Vector Search to the Rescue: How Dr. Avery Saved the World (with a Little Help from Moths)
Dr. Victor Avery's breath came in short, panicked bursts as he hovered over the lab counter, his gloved hands coated in a viscous mixture of neon-green fluid and strange bits of organic material. Alarms blared in the background, red lights spinning overhead. What should have been a routine experiment—merging chemical reagents with stray DNA fragments—had gone terribly wrong. A small glass vial labeled "Hyalophora-Chloroplast Compound" lay toppled on its rack, its contents boiling with toxic energy. In the air hung a faint, rotten odor, hinting at decay… or something far more sinister. Something alive.
Time was of the essence. Dr. Avery realized he had created a contagion which, left to its devices, could escape the laboratory and spread among the population faster than Dustin Hoffman could say Tootsie. He knew he needed to find a cure—and fast.
The Search for a Cure
Dr. Avery rushed to his desk, where two computers stood side by side: one with traditional fuzzy text search, the other with a new vector-based search that the IT guy Kevin—who never shuts up about AI—recently installed. "Now, I know in order to find a cure, I need to find DNA which is in the same family of the contagion's DNA. So let's search the lab's inventory to see what we can find."
First, I'll use the fuzzy search for the first ingredient—Hyalophora. The computer gives me back a list of similarly spelled inventory items:
- Hyaluronan (a cosmetic compound)
- Hydroxyproline (an amino acid)
- Hypochlorite (an ion commonly found in bleach)
"Maybe that's it... maybe that's the cure—Hyaluronan. It seems similar to what's in the compound. Fuzzy search works by seeing how many letters are similar to the word you're searching. So that must be what I'm looking for, right?"
Enter Vector Search
But hold on. What was Kevin saying? Vector search works by finding the relation between two words. So let me try the vector search database for the cure. If I type in Hyalophora, I get back:
- Callosamia (a moth)
- Actias luna (a moth)
- Columbia silkmoth (another moth)
Silk moth? Wait—these are all moths.
The Moth Man Cometh
Just then, Dr. Avery could hear the groaning of something inhuman outside the lab… evidence of his own nightmare come to life. The door flung open. The shadow of a... man? appeared. Dr. Avery called out, "Are you OK?" No response. The monster inched closer. Dr. Avery again called out, "Who's there?" The monster took a few more steps and walked under the last unbroken light, and there Dr. Avery could see—it was Kevin. "Kevin... are you OK?" But Kevin was not OK. He had been infected with the Hyalophora-Chloroplast contagion and was starting to transition into a six-foot-tall moth man.
Hang on, Kevin—your incessant droning on about AI is finally going to do some good in this world. As Kevin stares at the lamp, Dr. Avery leaps into action and starts mixing Callosamia moth DNA with some chlorophyll and throws it into Kevin's boba tea, which he also talks about a lot.
The cure is successful and the world has been saved—all thanks to vector database search.
The Moral of the Story
No moths or Kevins were harmed in the making of this article, and I'm pretty sure that's not how you cure a contagion, but hopefully that story helped explain what a vector database is and why you might want to use one.
When we started building Agile Luminary, we knew we needed to use vector search databases for searching User stories, project documentation, and reporting on your sprints.
Whether you're planning your next project, or saving the world from a lethal contagion, let Agile Luminary help you manage your workflow.