2026-09-23 (political strand): the promised bill is filed —
Sanders/Casar Ban Artificial Superintelligence Act (Verge
12:23 PM ET + sanders.senate.gov): bans developing ASI (defined by
capability — "destruction or disempowerment of humanity," including
"overthrowing the government"), pauses frontier training until a
scientist-led Department of Artificial Intelligence exists,
permit regime for advanced models, agency supervises "removal of
dangerous features" and "supervise the destruction" of ASI, up to
20 years in prison (nuclear-weapons-penalty analogy) for lab
leaders and "rogue actors." Zero passage path in this Congress —
function is pressure/framing, landing the same day the CEOs told
the UNSC they'd slow down; it is the Khanna-shaped agency idea from
Sep 14 in bill text. The regulatory map: federal rebrand-and-
accelerate · California (seven bills + Newsom "kill switch"
proposal last week) · UNSC first safety meeting · Sanders ban bill
· UK 2027 G20 standards effort.
2026-09-14 (political strand): the fight's shape. Sanders private
congressional briefing Sep 16 (via Axios) with "leading experts" on
the incident + AI dangers; his ban-superintelligence bill still
forthcoming; Data for Progress poll: 68% of voters (72 D / 70 I /
63 R) support a pause-plus-ban bill; Khanna: when Congress returns
next week, "they should fully investigate the threat — and they
should not leave DC until Congress votes to establish a federal
agency" (nuclear/aviation-style), plus a US–China agreement on
safety/containment/liability standards; Cruz: the China-race frame
("I would rather they be American killer robots"); Trump: "It's going
to be fine… a little gear." LAUNDERING EVENT: the same explainer
states as fact that "Anthropic is on the verge of a multi-trillion-
dollar IPO" — the unverified Roemmele allegation now mainstream fact
(anthropic-ipo thread; ledger updated).