Phase Lock

gist

Saturnin Pugnet's Phase Lock manifesto argues AI capability is outrunning human bandwidth through keyboards and screens, creating a closing Phase Lock Window to set the terms of human-machine integration. He prefers alignment-by-integration over temporary alignment-by-control, and bets non-invasive brain-computer interfaces will scale like modern AI—via hardware, neural datasets, and foundation models—so neural data becomes a frontier training modality and a higher-bandwidth interface that preserves human agency.

ideas

• The bandwidth problem is the mismatch. AI agency rises while humans still steer with screens and clicks, so direction and governance do not scale with capability.
• Integration outlasts control. Alignment-by-control can hold briefly; alignment-by-integration couples human values and intents to the machine directly before the Phase Lock Window shuts.
• BCI is a triple lever for alignment. It offers the brain as an architectural blueprint, supplies training signal for human intent, and widens mind-machine bandwidth so people can keep pace.
• Brain decoding inherits AI scaling laws. Better hardware, larger neural datasets, and foundation models can extract structure from noisy unlabeled brain signals—the ChatGPT moment of the brain.
• Consumer non-invasive BCI is the scale path. Devices without surgery can reach millions, finance further R&D through demand and lab value for neural data, and set a flywheel in motion.

quotes