Personal Cognition
Personal Cognition is a systems engineering firm with a public benefit mission. We’ve implemented and operated some of the largest cloud and enterprise computing infrastructure ever built. We help companies build infrastructure that is reliable, efficient and independent of any one cloud provider or jurisdiction.
We use the same foundation to build privacy-first software for individuals. Your data stays under your control. AI runs on your devices, and the applications work even if we shut down.
We use artificial intelligence carefully: to teach, and to improve the quality and reliability of the computing infrastructure we build.
What we can help with
We have experience building and scaling data management infrastructure:
- Outages and elevated data loss risk as distributed databases and storage grow
- Dependence on a single cloud provider or region
- Data residency and sovereignty requirements
- Increasing infrastructure costs
- Finding bugs in distributed systems before they show up in production
We provide assessments of new or existing infrastructure, embedded engineering work and infrastructure builds. We offer discounted pricing for open-source work that supports privacy-first applications and digital sovereignty, and for early-phase startups.
If you are interested in working with us, send an email to contact@personalcognition.com.
Technology
Our privacy-first software and consulting work makes heavy use of simulation testing, formal verification and declarative network protocols.
We are reimplementing cloud computing applications from the ground up with fault tolerance and cost-effectiveness in mind. Modern storage systems and databases are designed to operate for centuries without data loss, but are typically deployed in environments designed to last, at best, for decades and are not resilient against common failures. Equally bad, administrative errors or anti-abuse heuristics routinely kick tenants off those platforms.
Our offerings are fundamentally multi-cloud. We provide tools that validate small cloud computing firms so our customers are not limited to the largest clouds. Our applications are architected for true multi-region and multi-provider redundancy on day one. We are careful not to lock in our clients: We provide open deliverables designed for portability to other vendors, and do not introduce proprietary dependencies on our other offerings.
Privacy and sovereignty
Operating surveillance infrastructure is expensive and fraught with risk. Most existing tools were built for companies whose business model is surveillance. That suits those firms but imposes unnecessary cost and risk on everyone else.
They assume sprawling-yet-centralized data management infrastructure, impose onerous regulatory compliance processes, create unbounded liability in the face of data breaches, and force operators to write user-hostile terms of service agreements. Building systems that respect customers’ privacy sidesteps all of these issues.
Choosing providers that operate in appropriate jurisdictions is similarly simplifying: The US CLOUD Act and GDPR have created an impossible legal dilemma for multinational cloud providers: Each jurisdiction can force them to break the laws of the others. Ongoing cases are likely to worsen these problems: In the past, the industry thought it was legal to partition their businesses into local subsidiaries that obey local law. A Canadian court has ruled otherwise, and it will take years or decades to work these issues out.
We are building privacy-preserving alternatives not because it is easy, but because the status quo is unsustainable, and because it is the way we want computing to work in the future.