# n78 — the total assault: trying to kill everything at once
# (2026-07-09, the closing adversarial pass; verdict per vector)

## Vector 1 — the Lakatos charge (the meta-kill)
The accusation: this is not a theory but a family of toys plus a
dictionary rewritten three times (gravity->EM, cost->distance->energy,
passive->active). Each refutation was absorbed by reinterpretation;
the CORE (the vacuum audits a budget) has never itself been at risk —
the signature of a degenerating research programme.
VERDICT: the charge extracts a real concession but does not kill.
Counter-evidence: the reinterpretations paid tolls (n69 passed five
gates including a NEW measurement; the active vacuum passed stability,
trio and Caves); twenty-five kills were accepted as final. But the
concession is structural and now RECORDED AT THE TOP OF THE HOUSE:
the core is falsifiable through exactly ONE gate — the knee. A flat
null across the band kills the audit, the budget, and every reading
at once. THE KNEE IS THE FRAMEWORK'S EXISTENCE GATE; without it, the
Lakatos charge would be correct. Everything is bet on one experiment,
and the ledger must say so in one sentence.

## Vector 2 — F1's "zero freedom" is FALSE: a kill, landed
Count the unknowns: a knee sweep yields two numbers (the knee
frequency and the amplitude coefficient). The medium has g^2, kappa,
Delta PLUS the units mapping (lambda, substrate scale) — more unknowns
than measurements. "The derived G will match Cavendish with zero
freedom" was an overclaim: the extraction is UNDERDETERMINED by at
least one ratio. What survives: a one-unknown consistency RELATION
between knee, amplitude coefficient and G — still a real constraint
(a knee INCOMPATIBLE with Cavendish for every value of the free ratio
still kills), but not a parameter-free prediction. F1 IS AMENDED in
the forecast ledger, same day, at full prominence. The program's
oldest failure mode (semantic overclaim) caught once more — this time
by its own assault, before any referee.

## Vector 3 — the charge-sign catastrophe that becomes a doctrine
The EM reading ties charge SIGN to lattice-site parity: move a charge
one site and it flips sign?! Fatal for continuum electromagnetism —
EXCEPT this is precisely the Kogut-Susskind staggered-fermion
construction: sublattice structure encoding spinor components is HOW
lattice field theory represents Dirac matter. The kill converts into
a requirement: the EM reading commits the framework to matter as
staggered-fermion-like dressings (assumption A-EM, recorded). Not a
wound — a specification, and a testable one for the substrate program.

## Vector 4 — the cosmological bill (the 120-orders trap)
The pump maintains inversion across all space: naively, catastrophic
vacuum energy — the very problem the framework promised to address.
DEFLECTED WITH INTEREST: the Gaussian-order bill was MEASURED at
exactly zero (n74-A3: NESS carries vacuum noise precisely). Where QFT
overshoots by 120 orders at leading order, this vacuum's leading-order
idle cost is 0.000; Lambda must come from corrections — small because
the leading term vanishes, which is the right STARTING point for the
observed smallness. The trap the framework was built to escape, it
escapes at the order it has been tested.

## Vector 5 — F8 is a direction, not a bet
Conceded: without the pump's microphysics, "G drifts with dark energy"
has no coefficient and cannot lose. Re-graded in the ledger from bet
to registered direction.

## The assault's balance
One kill landed (F1 amended), two concessions extracted (the
existence-gate sentence; F8 regraded), one vector converted into a
specification (staggered matter), one deflected with interest (the
zero bill). The framework ends the assault smaller, sharper, and
honest about its single point of failure: EVERYTHING RIDES ON THE
KNEE. That is not a weakness discovered — it is the wager finally
stated at its true size.