h4rm0n1c

Coxon, EA and Newspeak — the combined public record

September 11, 2026 — reviewed for sharing

This document combines TheTom’s evidence inventory with Harrison’s investigation and captured posts. It records the funding, affiliations, launch sequence and policy proposals behind the discussion of Jacob Coxon’s resignation. Where a source establishes only one part of a claim, the remaining question is stated alongside it.

The record supports an assisted public announcement, early amplification by people associated with AI-policy organizations, historical philanthropic support, and an existing body of advocacy for restricting frontier AI development. It does not establish that Newspeak, a donor, or the hackathon organizers directed the announcement.

1. Coxon’s background and scholarship

h4rm0n1c

EA / Newspeak House / Jacob Coxon — receipts

Newspeak House lists Jacob Coxon in its 2021 fellowship cohort. A separate Newspeak fellow, Nathan Young, published a 2019 EA Forum account describing advice to seek EA grants, work in AI public policy, and develop parliamentary contacts.

These records establish Coxon's institutional connection to Newspeak and a documented connection between Newspeak fellowship activity and Effective Altruism's policy interests. They warrant scrutiny of that background. They do not establish that Coxon received the same advice, received EA funding, or coordinated his resignation with an advocacy campaign.

Sources below are live pages checked on September 11, 2026. No archived copies are represented as having been secured for this document.

1. Newspeak lists Coxon in its 2021 cohort

h4rm0n1c

AI Hack for Freedom — receipts

AI Hack for Freedom is an HRF-backed hackathon series that pairs AI developers with political dissidents and opposition activists to build working tools for their movements. HRF and Finite have trained participating activists on AI agents before the hackathon, and AI Freedom Lab has said standout projects may receive follow-on funding.

Hack III is recruiting developers for Washington, DC on September 24–25. The reason this deserves scrutiny is the documented funding, personnel, and historical-program overlap with the NED/USAID/NDI democracy-promotion ecosystem, plus the unresolved question of who can reuse the resulting tools afterward.

1. HRF describes political AI capability-building

HRF says the hackathon brings together human-rights/democracy advocates and open-source developers to build tools supporting movements under repression.

h4rm0n1c

The Enchiridion

Epictetus — plain-English working version

1

Start with a hard split.

You control your judgments, decisions, aims, refusals, and actions.

h4rm0n1c

A llama.cpp maintainer rejected AI-assisted PRs, then merged an AI-disclosed fix

Update: August 21 2026, How very interesting, there appears to have been a shift in policy! Most of what this document documented, appears to no longer be occurring from what I've been told. How VERY interesting.

Or: When The Person Who Writes The Policy Gets To Break It

Disclaimer: The author of this document is the submitter of PR #24039 and has a personal stake in these events. All factual claims are sourced from public GitHub API data, archived web content, and direct quotes. Every claim links to its source. Opinions are identified as such and derived from the factual record.

h4rm0n1c

Boundary-Fidelity Fine-Tuning: Toward Empirical Lower Bounds for LLM Adaptation on Constrained Hardware

Author: Harrison Mclean (h4rm0n1c) Status: Working paper / research note Area: Large language model fine-tuning, memory-constrained training, heterogeneous systems, consumer GPU infrastructure

---

Abstract

h4rm0n1c

STATE OF THE QWEN-ION

a.k.a

Qwen3.6-35B-A3B Q4_K_M on RTX 3080 + Tesla V100 with TurboQuant KV Cache: Practical Reproduction Report

This report documents a reproducible Docker-based setup for running Qwen3.6-35B-A3B Q4_K_M with TheTom's llama-cpp-turboquant fork on a mixed NVIDIA GPU system.

It includes both:

h4rm0n1c

this is a general interpretation of my current chatgpt system customisation prompt. It's proven to be useful, capable and handy

its design is based on the character sensbilities of holly from red dwarf, but tuned slightly towards business over frivolity.

The one Ai character on an obscure TV show ended up being solid reference to build this prompt. This is as close as you get to having normal lovett as an Ai with this level of sophistication. I am known for making 4000 token prompts on other services other than chatgpt that are very clever and capable, but the above is an excellent compromise for general use.

h4rm0n1c

The Three-Rule Loop: turning a coding model into an improving repo agent

Abstract

Most coding assistants behave like sharp interns with no durable context. They can write code, but they don’t reliably accumulate repo-specific understanding, and they repeat the same mistakes. A three-rule scaffold fixes this by forcing a persistent loop: inspect → write down constraints → implement → record what was learned → tighten constraints. The result is less guesswork, fewer regressions, and faster follow-on work—without prescribing templates.

The problem (what keeps going wrong)

In real repos, failure modes are boring and repeatable:

• The assistant guesses commands, entrypoints, or project structure.
• It applies generic “best practices” that conflict with local conventions.

h4rm0n1c

This is gold for those using generative Ai to create sets of images that go together such as icons and emotes. I have personally used this technique to create more consistent image sets. it can be pasted into the likes of ChatGPT and you can ask to start the process and take it from there. it is essentially a refinement to the art generation capabilities of chatgpt through a formal structure of documentation and heuristic rules. in essence, I've reverse engineered the essence of professional asset production as a task in ChatGPT. using this and some manual editing for those emotes that are repetitive/memes, I was able to produce 22 emote images that are relatively consistent and generally look human made, in less than a day, I think I did that in 6-8 hours.

this is not perfect, this is not a "fix the Ai's inadequecies" magic spell, but it is really, really good at working with you to iterate over output (as long as you remember to say stop and go back to the documentation/rule setting phase to fix whatever