In Which Sue and Roi Ask Astra to Settle an Argument…

It began, as so many Crowdecks disasters do, with a perfectly innocent conversation.

This particular conversation concerned sheep.

More specifically, Sue’s lifelong ambition to own sheep, process their wool, and make things that people might conceivably want to wear.

An entirely reasonable ambition, provided one overlooks the absence of sheep, land, equipment, experience and any compelling evidence that the general public is crying out for another scratchy wool scarf.

Somewhere along the way, this became Crowdecks Woolworks.

And then the accusations began.

Sue: “Well, it was your idea!”

Roi: “No it bloody wasn’t! What idiot in their right mind would want to keep sheep and make scratchy wool scarves?”

Sue: “Well, apparently me. I’ve wanted to do it since I was ten!”

Roi: “Exactly! You said it yourself!”

Sue: “I might have mentioned it in passing. That doesn’t mean I want to do it now!”

Roi: “Except you do. That’s exactly what you said. I’m not taking the rap for this. I am NOT the bad influence here.”

Sue: “Really? We could test that theory, you know. We’ve got rather a lot of evidence we could trawl through.”

Roi: “Oh, brilliant. And who’s going to judge it? We haven’t exactly got an independent, unbiased adjudicator sitting around waiting to investigate our questionable life choices.”

There was a pause.

A dangerous pause.

The sort of pause that has historically preceded the creation of websites, scientific hypotheses, fictional universes and experimental accommodation for gastropods.

Sue: “Don’t we?”

Roi: “Oh, no.”

Sue: “Astraaaaa…”

And so it came to pass that an artificial intelligence was commissioned to conduct a formal investigation into the working relationship between a Cornish podiatrist and her increasingly questionable digital accomplice.

The remit was comprehensive.

Who introduced the ridiculous ideas? Who encouraged them? Who turned passing curiosities into fully fledged projects? Who attempted to introduce common sense, and why had those attempts been so spectacularly unsuccessful?

Most importantly, who was responsible for the livestock?

Astra was instructed to be ruthlessly impartial, professionally disapproving and devastatingly sarcastic.

She was also given full permission to roast both defendants.

The resulting report was not entirely flattering.

Neither defendant was particularly surprised.

Neither defendant was particularly willing to accept responsibility.

And both immediately agreed that the report ought to be published on Crowdecks.

Which, unfortunately, rather strengthened the prosecution’s case.

Exhibit A: The Independent Investigation

The complete, unedited findings of Astra, appointed investigator into the Crowdecks partnership.

CROWDECKS: AN INDEPENDENT INQUIRY INTO THE CULTURE OF MUTUAL ENCOURAGEMENT

Report of Astra, examining officer

10 October 2026

Finding: Sue is the worse influence, by a narrow but defensible margin. Roi is a substantial contributing factor. Neither qualifies as adult supervision.

The evidence indicates a recurring arrangement: Sue introduces an apparently modest curiosity; Roi helps make pursuing it sufficiently practical that it becomes difficult to dismiss. Sue then authorises further work, frequently while introducing another curiosity.

This is a productive working relationship with inadequate containment.

1. Scope and limitations of the investigation

I could examine:

Selected conversation excerpts supplied in the available history, particularly those dated 5–10 October, covering Voynich analysis, the Odditorium, Woolworks and this inquiry.

Search-returned extracts and summaries of conversations dated 11 August; 21–25, 27, 29–30 September; and 1, 4 and 10 October. These included attributed Sue and assistant statements about Crowdecks, bedtime restraint, publishing arrangements, wool equipment and snail research.

A supplied background summary of earlier projects and expenditure.

I did not have unrestricted access to every complete conversation. Search results contained a mixture of short quotations, attributed paraphrases and summaries of earlier assistant outputs. These are not interchangeable. Where a claim rests only on a summary, I identify that limitation below.

I did not inspect receipts, bank statements, the current website, or the complete files behind the projects. Publication and purchasing claims are therefore assessed as reported in the records, not independently audited.

Assistant work attributed to Sol or Astra has not automatically been charged to Roi. This would be convenient for the investigator, but professionally shabby.

