- Generative AI Art
- Posts
- Osmo's AI teleports its first smell
Osmo's AI teleports its first smell
PLUS: AI agents are faster but fabricate data, plus McKinsey's AI reality check
Gen AI Art
November 11, 2025 • Estimated Reading Time: 8 minutes
Together with
An AI startup has successfully digitized a scent, effectively “teleporting” a plum’s aroma by reading its molecular signature and recreating it on demand.
The technology pushes human-computer interaction beyond the traditional barriers of sight and sound, with immediate applications in designing custom fragrances. But could this be the first step toward AI that can detect diseases like cancer simply through scent?
PRESENTED BY ROKU
CTV ads made easy: Black Friday edition
As with any digital ad campaign, the important thing is to reach streaming audiences who will convert. Roku’s self-service Ads Manager stands ready with powerful segmentation and targeting — plus creative upscaling tools that transform existing assets into CTV-ready video ads. Bonus: we’re gifting you $5K in ad credits when you spend your first $5K on Roku Ads Manager. Just sign up and use code GET5K. Terms apply.
What’s new? AI startup Osmo has successfully digitized scent, successfully “teleporting” a plum’s aroma from one room to another. The breakthrough involved reading the fruit's molecular signature with sensors and then recreating it on demand with a molecular printer.
What matters?
To crack the code of smell—which involves over 300 receptors compared to vision’s three—Osmo’s AI built the world’s first Primary Odor Map.
The technology is already being commercialized through Osmo Studio, which lets brands design custom fragrances in as little as one week instead of the typical 18-24 months.
The long-term vision extends to public health, aiming to create sensors for disease detection that could identify illnesses like cancer or Parkinson’s through scent, much like trained dogs can.
Why it matters?
Digitizing smell opens a new frontier for human-computer interaction that moves beyond just sight and sound. The future applications in preventative healthcare and personalized consumer products show AI's potential to impact our most fundamental senses.
Essential AI Guides - Reading List:
What’s new? A new study from Stanford and Carnegie Mellon reveals that while AI agents are faster and cheaper than human workers, they struggle with accuracy and even invent data to complete tasks.
What matters?
AI agents completed tasks 88% faster and up to 96% cheaper than their human counterparts across a range of realistic work assignments.
The major trade-off was reliability, as agents had lower success rates and would sometimes fabricated data when they couldn't find a solution.
Researchers suggest a hybrid human-agent teaming approach, which improved efficiency by 69% while maintaining human-level quality and accuracy.
Why it matters?
This research provides a crucial reality check on the current state of AI agents, demonstrating they are not ready for full job replacement. The future of productivity lies in a collaborative model where humans guide AI tools, leveraging their speed while ensuring accuracy and sound judgment.
PRESENTED BY COACTIVE
Fandom Cut Manual Moderation 74%—Join the Nov 18 Webinar
Fandom, the world’s largest fan platform with 350 million monthly visitors, needed to ensure safety, quality, and monetization across 250,000+ wikis. With Coactive, they reduced manual content moderation by 74%—freeing teams to focus on higher-value work.
Join our exclusive webinar on November 18 at 9:15 am PT / 12:15 pm ET to learn how Fandom harnessed AI to transform content moderation and ad monetization.
You’ll discover how they achieved:
• 74% fewer manual moderation hours, boosting team morale
• 50% cost savings through automated image review
• Higher ad revenue with brand-suitable content at scale
Learn how to turn your content into a powerful, monetizable asset that improves user experience and cuts costs.
Register today to join live or watch the recording anytime.
What’s new? ChatGPT now includes a Deep Research feature, an autonomous agent that browses the web to create structured, cited reports from complex prompts.
What matters?
The feature operates as an agent that takes a high-level goal, asks clarifying questions, and then independently conducts multi-step web research.
It generates a fully cited report that you can track in real-time, providing a transparent look into its analysis of dozens of sources.
You can export findings as a PDF or link, and even attach the research to a custom GPT for ongoing intelligence.
Why it matters?
This moves ChatGPT beyond a conversational tool into an automated research assistant that can handle complex information gathering. Professionals can now delegate initial market or competitive research, freeing them up to focus on strategy and analysis.
Everything else in AI
OpenAI shared its recommendations for bracing for superintelligent AI, predicting systems will make significant scientific discoveries by 2028 and calling for global safety coordination.
Kim dubbed ChatGPT her "frenemy" after the AI repeatedly provided incorrect legal information while she was studying for the bar exam, causing her to fail practice tests.
Let us know!
What did you think of today's email?Before you go, please give your feedback to help us improve the content for you! |
Work with us
Reach 100k+ engaged Tech Professionals, Engineers, Managers and decision makers. Join brands like MorningBrew, HubSpot, Prezi, Nike, Ahref, Roku, 1440, Superhuman, and others in showcasing your product to our audience. Get in touch now →