Scaling mathematical discovery of new optimization methods with a robust pipeline for tuning and testing LLM-generated optimization code on dynamically generated test functions.
Category Archives: Machine Learning
What I learned vibecoding my first iOS app
For the last few months, I’ve worked on Gigglebot, an iOS app for instantly putting AI images in your text messages. Starting with no experience in Swift or iOS app development, it’s been a huge learning experience, and I wanted to share some of the insights I’ve learned along the way. » Read More
Job destruction, Job creation, and the return of craftspeople in the era of AI
In the last post I spoke about how there will be a brief shiny period of AI solo-unicorns. This post examines what happens beyond that era of high AI valuations, as we enter a world where more companies with tiny AI-empowered teams are are competing with each other. » Read More
1P1B: The brief shiny era of the solo unicorn-chaser
As AI agents have gained traction over the past year, there’s a buzz in the Bay Area: Someone, somewhere, sometime soon, might have the first one-person company with a $1 billion valuation: 1-person-1-billion, 1P1B, or the solo-unicorn.
AI-first companies have already had some impressive exits (Base44 sold for $80M in June 2025), » Read More
Coding AI LLMs are 3-D Printers for Software
At a hackathon in 2023, I worked alongside a rising college Junior majoring in Computer Science who was worried that he should switch majors because AI would eliminate software development jobs. At the time, I brushed off his concerns, as I expected most of the usefulness to be in written language, » Read More
CityScaper – Rezoning tools
I took part in the SF10X hackathon in early August, and ended up winning second place and $1500 for a project which allowed for scalable visualizations of what the city will look like under the proposed rezoning.
The project was mentioned in the SF Standard’s coverage of the SF10x hackathon, » Read More
Privacy-Preserving Data Services: Beyond GDPR
For years, data-driven tech companies have used customers as their raw material: Facebook, Google, Amazon, and others track our moves across the web in order to sell that information and target advertisements. For a long time, users found this an acceptable tradeoff- but in the wake of the 2016 election and amidst the changes brought about by the General Data Protection Regulation (GDPR), » Read More
Trustless Federated Machine Learning
Machine learning has exploded as companies find ways to draw actionable insights from the data which consumers feed them. However, the efficacy of artificial intelligence algorithms is dependent on the size of the data pool- giving big companies like Facebook and Google a formidable advantage over small scrappy startups. » Read More
New Paper: Cybersecurity in Distributed Optimization
Conventional centralized optimization algorithms have challenge solving big optimization problems- at some scale, you simply can’t fit the problem on a single computer, let alone manage all of the variables. To solve this, researchers use decentralized optimization techniques to break the problem into a set of subproblems which can be rapidly solved on distributed computers or smart devices- but this exposes the optimization algorithm to cybersecurity threats from hacking these consumer-level devices. » Read More
The #DataOpen Journey
I’ve never been so happy sleeping so little. For five days this November, I taxed my mind and my laptop crunching through education policy data on the way to submitting an entry in the 2017 Data Open Championships, organized by Correlation One and Citadel. » Read More