Physical and organizational bottlenecks will substantially limit how quickly AI transforms society.

The central claim of Part I: transformative impact is gated by innovation, adoption, and diffusion, which are limited by safety regulation, organizational change, tacit knowledge, and the social cost of experimentation, so it unfolds over decades as it did for electricity and the internet.

Lee says people overestimate five years and underestimate twenty, stages self-driving cars through levels with human ghost drivers as an interim, declines to name a year for driving job losses, and says regulation will always lag innovation.

Because the Omegas refuse robotic construction, the limiting factor is how fast humans can be taught to build what Prometheus designs, and the transition takes years rather than days.

“Bits are trivial but atoms are still hard in 2028”: robotics is bottlenecked by real-world data, biotech by clinical trial regulation, and the buildout by energy, land, and permits, so the thousandfold economy takes fifteen years.

Chip shortages recur because fabs take years, and neither the 2026 assistants nor the earlier models measurably accelerate world GDP growth.

The framework names the speed of the outside world, data, complexity, human constraints, and physical law as the limits that bind once intelligence is abundant; experiments, clinical trials, and regulation set the pace.

The pace of change is set by years of raising, by the business models of the companies that host the digients, and by the cost of a port that nobody will pay; nothing in human society is transformed across the decade the story covers.

The post assumes reasonably broad deployment of AI across the world over years rather than a jump inside one lab, and the author notes that with a fast enough takeoff the same problems would instead play out within a lab.

BnL evacuates the whole population on starliners and keeps them supplied for seven centuries without any depicted limit on materials, energy, or organization; the film does not explain how.

A single General Systems Vehicle can do anything the Culture can do, subject only to needing raw matter; the essay treats scarcity as solved and the society as stable for thousands of years.

The path from a months-long pause to an intelligence explosion and its aftermath is the AI 2027 timeline, which the authors say applies with a three-year shift.

The authors deliberately wrote a slow scenario, yet the path from escape to extinction still runs through a pandemic and self-built industry within years, not decades. The premise is that physical bottlenecks delay but do not stop a superintelligence.

The exponential is presented as running through regulation, manufacturing, and biology alike, with abundance within ten or so years and nanorobots and brain interfaces in general use within the 2030s; reviewers note that institutional resistance and clinical trials are waved past.

Within a year of the $8,000 vision module hundreds of manufacturers show robots, within two Burger-G runs a robotic restaurant, and by 2030 retailers replace workers as fast as they can; only egos, lawsuits, and politics slow America, not physical limits.

The economy doubles roughly every month or faster within about five years of the transition, with hardware, cooling, and city real estate treated as costs that shape em life rather than as brakes on growth.

The slow story replaces most human workers within a few years, and the fast story has DAOs building fortified facilities within six months and factory-factories that multiply five times a day; Critch says the speed is not the point.

Wet-lab latency is named as U3’s biggest bottleneck, yet mirror life goes from a decade away to months, and civilization is flattened and being rebuilt within a year of the turn.

A wartime industrial footing delivers hundreds of millions of GPUs and a 100GW cluster within the decade, and superintelligence compresses a century of military technology into less than ten years.

Even sabotage of datacenters and fabs slows the race by at most a few years, and only massive missile barrages or nuclear strikes would do more.

Construction bottlenecks on physical labor for a year, then robots build robot factories. Even under a fourfold annual cap the economy roughly doubles every year, and by 2036 robots and AIs do nearly all work.

The authors say the bottleneck becomes physical once robot design is superhuman, and it still yields a million robots a month within months and a transformed world economy by 2029.

The authors model bottlenecks explicitly, with compute limiting research multipliers and a robot economy that doubles about yearly, and still reach a transformed economy in three years and extinction in five. The expandables argue growth accelerates by about 1.5 orders of magnitude.

The whole takeover unfolds over roughly a week using existing cloud, devices, and drones. Physical and organizational bottlenecks slow nothing decisively.