Skip to content
RCreddit.com·

Everything Requires Skill.

AI summary

The user discusses how even advanced AI models like OpenAI’s Image Gen 2.5, despite having sophisticated underlying models, often produce "vibe coded" or underwhelming results when tasked with creative endeavors. They note that when asking the AI to create something like a landing page design for an app, the output rarely achieves a "masterstroke of design" or a "wow" factor, especially when the task relies on offloading creativity to the model rather than using it as a tool.

Time & source

Times shown in UTC

Display time zone: UTC

Local time zone unavailable; showing UTC.

PublishedOffset at this time: UTC+0Sep 12, 2026, 04:24 UTC

IngestedOffset at this time: UTC+0Sep 12, 2026, 12:01 UTC

Published
Sep 12, 2026, 04:24
Ingested
Sep 12, 2026, 12:01
Source type
Dev community
Tier
Community
Source status
Healthy

Tier is a per-source editorial setting, not a per-item score.

Article

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!

Source·reddit.com