Coarse is better

Different image models, different purposes

• Several commenters say the comparison is between mismatched tools: older Midjourney-style models as “art toys” vs Nano Banana Pro as a precise, business‑oriented image editor.
• View: NBP is optimized for prompt adherence, realism, and editing; MJv2 for striking, loosely interpretable “happy accidents.”
• Some argue both are useful in their own domains: NBP for marketing and production, older/coarser models for exploration and concept art.

What is art? Intent, emotion, and authorship

• One camp: art requires human consciousness, intent, and struggle; models are “meaningless image factories,” so their output is not art.
• Another camp: art is defined by the emotions it evokes; if synthetic images move people, that experience is genuine.
• Others emphasize intention targeted at evoking emotion; without that, almost anything—clouds, commutes, car crashes—would become “art,” which some find too broad.
• Photography, collage, and “found art” are used as analogies: selecting from random or generated outputs can itself be an artistic act, though some see this as thin, low‑bandwidth authorship.

Process, effort, and gatekeeping

• Strong sentiment that the meaningful part of art is the human process, difficulty, and accumulated skill; AI shortcuts feel empty or manipulative.
• Pushback accuses this view of elitism: if everyone can make high‑quality images, that doesn’t inherently devalue art; tools are “elevators” for expression.
• Underlying anxiety: corporations using AI to cheapen or replace human creative labor.

Prompting, “coarseness,” and model behavior

• Multiple comments argue the article’s prompts are poor and exploit quirks of old models; modern encoders interpret phrases like “British Museum” literally as a location rather than as an aesthetic tag.
• Coarse, impressionistic looks are still seen as achievable by explicitly prompting for visible brushstrokes, texture, and looseness instead of relying on older models’ fuzziness.
• Some suspect aesthetic “mode collapse” and over‑tuning toward glossy, ad‑like realism; others say we’re just in a transition where legacy prompt styles no longer work.

Broader automation and labor concerns

• Debate over whether AI‑driven automation will cause meaningful job loss or just shift roles, with historical arguments about Luddites and deindustrialization.
• Several note that automation’s harms are amplified by unequal wealth distribution rather than by the technology alone.