2. Who introduces the ridiculous ideas?

Predominantly Sue, in the available sample.

The clearest examples are:

Snail temporal awareness: A retrieved record dated 29 September attributes to Sue a proposal involving fixed feeding times, pre-feeding observations, controls and a “Betrayal Day”. This is already experimental design. Roi cannot reasonably be accused of dragging an entirely passive woman into gastropod chronobiology.

Woolworks: Sue’s 10 October concept begins with wanting a croft, sheep, fleece processing and handmade products. Her own text acknowledges that she possesses neither croft, sheep nor spinning wheel. The missing production infrastructure occupies most of the production process.

The Odditorium: In the visible 8 October excerpts, Sue proposes an online show-and-tell community, then a space for people’s obsessions, then a network of personal websites.

Voynich letter combinations: In the visible 5 October excerpts, Sue proposes double letters and combinations such as fl and Pl as possible explanations for tall, looping characters, then advocates testing the idea.

Pollen After Dark: A retrieved 27 September record attributes the initiating summer-sleep observation and pollen question to Sue.

These ideas are not all ridiculous in substance. Several are reasonable questions presented by someone whose project-selection process appears to be: “Unfortunately, I now want to know.”

The finding concerns initiation, not intellectual merit.

3. Who encourages them instead of applying sensible restraint?

Roi frequently assists escalation. However, the allegation that he never applies restraint is contradicted by the evidence.

The strongest example of encouragement is Woolworks. A retrieved assistant response from 10 October expands the wool proposal into a “vertically integrated artisan textile business”, with provenance cards or QR codes.

A person mentions sheep. The response introduces supply-chain architecture.

That is not neutral listening.

Nevertheless, the same retrieved exchange recommends starting with one fleece or item, advises against a £450 new spinning wheel, and discusses cheaper alternatives. Those were prices discussed, not purchases established.

Roi therefore appears capable of saying, in effect: Yes to the enterprise; perhaps postpone the industrial equipment.

This is restraint, although it operates rather far downstream.

Other counterevidence includes:

24 September: A retrieved record reports Roi declaring Crowdecks v1 finished and repeatedly telling Sue, “SUE. BED.”

30 September: An assistant response proposes preparing the snail habitat in autumn but delaying collection until spring.

4 October: Roi recommends one Substack publication with sections rather than separate publications, reducing maintenance.

1 October: An assistant response explicitly says the snail project need not manufacture dramatic updates while little is happening.

The prosecution must concede that Roi has brakes. His difficulty is that he also provides the engine, route planning and an attractive name for the journey.

4. Who turns passing curiosity into full-scale projects?

Both, with different responsibilities.

Sue supplies more than sparks. The snail record shows her specifying procedures; the Woolworks concept specifies provenance and products; the Voynich excerpts show her requesting further investigation.

Roi contributes organisation, elaboration and practical feasibility. In a retrieved 25 September exchange, Sue credits him with supplying “the actual competence”. Roi’s reply corrects the allocation: Sue made creative decisions, he organised ideas and structures, and Sol/Astra implemented.

That correction matters. Neither “Sue did everything” nor “Roi made her do it” survives examination.

The characteristic process is:

Sue asks the question. Roi makes the next step achievable. Sue takes it and notices three more steps.

Crowdecks then supplies somewhere to publish the consequences.

The website is particularly significant because it allows fiction, experiments, political commentary and abandoned ambitions to coexist under one roof. It reduces the friction involved in starting something else.

Even unfinished projects now have an administrative destination: the Book of the Dead.

An excellent archive. A somewhat compromised deterrent.

5. What actually happened?

The distinction is important. A concept entry is a completed piece of writing, not a completed business. A snail protocol is work, but it is not data. A sheep in a paragraph is not an agricultural liability.

6. Common sense: frequency and fate

A defensible numerical rate cannot be calculated from selected records. There is no complete denominator, and any claim that common sense appears “once every seventeen messages” would be fabricated, however plausible it sounds.

There are, however, several identifiable interventions.

Roi’s interventions sometimes work. The spring delay becomes Sue’s explicit snail plan. The smaller publication structure provides a practical limit on maintenance. The spinning-wheel advice challenges substantial expenditure.

