ASI Timeline Tracker: Every Major Expert Prediction, With Dates

Direct-Answer Introduction

The ASI Timeline Tracker is the living record for Artificial Superintelligence predictions: every major public position on when Artificial General Intelligence — and then Artificial Superintelligence — could arrive, kept on one page, attributed to a named person or survey, and updated whenever a source moves. Predictions are opinions, never facts, and this page treats them that way.

Quick Takeaways

  • This tracker records who predicted what — it does not make predictions of its own, and no entry here is a promise or a fact.
  • Lab-leader positions summarised in our research span from powerful AI within 1–2 years to a 5–20 year superintelligence window; the researcher-survey median sits around 2042 for human-level AI alone.
  • Where our research summary records no specific statement date, the "Date said" column shows "—" rather than an invented date.
  • Every entry carries a source. Entries are only upgraded with a date when a dated primary source is verified.
  • The change log at the bottom records every addition, correction, and update, with dates.

Table of Contents

  • How to Read This Tracker
  • Our Methodology
  • The Prediction Table
  • What the Spread of Predictions Tells Us
  • How Updates Work
  • Change Log
  • Frequently Asked Questions
  • Internal Links
  • Sources

How to Read This Tracker

Read each row as a sentence with three parts: a named source, a claim, and the evidence status of that claim. "Sam Altman, OpenAI — internal AGI possible by end-2026" means: this is the position attributed to Altman in the research summarised in our Full Site Plan; it is a claim about possibility, made by the person running one of the labs in the race; and it has not happened or been independently verified. If you read every row that way — source, framing, status — the table cannot mislead you, even when individual predictions turn out to be wrong.

Three framings recur and should not be merged. Some sources predict AGI (human-level generality). Some, like Anthropic's Dario Amodei with "powerful AI", deliberately use their own term for very capable systems, which is related but not identical. And some, like Safe Superintelligence's Ilya Sutskever, address superintelligence itself. A "sooner" date for AGI and a "later" date for ASI are not necessarily in conflict — they are often answers to different questions.

Our Methodology

What Gets an Entry

An entry requires a named person, lab, or survey and a checkable position: a timeframe, a probability, or a clear conditional claim (for example, that a small number of further breakthroughs are needed). Anonymous claims, unsourced social-media summaries, and predictions attributed to "experts" in general do not qualify.

Attribution Before Everything

Every entry is published in the form "X predicted or said Y", following this site's Editorial Policy. We never convert a source's framing into our own wording where the conversion would strengthen the claim: "possible by end-2026" stays "possible", a roughly 50% probability stays a probability, and a range stays a range.

Dates — and Honest Gaps

Where a dated primary source (a lab post, essay, speech, paper, or reputable report with a publication date) is verified, the "Date said" column carries that date and the Sources section links it. The six seed entries below come from the research summary in our Full Site Plan, which records the positions but no specific statement dates; their "Date said" is therefore shown as "—". We regard an honest gap as strictly better than a plausible-looking invented date, and filling these gaps with verified primary sources is this page's first maintenance task.

One Canonical Page

This tracker is a static page, not a series of posts, by design. When a tracked source moves a timeline, the table is edited in place and the change is logged below — one canonical URL accumulating a record, instead of ten thin follow-up articles each capturing a fragment.

The Prediction Table

Expert / LabPredictionDate saidSource
OpenAI — Sam AltmanInternal AGI possible by end-2026 (claim as summarised in plan research)—ASI News World research summary, Oct 2026
Anthropic — Dario Amodei"Powerful AI" possible in 1–2 years (Amodei's own framing)—ASI News World research summary, Oct 2026
Google DeepMind — Demis Hassabis~50% chance of AGI within the decade; "one or two more breakthroughs" still needed—ASI News World research summary, Oct 2026
Safe Superintelligence (SSI) — Ilya SutskeverSuperintelligence in 5–20 years—ASI News World research summary, Oct 2026
xAI — Elon MuskClaims AI smarter than any human (claims, to be read against independent forecasts)—ASI News World research summary, Oct 2026
AI Impacts researcher surveyMedian forecast: 50% chance of human-level AI by ~2042—ASI News World research summary, Oct 2026

