Workflow intensity × tokens
Each point is a project. Both axes use log scales; hover a point for exact values.
10 100 1,000 1 10 100 1,000 10,000
Trading competition · intensity 3,291 · tokens 445 MTok · compute skew 1.87× Trading competition Wakterm.org · intensity 1,891 · tokens 224 MTok · compute skew 1.66× Wakterm.org UFOpedia.fyi · intensity 1,412 · tokens 74.03 MTok · compute skew 0.74× UFOpedia.fyi Personal Project Tracking · intensity 1,410 · tokens 37.09 MTok · compute skew 0.37× Personal Project Tracking FreeTranscribe.org · intensity 1,039 · tokens 55.70 MTok · compute skew 0.76× FreeTranscribe.org Panetone · intensity 1,003 · tokens 41.20 MTok · compute skew 0.58× Panetone Zola · intensity 765 · tokens 38.89 MTok · compute skew 0.72× Zola Autoresearch speedrun · intensity 509 · tokens 6,905 MTok · compute skew 194.23× Autoresearch speedrun gitrep.fyi · intensity 467 · tokens 18.86 MTok · compute skew 0.58× gitrep.fyi Agent Orchestration Process · intensity 354 · tokens 21.35 MTok · compute skew 0.87× Agent Orchestration Process Codex · intensity 318 · tokens 24.43 MTok · compute skew 1.11× Codex Options market-making research · intensity 305 · tokens 12.70 MTok · compute skew 0.60× Options market-making research Sycophant Bench · intensity 264 · tokens 12.61 MTok · compute skew 0.69× Sycophant Bench GPU price tracker · intensity 229 · tokens 7.77 MTok · compute skew 0.49× GPU price tracker MihaiCosma.com · intensity 223 · tokens 6.32 MTok · compute skew 0.41× MihaiCosma.com Refined GitHub · intensity 192 · tokens 5.61 MTok · compute skew 0.42× Refined GitHub GitHub activity tool · intensity 173 · tokens 3.33 MTok · compute skew 0.28× GitHub activity tool Parsync · intensity 153 · tokens 26.31 MTok · compute skew 2.51× Parsync Debate groupchat · intensity 142 · tokens 4.44 MTok · compute skew 0.46× Debate groupchat Popcharts · intensity 108 · tokens 3.57 MTok · compute skew 0.49× Popcharts Infobase · intensity 102 · tokens 16.64 MTok · compute skew 2.39× Infobase Clanker analytics · intensity 98.3 · tokens 7.60 MTok · compute skew 1.14× Clanker analytics Personal scripts · intensity 53.4 · tokens 2.87 MTok · compute skew 0.80× Personal scripts YTubic · intensity 32.1 · tokens 1.79 MTok · compute skew 0.84× YTubic Business opportunity evaluator · intensity 24.8 · tokens 2.59 MTok · compute skew 1.58× Business opportunity evaluator Automated video creation · intensity 11.6 · tokens 1.77 MTok · compute skew 2.33× Automated video creation Cheap food optimizer · intensity 6.5 · tokens 0.33 MTok · compute skew 0.78× Cheap food optimizer Autoresearch · intensity 6.5 · tokens 0.41 MTok · compute skew 0.99× Autoresearch ASR GPU inference optimization · intensity 3.9 · tokens 0.29 MTok · compute skew 1.18× ASR GPU inference optimization
↖ more compute-led
more attention-led ↘
Workflow intensity
Tokens (MTok)
log-log fit · ρ 0.92
Projects
Metrics use the 14d window unless labeled otherwise.
Project
Attention
Agent resources
Automation
Responses? Interactive Codex task completions followed by your next message within one hour.
Latency? Median time from an AI task completion to your next message, excluding delays of one hour or more.
Intensity? Total session time divided by median sub-hour response latency. It rises with longer project sessions and faster supervision cadence.
14d share? Share of total workflow-intensity scores across all measured projects in the 14d reporting window, including private projects omitted from this table. This is a relative engagement index, not a share of clock time. 1d share? Share of total workflow-intensity scores across all measured projects in the 1d comparison window, including private projects omitted from this table. Compare it with the main-window share to see where workflow focus has shifted.
Tokens? Millions of total tokens after discounting cache-read tokens by 90%, matching Clanker's normalized resource measure .
Compute skew? Actual normalized tokens divided by the tokens predicted from workflow intensity by the chart's log-log fit. Above 1× is compute-heavy; below 1× is workflow-heavy.
Sub-agents? Share of AI output generated by sub-agents or headless executions.
AOP? Share of AI output generated with aop .
Trading competition
166
1.5m
3,291
18.1% 0.0%
445
1.87×
55%
55%
Wakterm.org
340
1.8m
1,891
10.4% 0.0%
224
1.66×
0%
0%
UFOpedia.fyi
120
2.5m
1,412
7.8% 9.1%
74.03
0.74×
72%
72%
Personal Project Tracking
290
1.7m
1,410
7.8% 33.0%
37.09
0.37×
0%
0%
FreeTranscribe.org
195
2.4m
1,039
5.7% 8.0%
55.70
0.76×
0%
0%
Panetone
153
3.0m
1,003
5.5% 1.0%
41.20
0.58×
0%
0%
Zola
229
2.3m
765
4.2% 2.0%
38.89
0.72×
0%
0%
Autoresearch speedrun
43
1.7m
509
2.8% 0.0%
6,905
194.23×
100%
100%
gitrep.fyi
60
1.8m
467
2.6% 3.2%
18.86
0.58×
0%
0%
Agent Orchestration Process
85
2.7m
354
1.9% 0.0%
21.35
0.87×
13%
12%
Codex
44
2.5m
318
1.7% 0.0%
24.43
1.11×
0%
0%
Options market-making research
67
2.7m
305
1.7% 0.0%
12.70
0.60×
0%
0%
Sycophant Bench
63
1.8m
264
1.4% 0.0%
12.61
0.69×
16%
16%
GPU price tracker
56
1.3m
229
1.3% 0.0%
7.77
0.49×
0%
0%
MihaiCosma.com
53
2.8m
223
1.2% 0.1%
6.32
0.41×
0%
0%
Refined GitHub
48
1.9m
192
1.1% 0.0%
5.61
0.42×
0%
0%
GitHub activity tool
27
1.3m
173
0.9% 0.0%
3.33
0.28×
0%
0%
Parsync
29
2.1m
153
0.8% 0.0%
26.31
2.51×
0%
0%
Debate groupchat
26
1.5m
142
0.8% 0.0%
4.44
0.46×
0%
0%
Popcharts
12
1.0m
108
0.6% 8.2%
3.57
0.49×
0%
0%
Infobase
27
5.4m
102
0.6% 1.0%
16.64
2.39×
0%
0%
Clanker analytics
43
5.1m
98.3
0.5% 0.0%
7.60
1.14×
0%
0%
Personal scripts
24
5.3m
53.4
0.3% 0.0%
2.87
0.80×
0%
0%
YTubic
11
2.3m
32.1
0.2% 0.0%
1.79
0.84×
0%
0%
Business opportunity evaluator
5
2.9m
24.8
0.1% 0.0%
2.59
1.58×
0%
0%
Automated video creation
5
27s
11.6
0.1% 0.0%
1.77
2.33×
0%
0%
Cheap food optimizer
4
1.4m
6.5
0.0% 0.0%
0.33
0.78×
0%
0%
Autoresearch
4
3.2m
6.5
0.0% 0.0%
0.41
0.99×
0%
0%
ASR GPU inference optimization
1
26s
3.9
0.0% 0.0%
0.29
1.18×
0%
0%