Skip to content

Instantly share code, notes, and snippets.

View XyraSinclair's full-sized avatar

Xyra Sinclair XyraSinclair

View GitHub Profile
@XyraSinclair
XyraSinclair / Heesch_numbers.md
Last active August 17, 2026 20:16
The Heesch number world record (6, Bašić 2020) machine-decided exactly, with independently verifiable certificates

heesch-sat patches required at record scale

Upstream: github.com/isohedral/heesch-sat @ master, cloned 2026-07-29. Solver: cryptominisat 5.14.7. Build on macOS arm64, clang.

Patches applied to heesch-sat/src:

  1. Makefile: include/lib paths /usr/local + /opt/local -> /opt/homebrew.
  2. Makefile: -std=c++17 -> -std=c++20 (tileio.h uses templated-lambda NTTPs).
  3. isohedral.cpp: add #include and #include (libc++ only
@XyraSinclair
XyraSinclair / README.md
Created May 10, 2026 02:28
Model consulting as a small, reliable prompt tactic

Model consulting as a small, reliable prompt tactic

A compact public pattern for asking another strong model to make a plan, patch, or explanation simpler and more robust.

The point is not that one named model is always better. The point is that an independent model, given the right brief, often notices unnecessary complexity, missing invariants, brittle assumptions, and unclear handoffs. Used well, this is one of the cheapest ways to improve agent work.

The core move

After you have already decided that an outside consultation is worth the latency/cost, ask for a concrete review against a narrow outcome:

@XyraSinclair
XyraSinclair / hermes-auto-steward-gist.md
Last active April 16, 2026 23:10
Hermes auto-steward followthrough: canonical installation, behavior, and verification guide

Hermes auto-steward followthrough

Canonical installation, behavior, and verification guide for making Hermes continue obvious, safe work instead of stopping too early.

This document defines the feature contract, reference configuration, implementation shape, and verification procedure.

Purpose

Auto-steward adds bounded self-followthrough to Hermes so a chat does not stop after the first adequate answer when the user's intent is already clear.

@XyraSinclair
XyraSinclair / mwi-implications.tex
Last active February 1, 2026 14:12
Quantum Diversification under Many-Worlds: When and Why Timing Your Randomness Matter
% paper.tex
\documentclass[11pt]{article}
\usepackage[margin=1in]{geometry}
\usepackage{microtype}
\usepackage{mathtools,amssymb,amsthm}
\usepackage{graphicx}
\usepackage{booktabs}
\usepackage[hidelinks]{hyperref}
\usepackage[numbers,sort&compress]{natbib}
@XyraSinclair
XyraSinclair / dx-wishlist-late2025.md
Created December 6, 2025 21:03
AI building tools wish list
  • /search-history claude code command that helps claude rapidly search over history of claude code conversations, as well as a full screen nice visual way of human doing this. current terminal approaches like cass are not nearly ergonomic enough for me.
  • /fork agent chat.