Oct 22, 2019 · 21m · mad
Computing On Encrypted Data // Kurt Rohloff, Duality (Firstmark's Data Driven NYC)
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
At FirstMark's Data Driven NYC, Dr. Kurt Rohloff, CTO and co-founder of Duality Technologies, demonstrates how homomorphic encryption and lattice-based cryptography enable organizations to execute complex data science and AI analytics directly on encrypted data while maintaining strict privacy compliance.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Matt holds 3.5% of the talking time here. How this is scored →
Matt as informed peer 0.6 Guest teaching 2.0 Guest disagreement 0.2 Matt pushing back 0.2
speaking balance: gold is Matt, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 20:44 Politely correcting PKI substitution premise
In a very mild counter to an audience member's premise about homomorphic encryption replacing PKI, Rohloff teases apart the distinction, explaining that standard e-commerce PKI is not going away.
Hardest push from Matt ▶ 16:15 Probing single-player vs network use casesMatt Turck pushes past the general presentation by asking whether homomorphic encryption requires multi-company data sharing or if a single company can use it internally across departments.
Biggest teaching moment ▶ 15:15 Explaining differential privacy vs homomorphic encryptionRohloff educates the room on how differential privacy relies on adding noise to obfuscate data, which fails for fine-grained tasks like genomic research where homomorphic encryption excels.
Matt holds his own ▶ 15:06 Introducing differential privacy frameworkMatt Turck demonstrates domain awareness by introducing differential privacy as an alternative privacy-preserving analytical method and prompting the guest to compare the two.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Matt as informed peer | Guest teaching | Guest disagreement | Matt pushing back | Why |
|---|---|---|---|---|---|---|
| Addressing the Data Analytics and Privacy Conflict | 0 | 1 | 0 | 0 | This segment is part of Kurt Rohloff's solo presentation at Data Driven NYC, so host expertise and pushback are zero. Rohloff outlines the core trade-off between data privacy regulations and analytical utility, introducing homomorphic encryption as a solution. | |
| Balancing Data Utility and Privacy Without Zero-Sum Trade-Offs | 0 | 2 | 0 | 0 | Continuing his monologue, Rohloff explains the technical paradigm shift from binary operations to vector algebra primitives required for lattice-based crypto. Host activity remains non-existent during this monologue segment. | |
| Lattice-Based Cryptography and Post-Quantum Security | 0 | 2 | 0 | 0 | Rohloff details Duality's business strategy of open-sourcing the core cryptographic operations while commercializing data science wrappers like scikit-learn equivalents. The host does not speak in this monologue segment. | |
| PALISADE Open-Source Cryptographic Library | 0 | 2 | 0 | 0 | Rohloff discusses the open-source PALISADE C++ library and highlights practical deployments, including NIH rare-disease cancer research on encrypted genomic data. Host metrics remain zero due to monologue structure. | |
| Anti-Money Laundering and Financial Crime Collaboration | 3 | 3 | 1 | 1 | Matt Turck opens the Q&A by asking informed questions about differential privacy vs homomorphic encryption and single-player vs multi-party business models. Rohloff responds collaboratively, clarifying technological distinctions and answering audience questions politely. |
Statements from this episode (16)
Duality Technologies is fundamentally a data science company, not just security
“Though ostensibly duality technologies is really about security and cryptography, we're actually very much heart and soul at, a data science company, and how we can enable analytics on data that is protected and sensitive.”
Duality's encrypted computing tech originated from academic, DARPA, and DoD research
“So our solution is that we have a set of techniques that's, ah, been coming out of academia, out of DARPA, the DoD, where I personally came out of, and, ah, to enable computing on data while it's encrypted.”
Homomorphic encryption is on the cusp of solving major privacy issues
“And the core technology that we have is something that's called homomorph encryption and it's been growing in, in many different spaces and a diff, couple different areas, but it's enabling and on the cusp of addressing all these very, kind of, serious issues …”
Duality's co-founders include two MIT professors and a Turing Award winner
“And, ah, our, actually our co-founders include two MIT professors, including a Turing award winner.”
Traditional approaches to data security treat privacy and utility as zero-sum
“Generally there's been kind of like this all or nothing kind of approach, a zero sum game of providing a lot of privacy or providing a lot of utility from data, driving a lot of value.”
In 2009, an encrypted bitwise AND operation took half an hour
“Whereas in, in the old days, it would take, for example, you know, when the homework for crypto was first discovered, it basically took a half hour to run a bitwise AND on encrypted data.”
Duality runs sub-second genome-wide association studies on encrypted NIH data
“We're actually at a point now where we're actually encrypting genomic data for NIH and running full-on genome-wide association studies on encrypted genomic data with, ah, sub-second response times.”
Lattice cryptography is a leading candidate for NIST's post-quantum standard
“It's one of the leading candidates for the NIST's, ah, ah, public key standard coming forward for the post-quantum, post-quantum crypto.”
Lattice cryptography adoption is hindered by its unique compute model
“One of the reasons that so many, so few people have been able to actually use this technology so far is that it's a very, very different compute model.”
Duality open-sources its core homomorphic encryption and key generation operations
“We make all our, ah, crypto technology, the core crypto operations of key generation, encryption, decryption, everything else, open source.”
The best software engineers are lazy, designing modular and reusable systems
“The best software engineers are as lazy as possible, meaning they make things as modular as possible, and reusable as possible”
Kurt Rohloff developed the PALISADE cryptographic library over a decade with DARPA
“I personally have been working on it for a good 10 years or so primarily with support of DARPA and the other ARPAs in collaboration with folks at MIT and a number of defense contractors”
Duality enables encrypted anti-money laundering with the UK Metropolitan Police
“You know, another use case that we've actually had is that we're actually enabling anti-money laundering and fraud detection over encrypted data particularly as enabled by Duality Technologies running in collaboration with the CDA, a consortium of British bank…”
Differential privacy fails on fine-grained analysis like genome-wide association studies
“The challenge with that, it actually doesn't work very well if you need really fine-grained analysis as you would, for example genome-wide association studies.”