Imagine effortlessly remixing your favorite tunes, adding or removing instruments, all with simple text commands. This isn't science fiction; it's the reality AI music editing is bringing to life. Recent research has focused on giving us more control over AI-generated music. Systems like Loop Copilot let users generate music loops and refine them through conversation, effectively conducting an ensemble of AI musicians specialized in different tasks. A 'Global Attribute Table' keeps track of changes, ensuring smooth transitions and consistent musical attributes during editing. But what if you want to edit existing music? MusicMagus enters the scene, offering 'zero-shot' editing with diffusion models. By manipulating the AI's understanding of music, it changes attributes like genre or instrumentation without needing any retraining. Want to add a saxophone solo to a piano piece? Just type it in. While MusicMagus excels at swapping instruments or changing styles within a stem, it struggles with more complex edits like adding entirely new instruments. This is where Instruct-MusicGen shines. By combining text commands with audio input, it allows for sophisticated modifications, adding, removing, or tweaking individual instrument tracks. Instruct-MusicGen uses a dual approach, fusing text instructions with audio data to precisely guide the editing process. Tests show it's surprisingly good, producing high-quality edits that closely follow text commands. Although there are still hurdles to overcome, like achieving perfect signal-level accuracy and reducing reliance on large training datasets, these advancements suggest a bright future for music production. Soon, musicians and producers might rely less on complex software and instead use intuitive text commands to shape their sound, opening a new world of creative expression.