Speed ↓ · market economic
Compute governance
The compute supply chain is a stable chokepoint and state coordination on licensing, export controls, and reporting thresholds can govern capability indirectly.
Mechanism
Control flops via chip export controls, reporting thresholds, fab licensing, and compute accounting.
What this name has meant
vintage driftThe name is stable; the content has shifted. A reader acting on the label without asking which vintage is being meant risks arguing with a position nobody currently holds.
Export controls on frontier chips.
Broadened to reporting requirements, cloud access controls, and domestic audits, tools the original frame did not consider.
If it succeeds: what binds next
The compute chokepoint binds. Frontier capability is governed by whoever controls the flops accounting regime, a new concentration emerges inside the governance.
A strategy that produces a worse next problem than the one it solved has not done durable work.
Falsification signal
The capability-per-flop curve steepens faster than chip export controls tighten.
A strategy held without a falsification signal is not strategy; it is affiliation. Continued support after this signal lands is identity, not bet. See the identity diagnostic.
Self-undermining threshold
overshoot riskWhen capability-per-flop scales faster than the flops threshold tightens.
Aggressive compute governance pushes development toward distillation, algorithmic efficiency, and architectures that circumvent thresholds. The governance leverage decays as the metric becomes leaky. The Goodhart case.
Every strategy has a stable region where it reinforces itself and an unstable region where pursuit defeats it. The threshold between them is usually narrower than advocates acknowledge.
Historical analogue
Nuclear · Zangger Committee / Nuclear Suppliers GroupEvery strategy inherits a plausible ceiling from its precedent. The analogue conditions the realistic reach.
Produced
Input control regime that bounded proliferation inputs.
Did not produce
Persistent defection, smuggling, and domestic production routes.
People on the record
12Profiled figures appear first, with their tier in small caps. Each face links to the person and their full quote record. Tag: compute-governance.
Catherine Aiken
CSET researcher; China AI talent and capability
Cliff Young
Google; TPU principal engineer
David Patterson
UC Berkeley emeritus; Google AI hardware; 2017 Turing Award
Elsa Kania
CNAS adjunct senior fellow; China AI specialist
Jordan Schneider
ChinaTalk podcast host; Rhodium Group
Lennart Heim
Compute governance researcher at RAND
Pat Gelsinger
Former Intel CEO; chip-supply geopolitics
Paul Scharre
CNAS executive VP; 'Army of None', 'Four Battlegrounds' author
Saif M. Khan
Former NSC AI technology director
Stephen Witt
Author of 'The Thinking Machine' (NVIDIA history)
Tim Fist
Institute for Progress AI policy researcher
Zachary Arnold
Georgetown CSET; analytics lead
Coordinates
Conflicts, grouped by mechanism
3Frame opposition
incompatible premisesThe strategies accept different premises about what AI is or what the binding problem is. They conflict not on lever choice but on the frame that makes lever choice sensible.
Lever opposition
same lever, opposite pullThe pair's primary lever is the same; they pull it in opposite directions. A portfolio containing both is internally incoherent on that lever.
Complements, grouped by mechanism
5Same phase, different layer
same stage, distinct leversBoth are active in the same phase of the transition but act on different layers (model vs institution vs culture). They cover different failure modes inside the same window.
Same-lever twins
4Both use the same lever in the same direction. Usually redundant inside a portfolio: each dollar or effort unit only buys one lever pull, even if two strategies are named.
Axis position
Source note: Compute governance as strategy.md