Building systems that must survive the next era of computing, and breaking them first, so no one else can. Working at the intersection of post-quantum cryptography, artificial intelligence and offensive security.
Every encrypted message is a promise that it stays private for as long as it matters. These three figures show where that promise comes from, how a quantum computer breaks it, and what keeps it.
Multiplying 61 by 53 is instant. Recovering 61 and 53 from 3233 is the hard direction, and with a 2048-bit modulus (617 digits) it is hard enough to protect most of the internet. RSA turns that asymmetry into a lock anyone can close and only the owner of the factors can open.
Shor's algorithm turns factoring into finding the period of ax mod N. With a = 11 the period is 52, and gcd(1126 ± 1, 3233) returns 53 and 61. In 2025 Craig Gidney estimated that RSA-2048 could be factored in under a week with fewer than a million noisy qubits. Traffic recorded today only has to wait.
ML-KEM, standardized by NIST in August 2024, rests on the Module Learning with Errors problem: a lattice point blurred by deliberate noise, in 768 dimensions for ML-KEM-768 instead of the two drawn here. There is no known period for interference to amplify. In NIST's words, it is believed to be secure even against adversaries who possess a quantum computer.
Toy parameters for legibility. The arithmetic shown is exact.
Every computer ever built, from the abacus to the exascale supercomputer, works on the same assumption: the machine holds one definite state and steps to the next one. Quantum computing drops that assumption and computes with the rules physics actually runs on, where a system holds many states at once and those states can cancel each other out. It is the biggest change in what a machine can do since the transistor, and it is the reason everything else on this page matters.
A bit is 0 or 1. A qubit is a controlled combination of both, so n qubits carry 2n amplitudes at the same time. Add one wire and the state space doubles. That exponent is the entire story.
Two qubits can share a single state that cannot be written down by describing each of them separately. Correlations appear that no classical arrangement of parts can reproduce. Einstein called it spooky and hoped it meant the theory was incomplete. The experiments disagreed.
The part that never makes the headlines and the only part that does the work. A quantum algorithm is choreography: arrange the amplitudes so the wrong answers cancel and the right one survives to be measured. Shor's algorithm is that trick aimed at periodicity, and RSA is built on periodicity.
Classical hardware grows one transistor at a time. A quantum register grows one exponent at a time. Twenty more wires is a million times more state. Watch where that goes.
Small enough that a pocket calculator keeps up.
The honest footnote: this is not parallelism, and anyone selling it that way is selling something. You never get to read those states. Measure the register and it collapses to a single answer. The whole discipline is the craft of making that one answer the one you needed.
Two things follow from that exponent. One is a promise: chemistry, materials and drug design stop being guesswork and become computation, which is why quantum sits at the centre of the longevity argument further down this page. The other is a deadline, because the same machine that simulates a molecule also factors a key. That clock is already running.
A cryptographically relevant quantum computer will break RSA and elliptic-curve cryptography, the mathematics protecting nearly everything online today. Adversaries are already stockpiling encrypted traffic to decrypt later. This is harvest now, decrypt later, and it means the migration deadline is not Q-Day. It already passed.
Per NIST IR 8547: RSA-2048 and ECC at 112-bit strength become deprecated by 2030.
By 2035, RSA and ECDSA are fully disallowed for U.S. federal systems, and the world follows NIST.
Encrypted data intercepted today (medical records, state secrets, financial history) will be retroactively broken the day a large quantum computer exists. Long-lived secrets are already at risk.
In August 2024, NIST finalized ML-KEM, ML-DSA and SLH-DSA, the first post-quantum standards. The tools are ready. What most organizations lack is a map of where their cryptography lives.
At CYTE, Samuel's team built the Quantum-Safe Observatory, a live scanner tracking how ready the internet's infrastructure really is for the post-quantum transition.
Each of these looked impossible a decade before it happened. Notice the spacing between them shrinking.
Peter Shor proves a quantum computer could factor large numbers efficiently: a theoretical death sentence for RSA, written 30 years before the hardware to execute it.
Google's Sycamore performs in minutes a sampling task estimated to take a classical supercomputer millennia. The debate over the exact numbers misses the point: the era of quantum advantage begins.
DeepMind solves a 50-year grand challenge of biology. In 2024 it earns a Nobel Prize, the first awarded for a discovery made by an AI system. Biology becomes a computational science.
Large language models jump from research papers into daily life. Within two years, frontier models reason through mathematics, write production code and operate as autonomous agents.
NIST publishes the first post-quantum encryption standards. Months later, Google's Willow chip demonstrates error rates that fall as qubits scale, crossing the key threshold for fault tolerance once thought decades away.
Microsoft unveils topological qubits with Majorana 1; IBM publishes a concrete roadmap to large-scale fault-tolerant machines before 2030. Quantum computing shifts from "if" to "when", and "when" keeps moving closer.
The day a cryptographically relevant quantum computer runs Shor's algorithm against real keys. Nobody knows the date. Everybody will remember it.
The same exponential curves rewriting cryptography are converging on biology. Treat aging not as destiny but as a disease, with causes, mechanisms and, eventually, treatments. These are the three engines that could get us there.
AlphaFold mapped 200 million protein structures (every protein known to science) in months. AI systems now design drug candidates, simulate cell behavior and generate hypotheses faster than any lab can test them. The bottleneck of discovery is dissolving.
Chemistry is quantum mechanics, and classical computers simulate it terribly. Fault-tolerant quantum computers will model molecular interactions exactly, turning drug design from trial-and-error into computation. The first killer app of quantum computing may be medicine, not code-breaking.
Molecular machines that patrol the bloodstream, repair DNA damage and clear senescent cells. Today it's targeted drug delivery and CRISPR; tomorrow it's programmable matter operating at the scale where aging actually happens.
"The first person to live to 150 may have already been born. The interesting question is not whether these technologies converge. It's whether our security, our ethics and our institutions will be ready when they do."
Samuel Sabogal Pardo, on why he works on all three.
~$ whoami
samuel.sabogal.pardo
~$ id -G
crypto ai offsec quantum
~$ uptime
20+ years, three continents
~$ cat /etc/motd
build. break. harden. repeat.
tip: click here, this shell is real
~$ ▌
A Fulbright and RSA scholar and Carnegie Mellon University alumnus, Samuel has led high-impact projects across the US, Europe and Latin America, from developing cryptographic systems for banks to designing secure AI infrastructures.
He is Vice President at CYTE, a leading cryptography firm in Latin America. His academic journey includes advanced studies in Artificial Intelligence at Columbia University and deep learning certifications from deeplearning.ai.
With over a decade blending theory and practice, he has taught cryptography and secure software development to professionals, regulators and engineers. Whether building secure systems or breaking them to strengthen resilience, his work lives where logic never sleeps and security never fails.
A one-minute explainer on how zeros and ones become everything on a screen, from a single light bulb to a full 3D world.
> Handmade, pre-AI. Filmed on set, animated and composited by hand, years before a generative model could fake any of it.
