Generative AI: The Forest and The Trees
I am using the current iteration of artificial intelligence (AI) that is Generative AI and honestly, I love it.
One reason why I love Generative AI is that I can see the forest AND the trees, with a single knowledge base (the large language model) - I can query broadly (the forest) and deeply (the trees).
An example:
I want to get into music theory to help me understand music I listen to. So I start by asking for music theory basics.
- Pitch - the musical term for how high or low a note is.
- Rhythm - the timing and duration of musical sounds and silences.
- Harmony - the combination of multiple pitches played or sung simultaneously.
- Melody - a sequence of single notes that create a recognizable tune.
- Scales - a series of notes arranged in ascending or descending order that form the basis of much of Western music.
- Chords - a group of two or more notes played together that form the harmonic foundation of much of Western music.
- Key - the tonal center or home base of a piece of music, determined by the scale used.
- Form - the structure or organization of a piece of music, including its sections, phrases, and repetition.
- Dynamics - the volume or loudness of a musical sound.
- Texture -
That’s the broad stroke - and then I follow up with:
refers to the highness or lowness of a musical sound. It is determined by the frequency of sound waves, with higher frequencies corresponding to higher pitches and lower frequencies corresponding to lower pitches.
In Western music, pitch is typically represented by a series of 12 distinct pitches arranged in ascending or descending order, known as the chromatic scale. These pitches are named using the letters A through G, with sharps and flats used to indicate notes that fall between the main letter names.
Pitch is an essential component of melody and harmony, and is used to create contrast, tension, and resolution in music. It can also be used to create different moods and emotions, with higher pitches often associated with brightness, excitement, and happiness, while lower pitches are associated with darkness, sadness, and seriousness.
Musicians can manipulate pitch in a variety of ways, including by changing the tension on a stringed instrument, adjusting the position of valves or slides on brass instruments, or by changing the length of a tube or pipe on a wind instrument. Pitch can also be manipulated electronically using synthesizers and other digital instruments.
Focusing in on a topic, just a set of trees. Let’s focus on only one tree:
same song played in a different key. This means that the melody, harmony, rhythm, lyrics, and overall structure of the songs would be the same, but the pitch of the notes would be higher or lower.
For example, the song “Yesterday” by The Beatles has been recorded in a variety of keys, so we can use two versions of the song played in different keys as an example. The original version of “Yesterday” was recorded in the key of F, which gives the song a melancholy and introspective feel. If the same song were played in a higher key, such as G, the pitch of the notes would be higher and the song would sound brighter and more upbeat. However, the melody, harmony, and other musical elements would remain the same, with only the pitch of the notes changing.
After a quick search for these, I found these videos:
The Beatles - Yesterday (Key of F)
The Beatles - Yesterday (Key of G)
With this - I went from the broad to the detailed in just three questions! Getting examples solidified the concept in a way I would never thought of asking a master practioner in the space.
With Generative AI - asking these “why?” questions are great - the system won’t take it personally after the sixth “why?” or even sixtyth!
Large Language Models, LLM, power the current iteration of AI. I saw a post saying: “LLM - Lifelong Learning Machine” and from my experience with music, I whole heartedly agree!
What will you learn today?