A Taste of the Future
Episode 5: From Stories to Songs (and Maybe Taste?)
What if you could type:
“The taste of grandma’s apple pie… with a hint of starlight.”
And instead of a recipe, the AI handed you a little flavor pod that actually let you taste it? That might sound far-fetched, but so did text-to-images a few years ago.
We’ve Done This Before
First came text-to-story — you give an AI a prompt, and it spins out a tale.
Then came text-to-image — a single sentence can conjure a picture of a whale running across a desert or a horse made of origami.
In each case, you describe what you want, and the AI fills in the missing details.
What’s a “Modality”?
A modality is just a fancy word for a way of expressing information — like text, pictures, sound, or movement.
Even though these feel very different, the AI systems that generate them share similar underlying technology. Improvements in one area often spill over to others. That’s why AI progress feels like it’s happening everywhere at once.
A Few Modalities (So Far)
Text → stories, summaries, code
Images → art, photos, memes
Music → songs, jingles
Video → short clips, even movie trailers
Worlds → 3D spaces for AR/VR
Biology → proteins and food design
Someday? → Taste
Spotlight on Suno
This week I wanted to examine Suno, an AI that makes full songs from text prompts.
🎤 Big headlines: Recently, R&B artist Xania Monet secured a $3 Million recording deal. The thing is Xania Monet is an AI Artist. Xania’s creator is poet Telisha Jones who used an artificial intelligence based music generation service called Suno to create songs that are landing on Billboard’s charts.
⚖️ Big trouble: Meanwhile, flesh and bones artists are up in arms and major record labels are suing Suno, arguing its training data used copyrighted music without permission. Read more ›
🎧 Big fun: You can try it for yourself — free — at suno.com. When you sign in they will provide you a few credits that renew each day. Please note that if your Suno experimentation leads to a hit song, give me a call-out at the Grammies.
Closing Loop
So if we’ve already gone from text-to-story → text-to-image → text-to-song, what comes next?
Maybe one day instead of asking AI for a playlist, you’ll ask it for a flavor list:
“Alexa, give me a snack that tastes like victory after a hard-fought soccer game.”
Text-to-taste might not be so far-fetched after all.