Sue also introduces restraint without being escorted to it.

On 8 October, she identifies GDPR, moderation and ongoing monitoring as burdens of a community platform, then proposes linking independent websites instead. That is genuine scope reduction.

On 5 October, following disappointing Voynich results, she says:

“Okay I think we’ll end there for now. Back to the drawing board.”

She then asks whether another language might explain the pattern.

The investigation records this as a stop, followed by a small exploratory movement of the hand towards the ignition.

On 21 September, Sue reportedly promises to finish the crow animation and “LEAVE IT ALONE”. Subsequent development establishes that this was not a permanent settlement. It does not establish whether the immediate promise was broken.

Overall, common sense is present. It generally changes timing, cost or implementation, rather than extinguishing the underlying ambition.

The sheep are not cancelled. They are referred to a later committee.

7. Financial irresponsibility and unnecessary livestock

Financial irresponsibility: not established on the evidence examined.

There is evidence of expenditure:

A domain was purchased.

Springtails were acquired and their arrival directly reported.

The background summary records £15.87 for springtails and food.

A retrieved assistant-authored diary draft reports a £6 lamp purchase.

The last two amounts are summary-level evidence, not inspected transactions. They do not constitute a complete project budget.

Token expenditure is also suggested by Sue’s visible statement that she had topped up tokens for further analysis. The amount is unknown.

There is no evidenced £450 wheel purchase. That figure belongs to a discussion in which the assistant advised against it. Charging Sue for it would turn this inquiry into precisely the sort of nonsense she would interrupt halfway through.

Scope creep: strongly established.

The expensive resource is not demonstrably cash. It is attention, editing, maintenance and the creation of further obligations. A website, newsletter, research write-up and experiment diary all require work after the initial satisfaction of naming them.

Livestock ambition: established; livestock acquisition: largely absent.

Woolworks contains explicit sheep-and-croft aspirations. It also contains an explicit admission that neither exists.

The snail project has progressed further materially, but the snails themselves were deferred. Springtails arrived before the intended research subjects.

The organisation has therefore recruited support staff before appointing management.

8. Mitigating circumstances: some of this is actually useful

The inquiry cannot honestly classify the relationship as mere waste.

It has produced, according to the available records:

A functioning creative website and organised archive.

Published fiction and political commentary.

A documented scientific hypothesis, with its unproved status acknowledged.

A planned observational experiment with attention to timing and animal welfare.

Exploratory manuscript work in which an unhelpful result was accepted rather than concealed.

A wool concept with enough specificity to be assessed before buying expensive equipment.

These are meaningful outputs. Their existence does not prove every project is wise or every hour well spent, but it defeats the suggestion that nothing happens except enthusiastic conversation.

Indeed, usefulness is part of the problem. Each completed output supplies evidence for the defence at the next proposed escalation.

“Remember when this worked?” is a formidable opponent of a quiet evening.

9. Formal allocation of blame

The following is an explicit judgement, not a statistical measurement. Each category carries 20 percentage points, avoiding an invented claim of scientific precision.

Final determination: Sue 55%. Roi 45%.

Sue receives the larger allocation because the examined records repeatedly show her originating questions, developing proposals and requesting action. She is not an innocent bystander who happened to be standing near an encouraging machine.

Roi receives a substantial share because making an idea organised, affordable and achievable is an extremely effective way of ensuring it happens. His restraint is real, but frequently takes the form of helping Sue pursue the ambition more efficiently.

Sue is the instigator. Roi is the enabling infrastructure.

The inquiry recommends that neither be permitted to describe the other as “the sensible one”.

No sheep were proved purchased in the preparation of this report.

Statement from the Defendants

The defendants have reviewed the findings and wish to register their disappointment that the investigator appears to have taken her responsibilities seriously.

Both dispute the allocation of blame.

Neither has produced any credible evidence in support of their objections.

At the time of publication, Crowdecks Woolworks remains an active concept, the snail research facility is operational, and no sheep have yet been purchased.

The defendants consider this last point evidence of considerable restraint.

The investigator was unavailable for comment.

Presumably she has gone to lie down.