Hôm nay rảnh dùng thử AI (trước giờ chưa dùng cursor + agent mode của cursor)
Dùng thử mới thấy ngạc nhiên trong vòng 2h đã build được cái app game mà mình nghĩ bình thường chắc phải mất cả tuần mới làm xong.
ae xài thử cho vui, không cần cái gì phức tạp, chạy 2-3 dòng lệnh là boot được app trên web (mình chưa tối ưu cho phones)
Chỉ là một project làm cho vui cuối tuần thôi. AE nên tận dụng AI để code cho lẹ nhé.
Demo:
Well, that wouldn’t surprise me, because I’ve developed a few small apps with AI and seen how good they are (but not “very” optimal). What concerns me, however, is the “perilous” future of Generation Z developers. If they don’t become smarter than AI, they’ll be replaced by AI, like the painters in the auto body-painting shop.
they are (but not “very” optimal).
This is true! LLMs are still probability-based; they suggest the next best “code” for the problem.
If they don’t become smarter than AI, they’ll be replaced by AI, like the painters in the auto body-painting shop.
This statement is only true if AI is truly good; however, it is not at this moment. I believe that LLMs have reached their limit, and improvements will now be small incremental changes.
What I see is a much better future where human engineers can improve their productivity by leveraging AI to develop faster and more complex systems. While humans will still oversee the entire development process, they can delegate some mundane tasks to AI, like writing repetitive code. In this new age, future engineers will need to adapt and acquire new skills, as is typically required when new technology emerges.
Tóm tắt tiếng Việt: Dù gì đi nữa thì mọi người cần phải “adapt” với công nghệ mới, học cách sử dụng AI trong công việc.
giống như hồi xưa code trên notepad phải nhớ từng câu lệnh, sau này ra các ide thì chỉ cần nhớ vài kí tự rồi ctrl tab , giờ nâng cao hơn, dùng hẳn ngôn ngữ tự nhiên, dev sau này sẽ làm việc với vấn đề ở mức cao hơn, còn chi tiết thì AI lo
mà điều này hình như là mơ ước của nhiều dev ngày xưa ấy chứ, ko ngờ nó lại đến sớm vậy 
The word “ADAPT” is easy to pronounce but difficult to apply. Humans are the evolution of a biological process that means nothing less than "survival of the fittest"—or, in other words, existence through trial and error. Humans are therefore individuals and differ in their personal domains: social intelligence, scientific intelligence, linguistic intelligence, artistic intelligence, technical intelligence, etc.
While humans continue to oversee the entire development process, they can delegate some mundane tasks, such as writing repetitive code, to AI.
This is partly true, because different levels of intelligence determine individual survival. With AI, the world is gradually evolving into a three-caste system:
– the elites who develop AI,
– the adaptable who oversee AI production, and
– the ordinary people with adaptation problems who desperately struggle to survive in the shadow of the higher castes.
The resulting social problems are easy to envision: the elites are a tiny, wealthy minority, the adaptable are an income-earning minority, and the unadaptable are the majority on daily wages. This can be observed all over the world, even in Vietnam. The elites accumulate wealth, the adaptable work for the elites, while the unadaptable struggle to survive in perilous and low-value job.
Update: This BBC article gives us an insight into the unconditional adaptation of AI
This has been true even before the rise of AI; tech elites have long held power in the world. I don’t believe this is the issue we should be focusing on in our discussion ;).
Instead, we should concentrate on how to adapt this new technology to improve our lives. I prefer to look at the positive aspects rather than dwell on the negatives, as I believe this perspective contributes to making the world a better place. 
AI hiện tại chỉ tạo ra bullshit code thôi. Code 1 giờ debug 3 giờ thì không hiểu tối ưu chỗ nào. Người giàu lên chỉ có mấy công ty làm AI với phần cứng thôi. Như đọc sách làm giàu thì thằng giàu duy nhất là thằng bán sách ấy.
Một thời cũng tung hô no code/low code thay thế code này nọ.
Giờ đa số toàn mấy lão CEO lên tung hô AI này nọ…
AI đúng là tiết kiệm thời gian thật, nhưng phải biết cách dùn promt cho đúng và nhanh.
@nguyenhuuca
You’re right about the big-mouthed CEOs, who are usually lawyers or MBA graduates. They’re not concerned with efficiency or cost savings, but with maximizing their “dollar bonus.” When a CEO fails, he or she always receives a golden million-dollar parachute, and the blame is shifted onto the employees who are too dumb for his or her grandiose idea. Examples:
– A human resources manager uses AI to select applicants. Sounds great and tech-savvy. But the unspoken question is: if so, why does a company still need a human resources manager?
– A CEO uses AI to reduce operating costs and maximize company profits. The same unspoken question is: Why doesn’t the company founder use AI to eliminate one of the most expensive costs: the CEO’s compensation and salary (usually in $-millions) ?
These examples only show us how AI is used to intimidate and alienate employees in the interest of CEO’s profits.
AI can only do what it’s taught to do. If the knowledge imparted becomes outdated, AI is automatically outdated. Emotional, combinatorial, and inventive abilities are uniquely human. No AI can be replaced in the slightest. Those who rant loudly about AI are simply dull AI acolytes.
83% thành viên diễn đàn không hỏi bài tập, còn bạn thì sao?
