghua.edu.cn/info/3289/127053.htm