AI work vs human work
When something is done with the help of AI it is often considered “less valuable” compared to human only work: AI writing, AI code generation, AI generated images, AI translation. It’s hard to give a single-sentence-prompt-poem the same credit as work which took someone a week of thinking, trying different words, rhymes, sentence structures, even if the final artifact looks similar in size and even quality.
What it boils down to in the end is the effort and care put into your work. A half assed job is still a half assed job even if it was sprinkled with AI or if it’s 100% human work. A properly planned, executed, reviewed, and verified task is valuable even if AI was involved. If someone spent a week trying different prompts, combining and refining texts/images/you-name-it to make their work consistent, conform to an original style, fit the intended purpose, convey a coherent story – in my opinion it’s legitimate work.
IGORRR - ADHD was done with heavy AI usage, but it’s not a single-prompt-to-SORA slop, it’s a piece of art, which I want to return to and enjoy from time to time.
Now, however, things can take much less time. LLMs and AI Agents have given us a shortcut, like asking your computer genius friend why your PC started to misbehave (but now it’s always available and ready to help). Or how a good compiler output can make life of a programmer much easier.
The existence of LLMs doesn’t mean that quality suddenly became less important. It’s the operator – the human – who decides that quality can wait while more and more features are stuffed into the product (just like it always was). If you want, you can spend your token budget on making something better, more stable, secure.
But only if you want.