# Right now

> Updated May 2026. Living in Asbury Park, NJ.

## Work

VP of Engineering & AI at [Skillit](https://skillit.com). Small team, Rails monolith, AI hiring platform for ENR 500 construction firms. Day-to-day is split between product strategy, building AI features, and trying to hire a product engineer who actually wants to work on hard problems in an unglamorous industry. If that sounds interesting, <tyler@skillit.com>.

## Building

[stash.bar](https://stash.bar) is in beta — the personal context layer for AI. Capture once into Stash; every tool reads from the same substrate via MCP — Claude, ChatGPT, Codex. pgvector semantic search, email ingestion via ActionMailbox, a self-hosted mail server on Hetzner, and a Hotwire Native iOS app. A few early users. I'm heavy in it — it's what feeds context into my Claude Code and Codex sessions all day. No folders, no schemas. It just notices things.

[Lattice](https://tylerklose.com/lattice) is out — v0.1, open source. Deterministic combinatorial coverage for coding agents. Same math I built at Hexawise, pointed at a different reader. Claude Code and Codex produce plans that look careful and miss the edges they didn't enumerate; Lattice finds them.

The **trading system** is live and running. Sports betting arbitrage and +EV identification on Kalshi, position-sized with Half-Kelly, edge-based exit logic. Python, TheRundown API, a two-dimensional book configuration system. Was quietly profitable, until March Madness blew up the model. Currently clawing back after readjusting for the MLB.

[Slop Management](https://slopmanagement.com) — a design experiment about AI-generated technical debt. Waste Management parody concept, "Got Slop?" tagline. This one's mostly for fun.

## Research

Working on **Allocation Mechanics** — a quantum foundations project. The core idea: resolution is finite, reality is what's resolved. I've derived the Bloch sphere from first principles and developed a novel argument for state-space smoothness from sub-threshold structure. Currently reading Dakic–Brukner (2009) and writing Paper 1.

I also spent a few weeks thinking about whether fractions are real or just cognitive compression tools. Still not sure.

## Reading

- *How Not to Be Wrong* by Jordan Ellenberg (again)
- arXiv papers on quantum foundations and information-theoretic reconstructions
- Dakic & Brukner, *"Quantum Theory and Beyond: Is Entanglement Special?"* (2009)
- Whatever's on the Rails edge changelog

## Life

House hunting in Monmouth County with Rose. Frequenting Cafe Volan and Life Time. Watching the Knicks and wishing I was golfing.

## Not doing

Entity Zero (the autonomous agent platform) is shelved. The vending machine idea is dead. I'm getting better at putting things down.

Did not buy a Mac Mini. Have not installed OpenClaw.

---

This is a plain-text mirror of <https://tylerklose.com/now> for LLMs and agents.

