Everything Requires Skill.
用户指出,即使是像OpenAI的Image Gen 2.5这样先进的AI模型,尽管其内部包含复杂的模型,但在处理创意任务时,生成的结果往往显得“氛围编码”或不尽如人意。他们提到,当要求AI为应用程序创建登录页面设计时,输出很少能达到“设计杰作”或令人惊叹的效果,尤其是在任务依赖于将创造力完全卸载给模型而非将其作为辅助工具的情况下。
时间与来源
时间显示为 UTC
显示时区:UTC
本地时区尚不可用,暂时显示 UTC。
发布当时偏移:UTC+02026年9月12日 04:24 UTC
收录当时偏移:UTC+02026年9月12日 12:01 UTC
- 发布
- 2026年9月12日 04:24
- 收录
- 2026年9月12日 12:01
- 来源类型
- 开发者社区
- 档位
- 社区
- 信源状态
- 正常
档位是按信源手工设定的编辑判断,不是逐条打分。
Hello everyone, me again.
A lot of people liked my last post and felt it was well written and offered a good perspective. I hadn’t intended on posting again, but given the reception, I figured I’d share more.
Everything requires skill. For the foreseeable future, a better model will not replace a writer, a better image generator will not replace a human artist, and a better music generator will not replace a musician.
A good example is OpenAI’s recently released Image Gen 2.5, which has two different models under the hood. For all intents and purposes, I’ll treat them as one. When you ask it to create a landing page design for an app, you’ll often find that even the generated image itself looks “vibe coded.” You’d expect some masterstroke of design, something that wows you, but that’s rarely the case, especially when the task relies on offloading creativity to the model.
While I’m using web app mockups as an example, there are countless other scenarios where you cannot simply prompt your way into a better solution.
This isn’t a critique of OpenAI or its models’ overall capabilities. I think Image Gen 2.5 is a great jump, and I looked forward to it. But these shortcomings in creativity remain a consistent issue. Perhaps there’s a silver lining here, though. If I asked someone who loved designing pages to guide Astra, they could create the most beautiful things, because they can see what they want in their mind. They have experience and a perspective that can’t simply be mass-produced or replicated.
You may have heard the old saying, “The question is just as important as the answer,” or some variation of it. It isn’t enough for a model like Astra to be technically brilliant. It cannot see what you see, learn the way you do, or hold the full context of your experiences.
The universe seems to have a weird way of balancing things out. You can’t just let AI do everything for you. You have to use AI to help you do what you’re trying to do, and along the way, you’ll learn a lot.
Not to sound overly philosophical, but in a world full of artificial intelligence, having, acquiring, and growing organic intelligence is more valuable than ever. Being able to apply concepts to your work with AI, rather than expecting AI to supply all of that understanding for you, is an important skill that gets overlooked.
This isn’t to say AI can’t take jobs, or that it won’t eventually reach the point where it can do more of these things independently. Although, with music, I don’t believe it will ever fully replace the human connection.
When your favorite artist talks about their heart being broken, you know they, or even the ghostwriter, have actually felt that pain. No matter how great a generated song may be, I think there will always be an emptiness in knowing that the thing that made it doesn’t know what pain feels like. I might enjoy the song, but I’m not connecting with another person through it in the same way.
And that goes far beyond music. From art to cooking to writing, part of what we connect with is the person behind the work: what they’ve lived through, what they understand, and what they’re trying to share.
What I’m trying to say is that AI has unlocked a lot of potential for many of you reading this, just probably not in the way you anticipated. Calculators didn’t eliminate mathematicians. They allowed them to do more. The ceiling of what’s possible has been raised, while the path to reaching it has become shorter and easier but there is still a larger path beyond what the floor how now risen to.
That sounds like a hot marketing slogan an AI company would use, but there’s a lot of truth in it. It’s just more nuanced than “Use Astra to write your next paper” or “Use Image Gen to make your art.”
A better question I would implore you to think about is How can ChatGPT help you discover what you want to do, develop your skills, and bring out your potential? What could you, the person reading this, add to the world if you used AI to learn and refine whatever you’re drawn to?
Knowing how to know matters. So yes.
Everything still requires skill.
Would love to hear your thoughts below!