LiveHuman-AI & attentionWestern cycleAnchors pending

AGI-capability milestone (defined as: AI system demonstrating human-expert-equivalent performance across ≥80% of remote-computer-work tasks measured by GDPval or equivalent benchmark) achieved between 2027 and 2031?

As of 6 Oct · 22:12 UTCResolves 31 Dec 20272 movers0 pressure moments
Timeline A · AI RaceReality
55%
Crowd 70%Gap +15

By end of 2027, a frontier AI system passes a standardized benchmark showing it can handle 80% or more of remote computer work tasks at human-expert level—coding, data analysis, technical writing, complex troubleshooting across multiple domains. This system exists in a lab or deployed through an API, measurable by GDPval or an equivalent testing suite. The milestone marks the moment when AI can credibly bid on and execute high-skill remote jobs without human supervision on routine tasks.

Timeline B · Governed AIReality
45%
Crowd 30%Gap −15

Through 2027 and into 2031, AI systems improve rapidly but remain visibly limited: they excel at narrow task clusters (writing boilerplate code, summarizing documents, routine analysis) but stumble on task-switching, unexpected edge cases, and the judgment calls that real remote work demands. No published benchmark shows an AI clearing the 80% bar across diverse expert-level remote work. Capability gains flatten or plateau before reaching that threshold.

crowd 70
The quiet paththe one nobody’s watching

Regulatory freeze or geopolitical AI ban halts benchmark publication and deployment capacity before any milestone can be measured or claimed.

6%
The read
The underlying structure favors Governed AI (55%) over AI Race (45%) — the signal is real but noisy.

Two trajectories through human-ai & attention: Governed AI and AI Race. The Signals below describe what would be visible if each trajectory is landing; the Levers list what could bend it. The third timeline — the one neither Governed AI nor AI Race captures — is where the honest scope-limit sits.

Where reality is heading

A, 80% bandCrowd today
Projected reality support to 31 Dec 2028A rises from 41% to about 52% with an 80% band of 33 to 71. B falls from 53% to about 44%. Quiet path near 4 to 6%. Crowd today: A 55%.100%7550250Crowd today · A 70%A ~52%B ~44%P ~4%NowJan 2027Jan 202831 Dec 2028
Reality crosses toward A in early 2028. Today’s crowd number sits at the top of A’s band, possible but optimistic. View as table

The movers, ranked by leverage

01
AI safety-testing pre-deployment regulatory severityDetermines deployment-safety trade-off
leverage 0
Conditional
02
Chip export-control severityDetermines capability-race pace + China compute-substrate isolation
leverage 0
Conditional

What would change our mind

Signals toward A
  • attention-economy regulated (smartphone age-limits, algorithmic transparency, data-protection)
  • cognitive-dependency addressed via educational integration + digital-literacy
  • AI safety-testing pre-deployment regulatory severity: movement in the direction that supports Governed AI
Signals toward B
  • AI development treated as national-champion race
  • safety-frameworks minimized as constraint-on-innovation, sovereignty-first (US-AI, China-AI, EU-AI, potentially India-AI, Gulf-AI as distinct capability stacks), export-controls hardened, capability-race accelerated
  • AGI-timeline compressed as race dynamics dominate
  • attention-economy less regulated except where national-security implications
  • cognitive-dependency mostly market-mediated with cultural-restorationist opt-out

Evidence