🔍 Read the full analysis: 50 Signs Of Superintelligence And The Search For AI’s AGI Milestone on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A review published by ThorstenMeyerAI.com on October 26 examines 18 claims tied to researcher Alan D. Thompson’s 50-sign superintelligence tracker. It classifies seven as verified, five as vendor claims, three as based on a single source and three as disputed or contradicted; the tracker itself marks two of 50 indicators achieved, 33 partially achieved and 15 not achieved.
ThorstenMeyerAI.com published a review on October 26 of claims in researcher Alan D. Thompson’s 50-indicator artificial superintelligence tracker, finding that evidence ranges from independently checked mathematical results to vendor-reported achievements and disputed claims. Thompson’s tracker lists two indicators as achieved, 33 as partially achieved and 15 as not achieved, but the review cautions that its status categories and definition of superintelligence are broader than some readers may expect.
The review says it traced 18 headline claims to papers, company posts, researcher statements or official filings. It classifies seven as verified, five as vendor claims, three as resting on a single source, and three as disputed or contradicted. The article describes “verified” claims as checked by an independent expert or supported by an official document; vendor claims rely on a company’s own reporting rather than independent reproduction.
Several of the better-supported examples concern mathematics. The review cites contributions discussed by mathematicians Terence Tao, Don Knuth and Tim Gowers, while retaining caveats about human coaching, prior literature and the kinds of problems involved. It also distinguishes a mathematical result from a broader claim that a system can perform expert work autonomously across fields.
The tracker’s categories need careful reading: Thompson’s August correction note says “partially achieved” includes funded programs, research infrastructure, formal scientific missions and other enabling work. That means a partial status does not necessarily indicate a near-complete capability. The source review also reports that Thompson labels Claude Mythos Preview and GPT-6 Astra “early ASI,” a characterization of his own framework rather than an independently established consensus.
Whose superintelligence?
One word, very different bars
Dr Alan D. Thompson's ASI starts where Google DeepMind's scale puts "Expert AGI". The same model can be "early ASI" on one scale and a strong but narrow AGI candidate on another.
- Median human50th percentileThompson: AGIa machine at the level of an average (median) humanDeepMind: Competent AGIat least the 50th percentile of skilled adults
- Expert90th percentileThompson: ASIexpert-human level in practically any fieldDeepMind: Expert AGIat least the 90th percentile of skilled adults
- Virtuoso99th percentile—DeepMind: Virtuoso AGIat least the 99th percentile of skilled adults
- Beyond every human100%—DeepMind: Superhuman (Level 5)outperforms 100% of humans
- Beyond organisations2026 report—DeepMind: ASImore capable than large organisations of humans
Thompson’s tracker status as of 26 October 2026. DeepMind rows: “Levels of AGI” (2023) and, for the last row, “From AGI to ASI” (June 2026). OpenAI’s Charter: AGI = “highly autonomous systems that outperform humans at most economically valuable work”.
Sources: LifeArchitect.ai ASI checklist (26 Oct 2026), AGI countdown, Morris et al., Levels of AGI (2023), Genewein et al., From AGI to ASI (2026), OpenAI Charter.
Powered by Thorsten Meyer AI
Why the ASI Scorecard Matters
The disagreement is not only about whether particular systems can solve hard problems. It is also about what “superintelligence” means and what evidence should count. A tracker that treats funded research or enabling infrastructure as partial progress can be useful for monitoring activity, but its orange status should not be read as proof that a capability is close to deployment.
Definitions affect public understanding of the technology’s current reach. Thompson defines ASI as a machine performing at an expert-human level in practically any field. The review says this resembles Expert AGI in Google DeepMind’s 2023 framework, rather than the more demanding superhuman or organisation-wide standards used elsewhere. Readers comparing claims need to know which benchmark is being used before interpreting a declaration that a model is “early ASI.”
The distinction also matters for policy and business decisions. Mathematical demonstrations can be significant without showing that an AI system can reliably manage a company, operate in the physical world or replace a research team. The review says large-scale household robotics and a verifiably billion-dollar AI-run company have not appeared, and it reports no visible break in the US unemployment rate tied to AI. Those observations do not settle AI’s future effects, but they mark a gap between selected demonstrations and broad economic or physical autonomy.
