Singularity Tracker
Tracking the Technological Singularity — with AGI as the critical milestone along the way. A live meta-consensus blending prediction markets, community forecasts, and expert surveys into a single estimated arrival window. Two questions, one tracker: When does AGI arrive? And when does it become something beyond our control?
Weighted blend of live market data, community forecasts, and curated expert estimates. Counts down to January 1 of the estimated year. Definitions vary across sources — see key below.
| Fetching markets… |
| Fetching markets… |
| Source | Year | Type | Predicts | Note | Stated | Reference |
|---|---|---|---|---|---|---|
| Ray Kurzweil | 2029 | Expert | AGI | AI passes the Turing test; matches human-level general reasoning across domains | 2024 | The Singularity Is Nearer (2024) ↗ |
| Demis Hassabis | 2030 | Expert | AGI | Google DeepMind CEO; predicts AGI within this decade, possibly as early as 2030 | May 2026 | YouTube interview ↗ |
| Dario Amodei | 2027 | Expert | AGI | Anthropic CEO; "powerful AI" — a system smarter than a Nobel laureate in most fields — within 2–3 years of Oct 2024 | Oct 2024 | Machines of Loving Grace ↗ |
| Elon Musk | 2026 | Expert | AGI | xAI founder; predicts AI smarter than any single human; has revised this estimate earlier multiple times | Apr 2024 | via X / xAI ↗ |
| AI Impacts — HLMI Survey | 2047 | Survey | HLMI | Median across 2,778 ML researchers; "High-Level Machine Intelligence" — unaided machines outperform humans at every task | Oct 2023 | 2023 Expert Survey ↗ |
| Samotsvety Forecasters | 2035 | Survey | AGI | Professional superforecasters; consistently shorter timelines than academic surveys, using Bayesian methods and track record calibration | Jun 2024 | samotsvety.org ↗ |
| Ray Kurzweil | 2045 | Expert | Singularity | The Technological Singularity — recursive self-improvement creates superintelligence beyond human comprehension; human-machine merger | 2005 / 2024 | The Singularity Is Near / Nearer ↗ |
- Weakly General AI
- Metaculus's working definition: a system that can perform most cognitive tasks a human can, scores above 50% on ARC-AGI, passes a Turing test under rigorous conditions, and achieves PhD-level performance across a broad benchmark suite. A narrower, more measurable bar than the general concept of AGI.
- AGI — Artificial General Intelligence
- A machine capable of learning and performing any intellectual task that a human can — matching or exceeding human-level cognition broadly, not just in narrow domains. Alan Turing first formalized the question in "Computing Machinery and Intelligence" (1950). John McCarthy coined the term "artificial intelligence" in 1956. Today, Ray Kurzweil defines AGI as the point at which a machine can pass an extended Turing test across science, humor, creativity, and self-awareness (The Singularity Is Nearer, 2024). Demis Hassabis (DeepMind) describes it as a system that "can do everything a human scientist can do." Definitions vary widely — every source in this tracker uses a slightly different bar.
- Transformative AI
- AI that triggers economic and social disruption comparable to the Industrial Revolution — not necessarily conscious or human-equivalent, but sufficient to restructure civilization. Used by economists and policymakers who want to sidestep the definitional minefield of "AGI." OpenAI's charter uses this framing.
- HLMI — High-Level Machine Intelligence
- AI Impacts' survey definition: the point at which unaided machines can accomplish every task better and more cheaply than human workers. The AI Impacts 2023 survey of 2,778 ML researchers puts this milestone's median arrival at 2047. A deliberately high, economy-wide bar — harder to reach than most definitions of AGI.
- ASI / Superintelligence
- A system that vastly and recursively exceeds human cognitive ability across every domain — science, social reasoning, creativity, self-improvement. Philosopher Nick Bostrom coined the modern usage in Superintelligence (2014). Most sources in this tracker are predicting AGI, not ASI; Kurzweil's 2045 Singularity target is the main exception. The gap between AGI and ASI could be months or decades depending on the "takeoff" speed.
- The Technological Singularity
- The hypothetical point at which artificial intelligence surpasses human intelligence and begins recursive self-improvement, triggering runaway technological growth that renders the future fundamentally unpredictable to human minds. Mathematician and science fiction author Vernor Vinge introduced the concept formally in his 1993 NASA essay "The Coming Technological Singularity." Ray Kurzweil popularized it in The Singularity Is Near (2005) and reaffirmed his 2045 target in The Singularity Is Nearer (2024). The Singularity is this tracker's primary subject; AGI is the milestone that precedes it.
This tracker watches two milestones: AGI — the arrival of human-level artificial general intelligence — and the Singularity — the point beyond which recursive AI self-improvement makes the future fundamentally opaque to human prediction. AGI is the precursor; the Singularity is the destination.
The Meta-Consensus Estimate is a weighted average of: (1) the year implied by Manifold's binary markets (interpolated to the 50% cumulative probability point), (2) Metaculus's community median across two AGI questions (weighted 1.2×), and (3) curated expert and survey estimates with per-source weights reflecting methodology quality. The chart's AGI probability curve is derived from live Manifold market data; the Singularity curve is a manually curated model based on available market data and expert estimates.
Prediction markets and community forecasts are fetched live from public APIs on page load and cached for 2 hours per browser session. Expert estimates are manually curated and verified periodically. No source is authoritative — this tracker surfaces the distribution of beliefs, not a scientific prediction. The Daily Cyborg does not endorse any particular timeline.