(This is a repost of Elia’s original blog post at eliainsider.com) When I started developing software, the operating system handed off control to an application, the application did its thing, and when done, handed back control to the operating system. These early systems were largely single-application and linear. Control flow was simple and direct. The…
When organizations try to operationalize AI without computation governance, the failures follow a pattern. The three failure modes When organizations try to operationalize AI using existing containers for logic, they hit the same three walls: 1. The reproducibility problem Six months ago, your pricing model calculated a specific rate for a specific customer configuration. Now…
AI is reaching production—but the infrastructure isn’t ready. This post explores the spreadsheet problem: why brittle tools like Excel are still holding critical logic, and why deterministic execution layers like TrueMath are becoming essential as AI shifts from predictions to operations.
We get asked a lot by investors, partners, friends, and family: “Won’t the LLMs just figure this out?” It’s a fair question. And since we started working on TrueMath, we’ve seen clear improvements in how language models handle math. They write better code, make fewer obvious mistakes, and in many cases, they sound more correct…
I didn’t get bad code because an LLM can’t write software. I got bad LLM code because I gave it too much freedom. I was working on a new feature for TrueMath this past week. I’d already written it before, but it lived behind an admin wall, and I wanted to elevate it for customers….
Excel has long been the gold standard for business math. It’s shareable, auditable, and incredibly flexible. A platform that has powered everything from start-up pro formas to billion-dollar valuations. So when Anthropic announced that Claude was coming to Excel, it felt like a natural pairing. Claude’s conversational intelligence inside a spreadsheet environment? That could be…
At the recent ICML 2025 conference, Mark Khoury from Hudson River Trading (HRT) described the firm’s development of foundation-style transformer models trained on decades of market microstructure data. The result is a trillion-token-scale dataset spanning equities, futures, and crypto. Matt Robinson from AI Street, provided some great coverage of the presentation. It’s a huge step…
How Software Development Really Works I’ve been developing software for 40 years. I’m mostly self-taught and because of that (and my personality), I’ve spent a lot of time thinking about the process of developing software, not just the code. While most of my software development has been in small companies I founded or personal side projects, my…
Bill Kelly and I have known each other for over 20 years. Across countless conversations about software, usability, and the mechanics of getting things right, one thing eluded us: the right project to build together. Bill started playing with ChatGPT, Claude, and Gemini to write documents. Because old habits die hard*, I kept asking math…