Tolkien Forum: 'AI Now Outperforms 98% of Humans'

AI Now Illustrates Tolkien as Creative Sector Counts the Cost

Can AI paint better than 98% of humans? In a forum for Lord of the Rings fans, the answer circulates without nuance and with proof: fantasy images generated in seconds with Midjourney and similar programs. The overarching argument is broad—writing, music, image, video, and 3D rendering in any style and in an incredibly short time—and it always ends the same way: the technology has barely begun.

The classic counterargument, that the machine has no soul, is dismissed as relative. The models don't improvise: they abstract millions of human works and, with millions of parameters, they increasingly succeed at what they are asked. With that foundation, the question ceases to be artistic and becomes accounting.

What Image AI Does Today and What It Misses

The tools have split into two leagues. Dedicated ones—Midjourney, Novel AI, Mage—produce illustrations with a commissioned finish. Generalist ones, like Grok, Gemini, or ChatGPT, yield decent results when the user knows how to ask. With a good AI, the enthusiastic side argues, you can even simulate an oil painting.

The critical side looks at the details and finds seams. Characters with an extra leg. The Cameron blue in armor. The hazy halo around faces. Overlapping textures no painter would apply, decorative brushstrokes without volume or thickness, absent gradients. These are the factory defects, and anyone who has flipped through many art books detects them immediately.

The forecast, even from skepticism, is that this list will shorten week by week. When the technique no longer betrays itself, judgment will have to be made another way: through personal style, which still doesn't exist. For now, the machine reproduces others' styles, and it's assured that training it with a few high-resolution images is enough to clone any author's style.

The Jobs That Are Leaving: From Ten Artists to Two

The most repeated calculation isn't about art, it's about payrolls. Where ten artists used to work, two artists and two prompters will work, producing triple. The conclusion is harsh and admits no embellishment: far more jobs will be destroyed than created.

The figures circulating are difficult to verify and easy to repeat. One claims that the previous year one million intellectual, artistic, and creative jobs were destroyed in the United States alone. Another speaks of works made between 90% and 100% with AI slipping into catalogs and commissions as their own.

Specific cases are more eloquent than any statistic. According to one case mentioned in the thread, a company producing interactive content for children laid off the person who created the images. And in small and medium-sized businesses where only a logo or a simple sketch is needed, the question is no longer who will do it best, but who will pay hundreds or thousands of euros for something an AI delivers for free.

Who Will Pay a Dime for an AI-Generated Illustration?

Here the consensus breaks down curiously. Even among proponents, it's admitted that generated content has no market: there's no original to buy, only a printed sheet.

Some argue that graphic design was never a stable profession, but something circumstantial tied to specific technology, and that illustration has more potential because it retains a component of craftsmanship. At the other extreme, it's stated that an artist with years of experience and above average skill has already thrown in the towel. Hard times for lyricism and for living, or half-living, from art.

Suno, Grok, and the Book of Kings: Uses Almost No One Disputes

Outside the labor dispute, practical uses are accumulating. One father recounted generating an illustrated book for his daughter with AI: the story of two siblings saving axolotls, an adaptation of Campbell's monomyth that works because it resonates with the unconscious, with the children's traits described in words and a kawaii style. In minutes, he had the formatted manuscript, with minor corrections.

A 3D animator with nearly 30 years of profession plays with Grok and reuses material that already works. And the case of a person with higher music education who, due to lateral paralysis, cannot play an instrument and has found a way to compose again with Suno is related. He's already on track for a quadruple album.

In music, moreover, the technical discussion is different. Harmony trinc recognizable patterns, and composition has always relied on trial and error, two areas where the machine moves comfortably. Millions of tracks can be published per minute, and not all of them bad.

In the end, the dispute that began by asking if the machine has a soul has devolved into something more uncomfortable. If a program does 98% of the job well, perhaps the problem isn't the machine's sensitivity, but that of what we claimed to admire.