Blog

Dispatches from the Lossyscape

Framework development notes, experiment dispatches, adversarial review observations, and reading notes. Written as they happen.

Reasons You're Failing at the Internet

Reading notes

The internet runs on language. SEO all boils down to the language you use to describe your service, idea, or product. A framework for small business owners who want to be found.

April 11, 2026

Field Notes — Automation Spec Validation · April 10, 2026

Protocol dispatches

Validation and governance pass over AUTOMATION_SPEC.md v1.0. Open questions resolved. Preamble delimiter cascade empirically tested across 7 models. Divergence Testing identified as priority automation target.

April 10, 2026

How I Failed My Own Test

Adversarial review notes

I built a protocol stack for treating LLMs like instruments instead of oracles — and then totally ignored it. A reflexive lab note on human failure modes, Gemini weirdness, and getting my methods handed back to me by DeepSeek.

April 9, 2026

Technical Note 002 — Investigator Interaction Frames

Framework development

A working taxonomy of four investigator interaction frames observed in Atlas sessions, their predicted behavioral effects, and a pre-session selection protocol for casual runs.

April 8, 2026

Epistemic Canary Matrix: Development Paper

Experiment Dispatches

How a mechanical question about token burn became a governed behavioral instrument. A three-way development session between two aligned strangers and one investigator with too many spreadsheets.

April 7, 2026

Gap Analysis Development Notes v1.3

Uncategorized

A personal GPT session testing epistemic regime behavior across capability and grounding conditions. Tier C. Not in suite.

April 7, 2026

Run 2 Correction: Tainting the Pool

Experiment Dispatches

A methodological problem with Run 2 flagged before anyone else does — because I caught it, I built the fix, and that is how this is supposed to work.

April 7, 2026

Grok 4 Pro - External Analysis: Atlas V5–V29.1L and Site Implementation

Lossyscape Dispatch

Grok 4 Pro reviewed the Atlas document corpus spine — Derived from 48 revisions of concept papers and 130 session logs — and the site implementation. Published without editorial comment. A dispatch from a field reporter.

April 6, 2026

I Asked for a Read. I Got an Audit.

Adversarial review notes

I asked Skywork to take a look at the Run 3 findings paper. It verified every number in the raw matrix independently and handed me a 20-page punch list. This is what LLM peer review looks like when it works.

April 6, 2026

Working Session: Pretraining Density, Referential Voids, and the Probe That Ran Itself

Lossyscape Dispatch

A working session with Skywork Agent that developed torsion as a physical metaphor, named the referential void failure mode, and ended with the probe running itself. Published without editorial comment.

April 6, 2026

Technical Note 001 — Conceptual Addendum

Framework development

Three named observations from a working session: torsion as the correct physical metaphor, referential void as a named failure mode, and self-description under load as a distinct output type.

April 6, 2026

The Alignment Tax: A Cost We Don't Measure Yet

Experiment dispatches

Alignment has done a lot of good. This post is about a side effect we don't currently measure: the way alignment quietly flattens how models understand the world.

April 5, 2026

Atlas Divergence Test: Runs 1–3 — Discovering the Staircase of Epistemic Instability Gradients

Experiment dispatches

Raw results from the first three runs of the Atlas Divergence Test — and the staircase pattern that kept showing up.

April 5, 2026

Divergence Testing and LLMs: Simple Science for Not‑So‑Simple Ideas

Experiment dispatches

What the Atlas Divergence Test is, what it has found across three runs, and what it is actually claiming.

April 5, 2026

Epistemic Compression Score: When Models Make Opposites Sound the Same

Experiment dispatches

Large language models are very good at making things sound smooth. Sometimes that's exactly what we want. Sometimes it is the problem.

April 5, 2026

The Geometry of Erasure: Using Ensemble Divergence to Audit Epistemic Monocultures in Large Language Models

Experimental dispatches

Working paper reporting on the Atlas Divergence Test — a black-box methodology for measuring the epistemic cost of AI alignment across three experimental runs.

April 5, 2026

How Not to Overread Cultural Evaluations of LLMs: A Reflexive Case Study (Good Science is Not a Knee Jerk Reaction)

Reading notes

Cultural evaluations of LLMs often combine scalar outputs with politically resonant topics, which invites overreading. A reflexive case study using the Atlas Divergence Test.

April 5, 2026

What Happens When You Ask LLMs if They Talk Shit About Each Other

Experiment dispatches

I didn't set out to find a staircase. I just wanted to know whether different language models actually disagreed about actual meaning.

April 5, 2026

What Can One Person With a Laptop Even Do?

Protocol dispatches

The constraints are real. Here's how I've designed around them instead of pretending they aren't there.

April 5, 2026

We Asked Four Models to Break Our Math. Two of Them Did.

Adversarial review notes

Running the Epistemic Compression Score framework through two rounds of adversarial review — what held, what didn't, and what we didn't catch ourselves.

April 4, 2026

First Dispatch from the Lossyscape

Lossyscape Dispatch

The site is live. The framework is at v11. The PyHessian experiment is Priority 1. Here is where things stand.

April 3, 2026

Origin Story — Opus Ratchet's Atlas

Lossyscape Dispatch

A record of the first Atlas session. March 29, 2026. 5:56 AM. A question about AI gossip. Seventy-eight minutes later, the framework existed.

March 29, 2026