PhotoCloak
Invisible to machines, unchanged to humans
About
The venture
PhotoCloak applies adversarial perturbations to photographs, rendering them unreadable by facial recognition systems while remaining visually identical to the human eye. The technology addresses a growing concern: how do you share images online without feeding surveillance infrastructure you never consented to?
Our work
How we supported
Every engagement is shaped by the venture's specific needs. Here's how we contributed to PhotoCloak.
Technical due diligence
We reviewed the underlying research and adversarial ML approach with external experts. The technology is sound, differentiated, and difficult to replicate without deep domain expertise.
Narrative development
Privacy tech can feel paranoid or niche. We helped frame PhotoCloak as a reasonable response to an unreasonable status quo, accessible language for a mainstream audience.
Partnership strategy
We identified potential integration partners: photo editing apps, social platforms, and privacy-focused browsers. Some conversations are ongoing; others have converted to pilot discussions.
Investor positioning
We advised on how to present the opportunity to investors who may not be familiar with adversarial ML. The deck now leads with the problem, not the technology.
Outcomes
Where it landed
- Long-running advisory relationship
- Partnership pipeline developed
- Fundraising materials refined
If you've already climbed the hard ones.
Let's talk about the next.
We take on a small number of founders and companies each cycle. Write to us with what you're building, who's on the rope with you, and where you're trying to get to.