What the Spread of Predictions Tells Us

The most important information in the table is not any single row; it is the distance between the rows. The people leading frontier labs describe powerful or general systems in windows of years. The aggregated judgement of the wider research field, as summarised from the AI Impacts survey, places even the human-level step around 2042 at median. Both cannot be straightforwardly right about the same thing — and noticing that they may be answering different questions (different definitions, different terms like "powerful AI", different incentives) is the beginning of reading this field well. This tracker exists so that, over time, the record itself arbitrates: dated entries, kept honestly, show whose framing survives contact with events.

How Updates Work

  • Same-day rule: when a news post on this site reports a tracked source moving a timeline, this page is updated the same day.
  • Log, don't overwrite silently: superseded positions are not deleted without trace; the change log notes what changed and when.
  • Corrections: errors in a name, figure, or framing are corrected in the table and recorded in the change log, per our Editorial Policy.
  • Reader submissions: timeline updates with a primary-source link can be sent via the Contact page; unverifiable claims are not added.

Change Log

  • Tracker created — October 2026: seeded with the six positions summarised in the ASI News World Full Site Plan research (OpenAI/Altman; Anthropic/Amodei; Google DeepMind/Hassabis; SSI/Sutskever; xAI/Musk; AI Impacts survey). "Date said" shown as "—" for all six: the plan records no statement dates, and none have been invented. First maintenance task: attach dated primary sources to each entry.

Frequently Asked Questions

Is Artificial Superintelligence predicted to arrive on a specific date?

No. No entry on this page is a date-certain prediction, and this site makes none of its own. Entries are attributed positions — claims, ranges, and probabilities — recorded so readers can compare them and check them over time.

Why does the "Date said" column show "—" for every entry?

Because the research summary these seed entries come from records the positions but not the specific dates on which they were stated. Rather than invent dates, we show the gap and fill it only when a dated primary source is verified for that entry.

Why do the lab leaders and the researcher survey disagree so much?

They may be using different definitions (AGI vs "powerful AI" vs superintelligence), weighing different evidence, or facing different incentives — lab leaders speak for organisations racing to build these systems, while surveys aggregate a wider, more cautious field. The tracker does not resolve the disagreement; it documents it, dated and attributed, so the record can.

How is this tracker different from prediction lists on other sites?

Three ways: one canonical page that is edited rather than replaced; an attribution rule that preserves each source's exact framing ("possible", probability, range); and a public change log, including honest gaps where a date is not yet verified.

Can I suggest an entry or a correction?

Yes — via the Contact page, with a link to the primary source (the lab post, paper, speech, or reputable report). Suggestions without a checkable source cannot be added.

Internal Links

  • What Is Artificial Superintelligence (ASI)? — the definitions behind every term in this table
  • When Will AGI Arrive? What Altman, Amodei, Hassabis, Musk, LeCun and the Researcher Surveys Actually Say
  • OpenAI and Superintelligence: What OpenAI Has Actually Claimed (and Not Claimed)
  • Anthropic's "Powerful AI": Amodei's Timeline in His Own Framing
  • Google DeepMind and AGI: Hassabis's "One or Two More Breakthroughs"
  • Safe Superintelligence (SSI): Ilya Sutskever's 5–20 Year Bet, Explained
  • xAI and Superintelligence: Musk's Claims vs Independent Forecasts
  • Editorial Policy — the attribution and corrections rules this tracker follows

Sources

  • ASI News World, Full Site Plan (20 October 2026), §2: the six seeded positions as summarised in plan research (OpenAI/Altman; Anthropic/Amodei; Google DeepMind/Hassabis; SSI/Sutskever; xAI/Musk; AI Impacts survey median ~2042 for human-level AI). No statement dates are recorded in the plan; none are asserted here.
  • Dated primary sources will be added to this section as each entry is verified.

Author: ASI News World


 

0 Response to "ASI Timeline Tracker: Every Major Expert Prediction, With Dates"

Post a Comment