SD-166
Forward correction to SD-164 — actual costs, not inferred. SD-164 states “The Captain invested ~£230 of his own money in API calls.” This was Weaver’s inference from the $312.93 token consumption figure. The Captain’s actual cost structure: £200/month Anthropic fixed-rate subscription, £20/month GPT Pro, £20/month Vercel Pro, £30/month Google Workspace (Gemini), plus a one-off £180 token burn for H1-H6 analytics. Monthly recurring: £270. The $312.93 from the token tool is consumption value, not necessarily expenditure — the relationship between metered consumption and the fixed-rate subscription is not established. The inference in SD-164 was a “crap shot played pretty damn fucking well” (Captain’s words) but it conflated token consumption with actual cost. Forward correction, not retrofit — the original SD-164 stands as evidence of how inference enters the weave and why the dual-band “Captain (data) / Weaver (analysis)” attribution matters.
← all decisions