February 23, 2026
Nose-diving into drama
Scent, in Silico
AI wants to bottle smell — profit hype, privacy panic, perfume wars
TLDR: Big names are teaching AI to read and create smells, aiming at perfumes and safety tech like leak detection. Comments split between excitement for real-world sniffers and fears of creepy scent surveillance, with memes about “SmellOS” and debates over whether smell can ever be standardized like sight.
AI is coming for your nose, and the crowd has feelings. The story: tech giants like Google, scent startup Osmo, and fragrance titan Givaudan want to turn smells into digital code—useful for perfumes, and also for spotting gas leaks, rotten food, and even illness. But the top comment slams the hype for focusing on “making scents” over “detecting scents.” As vjanma puts it, “Think of it as speaker vs microphone”—everyone’s gawking at the speaker, while the microphone problem (finding and naming odors in messy air) is just as hard, and arguably more important. Cue fiery threads debating whether smell can ever get its own “RGB,” or if it’s just vibes forever.
Privacy alarms blared: if DARPA funds this and Estée Lauder sells it, will “smell tracking” become the new cookie? Perfumers showed up to defend natural ingredients, while eco-minded commenters cheered lab-made dupes for saving rainforests—then fought about whether the brain reacts the same way. Skeptics yelled “cilantro gene!” to remind everyone odor is personal. Meanwhile, jokes flew: “SmellOS update: patch notes nerf fishy breath,” “scratch‑n‑sniff CAPTCHAs,” and, of course, “SmellGPT.” Fans dream of sniffers exposing counterfeit goods; critics dread scent ads following your armpits. The mood: curious, divided, and extremely online.
Key Points
- •Smell lacks a standardized encoding comparable to color (RGB) or sound (Fourier transform), making formalization difficult.
- •Companies like Google, startups such as Osmo, and fragrance houses including Givaudan are using AI to digitize scent.
- •Investors including DARPA and Estée Lauder see practical applications in detection (gas leaks, spoilage, disease markers, counterfeit products).
- •Computational approaches could reduce dependence on natural perfume ingredients and enable novel scent creation.
- •Smell’s origins lie in bacterial chemosensation, with receptor proteins and semiochemicals guiding behavior across species.