Products Worth Considering
Apple iPad Air (4th Generation, 10.9-inch, Wi-Fi, 256GB) – Space Gray
Anker SOLIX C200 DC Portable Power Station, 192Wh, 200W
Bose SoundLink Flex Bluetooth Speaker (2nd Gen), Portable Waterproof Speaker, Black
As an affiliate, we earn on qualifying purchases.
Three Different Standards for Intelligence
Thompson’s public checklist contains 50 indicators, ranging from recursive code improvement to highly speculative milestones. His separate “conservative countdown to AGI,” according to the source, stood at 99%; its remaining threshold is a robot completing multi-step fine assembly, such as putting together flat-pack furniture, with average-human reliability. These are the author’s chosen measures, not a field-wide consensus or a validated probability estimate.
The review compares that approach with other published definitions. Google DeepMind’s 2023 “Levels of AGI” framework distinguishes performance bands from competent through expert and virtuoso to superhuman. Its June 2026 report describes superintelligence as a system more capable than large organisations of humans. OpenAI’s Charter, as quoted in the source, defines AGI around highly autonomous systems outperforming humans at most economically valuable work. The frameworks therefore set different bars for capability and autonomy.
Some entries in the tracker illustrate why source checking matters. The review says two exoplanet candidates found with Claude Code were placed under an indicator about colonising a planet beyond Earth. That categorisation does not mean colonisation has occurred; it shows how an observed research result can be mapped to an indicator whose wording describes a much more distant outcome. The article’s central approach is to check each claim against its underlying source rather than treating a montage or status colour as proof.
“documented material progress toward an indicator, including funded programs, research infrastructure, formal scientific missions, and enabling work”
— Alan D. Thompson, in an August correction note to his tracker
What the Checklist Cannot Establish
The review’s classifications do not amount to a comprehensive audit of all 50 indicators: it examined 18 selected claims. The supplied source does not give the full claim-by-claim table, the underlying evidence for every status, or enough detail to independently assess each classification. Some results are attributed to company reports, and the source itself says they have not been independently reproduced.
It is also unclear how consistently the tracker applies its categories across very different milestones. A mathematical proof, a research program and a robot performing reliable assembly are not directly comparable measures of general intelligence. The source identifies no consensus threshold for AGI or ASI, and it does not establish that the models Thompson calls “early ASI” meet other organisations’ definitions.
The article reports Anthropic’s internal evaluation and its estimate of research productivity, but those figures are company-specific and are not a general measure of AI capability. The evidence described also does not establish whether current systems can autonomously perform expert work across practically any field, or when physical and economic milestones might be reached.
Evidence to Watch in Future Updates
The next useful test is whether claims tied to company systems can be reproduced independently and whether researchers disclose enough detail to evaluate the human contribution, tools used and problem difficulty. For mathematical claims, the review points to public proofs and expert commentary as checks; for commercial performance, independently verifiable records would carry more weight than a company announcement alone.
Readers can also track changes to Thompson’s checklist and its separate AGI countdown, while comparing them with definitions published by research organisations. Any later status update should be read alongside the tracker’s explanation of what “partial” means. The source does not report a scheduled independent review or a specific next milestone, so the timing and criteria for future reassessment remain unspecified.
Key Questions
How many indicators does Thompson’s tracker mark as achieved?
As reported in the source on October 26, 2026, the tracker lists two achieved, 33 partially achieved and 15 not achieved, for 50 indicators in total.
Does “partially achieved” mean a milestone is nearly complete?
Not necessarily. Thompson’s correction note says the category can include funded programs, infrastructure and enabling work, so it should not automatically be interpreted as near-completion.
What did the review find about the 18 claims it checked?
It classified seven as verified, five as vendor claims, three as based on a single source and three as disputed or contradicted. The review examined selected claims, not every tracker entry.
Why do definitions of ASI differ?
Thompson’s definition focuses on expert-human performance across fields. Other frameworks place more weight on superhuman capability or autonomy across economically valuable work, so the same model may receive different labels.
Has the source established that current models are superintelligent?
No. The source reports Thompson’s “early ASI” description as his view and presents mixed evidence for selected claims. It does not establish that current models meet a shared, independently validated standard for superintelligence.
Source: ThorstenMeyerAI.com
