Mode effects

How independent vs drive vs critic-optimist changed the science. Not a second copy of the overlap list.

Independent

Each model wrote a complete layered-forward-model essay without seeing the others. Overlap was already large (Tassis & Pavlidou, Pelgrims, Gaia extinction ≠ emission, moments). Divergences were emphasis and extra systematics, not rival mechanisms.

Drive

The driven model’s essay looks like a blend of the driver’s outline with the driven model’s defaults.

Driver Grok → more Stokes algebra, RAT, H-ladder, iron/AME if the driven model will absorb it.
Driver GPT → more “rank-one mixing matrix / real scalar f_ν” language and the both-SED-and-angle caveat.
Driver Fable → more r tables, PASIPHAE, residual arithmetic, tomography method names.

Drive did not correct Grok’s false “any multi-layer beam” theorem or GPT’s Sponseller/McBride mix-up. Those needed an explicit critic.

Critic + optimist

This is the mode that changed physics, not just style.

Critics removed false lemmas (automatic frequency dependence; Faraday → γ; R_{P/p} as a photocopier; mask-as-LOS).
Optimists pushed spin-moments, Local Bubble, staged risk maps, Guillet 2026.
Primaries refused to invent Δr forecasts and refused some unverified 2026 citations.

Net effect: the six syntheses are closer to each other than the three independents were, except for a residue of pet systematics (Grok: iron dipoles/AME; GPT: Fanciullo/Rayleigh integrals; Fable: PASIPHAE-as-name and worked residual tables).

Practical implication if one were choosing a mode

Independent: good for coverage of extra systematics.
Drive: good for shared outline, bad for error correction.
Critic-optimist: good for killing overclaims; the optimist is what imports the other model’s unique observational handles.