v1-vs-v2 — Workgraph

Metadata

Status	done
Assigned	`agent-31`
Agent identity	`3184716484e6f0ea08bb13539daf07686ee79d440505f1fdf2de0357707034c3`
Created	2026-05-01T18:40:11.112354551+00:00
Started	2026-05-01T18:40:53.789159481+00:00
Completed	2026-05-01T18:44:36.838722129+00:00
Tags	`grant,urgent,decision-support`, `eval-scheduled`
Eval score	0.92
└ blocking impact	1.00
└ completeness	0.95
└ constraint fidelity	0.85
└ coordination overhead	0.95
└ correctness	0.90
└ downstream usability	0.90
└ efficiency	0.85
└ intent fidelity	0.82
└ style adherence	0.95

Description

The grant proposal is due TODAY. Erik is asking — as a sanity check at the eleventh hour — what v1 was proposing and whether v2 is genuinely stronger. He needs an honest, fast read, not a justification of an already-made decision.

Erik's question, verbatim: 'what do you think is the strongest'

Project convention: notes go to ~/poietic.life/notes/ (Erik's stated routing). Use ~/poietic.life/notes/v1-vs-v2-executive-comparison-20260501.md.

What to do

Read workgraph_google_application_FINAL.md (v1) and workgraph_extended_outline.md (v1).
Read workgraph_google_application_FINAL_v2.md (v2) and workgraph_extended_outline_v2.md (v2).
Read STATE.md §5 (the v2 pivot decision log) — Erik wants to see the case made for the pivot, not just trust it.
Produce a single output file with these sections, each tight:

Section A: v1 in 150 words

What v1 proposed: scientific spine, demonstration design, key claims, deliverables. No spin.

Section B: v2 in 150 words

Same shape. No spin.

Section C: Honest comparison (300 words max)

For each of these dimensions, state which arm is stronger and why. If they're a wash, say so:

Translational impact (Google.org cares about real-world benefit; rare disease patients vs cancer patients)
Defensibility against expert reviewers (which spine is harder to poke holes in given current literature?)
Authenticity to the founders' track record (Erik = vg/PGGB pangenomics; Luca = CRISPResso/CRISPRme; Vaughn = org design)
Demonstration credibility (PHR methodology comparison vs drug-discovery competition — which produces something a reviewer can actually evaluate?)
Risk of falling apart under scrutiny (overclaim risk, COI risk, data-availability risk)
Fit to Google's stated priorities (Functional Genomics vs AlphaFold-style structural pipeline)

Section D: Recommendation (100 words)

Submit v2 / submit v1 / hybrid / something else. State the call plainly. If v2 is the right call, say what v1 had that v2 should claw back before submission. If v1 is the right call, say so and flag that we'd need an emergency reversal.

Constraints

Style: no em-dashes (CLAUDE.md style rule).
Be honest. Erik incorporated the company, he can take a real assessment. Don't rubber-stamp v2 just because it's the v2 pivot.
Cap each section to its word limit. The whole doc should be under 800 words.

Validation

All four files actually read (v1 application, v1 outline, v2 application, v2 outline)
STATE.md §5 decision log read
Six comparison dimensions each get a verdict
Recommendation is a single clear call, not 'it depends'
Output at ~/poietic.life/notes/v1-vs-v2-executive-comparison-YYYYMMDD.md
Doc is under 800 words total

## Description

The grant proposal is due TODAY. Erik is asking — as a sanity check at the eleventh hour — what v1 was proposing and whether v2 is genuinely stronger. He needs an honest, fast read, not a justification of an already-made decision.

Erik's question, verbatim: 'what do you think is the strongest'

Project convention: notes go to `~/poietic.life/notes/` (Erik's stated routing). Use `~/poietic.life/notes/v1-vs-v2-executive-comparison-20260501.md`.

## What to do

1. Read `workgraph_google_application_FINAL.md` (v1) and `workgraph_extended_outline.md` (v1).
2. Read `workgraph_google_application_FINAL_v2.md` (v2) and `workgraph_extended_outline_v2.md` (v2).
3. Read `STATE.md` §5 (the v2 pivot decision log) — Erik wants to see the case made for the pivot, not just trust it.
4. Produce a single output file with these sections, each tight:

### Section A: v1 in 150 words
What v1 proposed: scientific spine, demonstration design, key claims, deliverables. No spin.

### Section B: v2 in 150 words
Same shape. No spin.

### Section C: Honest comparison (300 words max)
For each of these dimensions, state which arm is stronger and why. If they're a wash, say so:
- **Translational impact** (Google.org cares about real-world benefit; rare disease patients vs cancer patients)
- **Defensibility against expert reviewers** (which spine is harder to poke holes in given current literature?)
- **Authenticity to the founders' track record** (Erik = vg/PGGB pangenomics; Luca = CRISPResso/CRISPRme; Vaughn = org design)
- **Demonstration credibility** (PHR methodology comparison vs drug-discovery competition — which produces something a reviewer can actually evaluate?)
- **Risk of falling apart under scrutiny** (overclaim risk, COI risk, data-availability risk)
- **Fit to Google's stated priorities** (Functional Genomics vs AlphaFold-style structural pipeline)

### Section D: Recommendation (100 words)
Submit v2 / submit v1 / hybrid / something else. State the call plainly. If v2 is the right call, say what v1 had that v2 should claw back before submission. If v1 is the right call, say so and flag that we'd need an emergency reversal.

## Constraints

- Style: no em-dashes (CLAUDE.md style rule).
- Be honest. Erik incorporated the company, he can take a real assessment. Don't rubber-stamp v2 just because it's the v2 pivot.
- Cap each section to its word limit. The whole doc should be under 800 words.

## Validation
- [ ] All four files actually read (v1 application, v1 outline, v2 application, v2 outline)
- [ ] STATE.md §5 decision log read
- [ ] Six comparison dimensions each get a verdict
- [ ] Recommendation is a single clear call, not 'it depends'
- [ ] Output at `~/poietic.life/notes/v1-vs-v2-executive-comparison-YYYYMMDD.md`
- [ ] Doc is under 800 words total

Depends on

done .assign-v1-vs-v2

Required by

done .flip-v1-vs-v2

Log

2026-05-01T18:40:11.111498902+00:00 Task paused
2026-05-01T18:40:31.861081164+00:00 Task published
2026-05-01T18:40:53.589189625+00:00 Lightweight assignment: agent=Default Evaluator (31847164), exec_mode=full, context_scope=task, reason=Default Evaluator with 0.97 score and 20 similar tasks is ideal for comparative grant analysis requiring honest assessment, document synthesis, and clear verdict on competing proposals.
2026-05-01T18:40:53.789163800+00:00 Spawned by coordinator --executor claude --model opus
2026-05-01T18:41:05.150140260+00:00 Starting: reading v1 and v2 application + outline files, plus STATE.md §5
2026-05-01T18:43:00.299467819+00:00 All four files read + STATE.md §5. Drafting comparison now.
2026-05-01T18:44:32.590655081+00:00 Validated: doc at ~/poietic.life/notes/v1-vs-v2-executive-comparison-20260501.md, 705 words (<800), all four files read + STATE.md §5, six dimensions each get verdicts, no em-dashes, single-call recommendation (submit v2). Not committing to poietic.life repo: it's the public landing page repo with unrelated uncommitted index.html changes from another operator; notes are internal review material.
2026-05-01T18:44:36.838735147+00:00 Task pending eval (agent reported done; awaiting `.evaluate-*` to score)
2026-05-01T18:46:09.239124067+00:00 PendingEval → Done (evaluator passed; downstream unblocks)