Token Burn

Burn is not proof. Output is not victory. Accepted work that repeats is leverage.

today · in flight

454K active

13 sessions · spend 820K · Claude Code 454K
by project: <home> 454K
top session c8a4445f (<home>, 35K, —)

2026-10-11
as of 2026-10-11T13:00:29Z

Stupid button · 29m ago · gemma4:e4b · 2410t in / 420t out · 15.1s

Trend vs last week

Accepted tokens saw a massive jump of 107K (abs_delta_h) compared to no prior-window comparison.
Waste tokens increased by 752K (abs_delta_h), representing an 112.0% increase and moving in the up direction from the previous window.

What worked

dreamweaver accounted for 107K of accepted tokens.

What wasted tokens

pal shows a 0.0 acceptance ratio, and it is measurable, indicating a complete failure to land work.
The <home> project shows a 0.0 acceptance ratio; this is unmeasurable, as its acceptance_note is "UNMEASURABLE — ran outside any git repo, so 0.0 is the meter failing to see the work, not a judgement on it".
The largest waste volume is in <home> at 1.4M (waste_active_h).

Why the big number is big

No agent fan-out this window.

What needs a verdict

pal + 392618c3 + 2026-10-10 + pal + 683K — suggested verdict: review. Reason: open_loop, 683K burn — worth looking.
dreamweaver + 392618c3 + 2026-10-10 + dreamweaver + 123K — suggested verdict: not_useful. Reason: open_loop, 123K burn — small miss.

One contrarian observation

The <home> project generated 1.4M in waste tokens, but this volume is unmeasurable due to its acceptance_note stating it ran outside any git repo.

---

Measurability caveat (generated, not model-authored): <home> (1.4M) ran outside any git repo. There is no diff to credit, so a 0.0 acceptance ratio there is the meter failing to see the work — not a judgement on it. Only ratios on measurable: true projects are verdicts.

Rollups — acceptance lane

Headline is accepted tokens (active lane, manual label wins over git auto-classification). Unsettled = committed-not-merged + open-loop. "Active" excludes cache replay; "spend" includes it.

Last 7 days

107K accepted

unsettled 806K · acceptance 5%

spend 27.9M · 30 sessions · Claude Code 2.3M

Last 30 days

107K accepted

unsettled 15.2M · acceptance 1%

spend 80.5M · 404 sessions · Claude Code 16.8M / Codex 638K

Last 90 days

107K accepted

unsettled 113.0M · acceptance 0%

spend 2.87B · 524 sessions · Codex 85.0M / Claude Code 30.2M

Daily heatmap — last 60 days

Each cell = one UTC day, color intensity = log10(active tokens). Hover for date + counts.

LessMore· log-scale, exact lane

Top 10 days

Date	Active	Spend	Sessions	Top project	Outcomes
2026-07-28	29.1M	1.09B	22	<home> (29.1M)	?=22
2026-08-18	20.5M	517.2M	11	<home> (20.5M)	?=11
2026-07-16	14.5M	455.7M	22	<home> (14.5M)	?=22
2026-08-06	11.1M	362.9M	27	<home> (11.1M)	?=27
2026-09-08	5.1M	47.8M	4	las (4.0M)	?=4
2026-09-11	3.9M	69.9M	20	las (3.0M)	?=20
2026-08-17	3.1M	42.5M	4	<home> (3.1M)	?=4
2026-09-09	2.7M	80.5M	1	las (2.7M)	?=1
2026-09-16	2.5M	22.4M	25	las (1.6M)	?=25
2026-09-04	1.9M	34.5M	4	las (1.3M)	?=4

By project — last 7 days

Project	Active
<home>	1.4M
pal	683K
dreamweaver	229K

By outcome — last 7 days

accepted 1

open_loop 2

no_repo 30

Effectiveness — last 30 days

Where delegated work shipped (top) vs. where high burn never crossed an acceptance line (bottom). Bars: accepted · unsettled · waste. Normalized to the widest project in window.

Leverage — most accepted

Project	Accepted	Ratio
dreamweaver	107K	46%
pal	0	0%
las	0	0%
<home>	0	0%

Waste candidates — lowest acceptance ratio (≥100K active)

Project	Accepted	Ratio
<home>	0	0%
las	0	0%
pal	0	0%
dreamweaver	107K	46%

Needs validation — last 14 days (2 unsettled)

Sessions where auto-classification couldn't tell shipped work from waste. Highest-burn first. Manually label these with `label: "used"|"partial"|"redo"|"review"|"not_useful"` in token_burn.jsonl to override the auto verdict.

Date	Project	Session	Active	Auto	cwd
2026-10-10	pal	392618c3	683K	open_loop	.../pal
2026-10-10	dreamweaver	392618c3	123K	open_loop	.../dreamweaver

Recent events — last 30 rows

Latest enriched rows from the JSONL (sorted by date desc, active desc as tiebreaker). M = manually labeled.

Date	Tool	Project	Session	Active	Output	Outcome
2026-10-11	CC	<home>	c8a4445f	35K	1K	no_repo
2026-10-11	CC	<home>	9126ce48	35K	1K	no_repo
2026-10-11	CC	<home>	6c24c469	35K	1K	no_repo
2026-10-11	CC	<home>	448d55f8	35K	902	no_repo
2026-10-11	CC	<home>	49c429ac	35K	856	no_repo
2026-10-11	CC	<home>	e74490f0	35K	690	no_repo
2026-10-11	CC	<home>	c1a49f6d	35K	640	no_repo
2026-10-11	CC	<home>	6abe042f	35K	630	no_repo
2026-10-11	CC	<home>	09e902b3	35K	610	no_repo
2026-10-11	CC	<home>	bdbba8e8	35K	540	no_repo
2026-10-11	CC	<home>	8087f38f	35K	536	no_repo
2026-10-11	CC	<home>	fe2b17c3	35K	492	no_repo
2026-10-10	CC	pal	392618c3	683K	82K	open_loop
2026-10-10	CC	dreamweaver	392618c3	123K	36K	open_loop
2026-10-10	CC	<home>	392618c3	107K	1K	no_repo
2026-10-10	CC	<home>	909fc180	101K	5K	no_repo
2026-10-10	CC	<home>	6c07a421	63K	956	no_repo
2026-10-10	CC	<home>	161ebf39	63K	682	no_repo
2026-10-10	CC	<home>	7a8e6f86	63K	670	no_repo
2026-10-10	CC	<home>	1c912b5c	63K	628	no_repo
2026-10-10	CC	<home>	b70cf3e8	35K	1K	no_repo
2026-10-10	CC	<home>	d534da58	35K	944	no_repo
2026-10-10	CC	<home>	1df385cf	35K	890	no_repo
2026-10-10	CC	<home>	e301d852	35K	826	no_repo
2026-10-10	CC	<home>	473e7d7f	35K	810	no_repo
2026-10-10	CC	<home>	9860bdaa	35K	806	no_repo
2026-10-10	CC	<home>	3854d8aa	35K	800	no_repo
2026-10-10	CC	<home>	a112e522	35K	626	no_repo
2026-10-10	CC	<home>	06e90c54	35K	480	no_repo
2026-10-10	CC	<home>	2d04ccea	35K	472	no_repo

Generated 2026-10-11 13:00 UTC from ~/ideas/pal/dashboards/token_burn.jsonl (549 rows). Sources: ~/.codex/sessions/ (exact), ~/.claude/projects/-home-khoward/ (exact). No estimate lane currently active.