Sitemap
A list of all the posts and pages found on the site. For you robots out there is an XML version available for digesting as well.
Pages
Add invisible noise to a digit and watch a neural network change its mind — adversarial examples, live in your browser. Get details
An interactive look at base rates and Bayes’ rule: test 10,000 people at once and watch why a positive result on an accurate test can still usually be a false alarm. Get details
An interactive look at why a bond’s price drops when interest rates go up. Get details
Compare two texts, files, or PDFs privately in your browser. Nothing is uploaded. Get details
An interactive look at how holding many investments narrows the range of outcomes, cutting the worst cases while keeping most of the upside, and how fees skim off the top. Get details
A stochastic SIR outbreak on a grid: poke it, tune it, vaccinate it, and watch the curve. Get details
Draw any shape and a chain of rotating circles traces it back. Get details
A small, growing collection of interactive things, built as an ongoing exercise in seeing what Claude can do. Get details
A full-screen 2D n-body gravity playground — set a mass, click to drop it, drag to fling it into orbit. Get details
Watch SHA-256’s avalanche effect: change one character, flip half the bits. Get details
Resize an image to a target file size, crop, rotate, strip metadata, and generate a favicon pack — all privately in your browser. Get details
A reinforcement-learning agent lands a rocket. Change gravity, wind, or the pad and watch it cope. Get details
An interactive, full-screen Conway’s Game of Life. Get details
An interactive look at how loan payments split between interest and principal — and how paying extra (especially early) saves you money. Get details
A full-screen sandbox: drop bar magnets, watch their field lines, and let them snap and repel. Get details
Merge, split, reorder, redact, sign, fill, OCR, watermark, compress, and password-protect PDFs privately, in your browser. Get details
Chaos you can grab: two pendulums start 0.001 radians apart and end up nowhere near each other. Get details
Generate and scan QR codes in your browser. The codes are static, so the data lives in the pattern itself, with no redirect service, and a printed code keeps working forever. Get details
A timing-attack demo: recover a hidden token one character at a time, using nothing but how long a comparison takes. Get details
A tiny GPT writes Shakespeare one character at a time in your browser, with every probability and attention head on display. Get details Posts
portfolio
publications
R. Gupta, B. Goodman, N. Patel, R. Hosfelt, S. Sajeev, E. Heim, J. Doshi, K. Lucas, H. Choset, M. Gaston. CVPR Workshops '19. [BibTeX]
Get details Z. Huang, N. G Marchant, K. Lucas, L. Bauer, O. Ohrimenko, B. IP Rubinstein. NeurIPS '23. [BibTeX]
Get details W. Lin, K. Lucas, N. Eyal, L. Bauer, M. K. Reiter, and M. Sharif. NDSS '24. [BibTeX]
Get details talks
K. Lucas, A. Jasen, and L. Bauer. In Neurips 2020 NewInML Workshop. [BibTeX]
Explored surrogate neural network capability to aid adversarial training for hard to differentiate perturbations Get details In DEF CON 33.
Over the past few months, we’ve thrown Claude into the digital trenches of multiple cybersecurity competitions—from defending vulnerable networks at CCDC to cracking challenges in PicoCTF and HackTheBox. In this talk, I’ll take you through our journey deploying an AI assistant against human red teams and live CTF challenges. I’ll show you Claude’s unexpected wins (landing in the top 3% globally in PicoCTF and successfully fending off red team attacks at CCDC) alongside its entertaining fails (devolving into security philosophy when overwhelmed, making up flags for PlaidCTF when stuck). Get details teaching