🔍 Read the full analysis: A Checklist Of 50 Superintelligence Signs: How Close Is AI To AGI? on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
AI consultant Alan D. Thompson’s October 26 checklist marks 2 of 50 indicators achieved, 33 partially achieved and 15 not achieved. A review of 18 headline claims found a mix of independently checked results, company-reported claims and disputed or single-source assertions; the checklist is one researcher’s framework, not a standard measure of AGI.
Australian AI consultant Alan D. Thompson marks 2 of 50 indicators on his public artificial-superintelligence checklist as achieved, with 33 partially achieved and 15 not achieved, according to a status grid last updated October 26. A review by ThorstenMeyerAI.com of 18 prominent claims in Thompson’s 2026 video found seven verified, five based on company claims, three resting on a single source and three disputed or contradicted — a mixed record that does not establish that current AI systems meet a shared definition of AGI or superintelligence.
Thompson’s checklist ranges from mathematical discoveries and software improvement to space exploration and human-level robotics. He defines AGI as average-human performance and ASI, or artificial superintelligence, as expert-human performance across practically any field. His tracker labels the current models “early ASI,” but that is his assessment, not a finding from an independent measurement standard.
The source review found its strongest support in claims that can be checked against research or formal records. It identified mathematical contributions involving Terence Tao, Don Knuth and Tim Gowers, while also retaining important qualifications: some work involved coaching, human-written proofs, prior literature or particular problem formulations. Other claims, including results announced by AI companies, were classified as vendor claims rather than independently reproduced findings.
The tracker’s “partially achieved” category is broad. Thompson says it includes documented progress, funded programmes, research infrastructure and enabling work, rather than only systems close to completing an indicator. The review notes, for example, that two exoplanet candidates identified with Claude Code were placed under an indicator about colonizing another planet. That status does not mean colonization has occurred.
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 Checklist’s Labels Matter
The checklist offers readers a way to track concrete developments, but its totals can be misleading without the definitions behind them. With 33 indicators marked partial, the label spans very different kinds of progress, from research activity to capabilities that may remain far from practical use. It should not be read as a countdown showing that superintelligence is nearly here.
Definitions also change the conclusion. Google DeepMind’s 2023 “Levels of AGI” framework distinguishes competent, expert, virtuoso and superhuman performance. Thompson’s definition of ASI — expert-level performance across fields — is closer to DeepMind’s “Expert AGI” category than to its later description of superintelligence. DeepMind’s June 2026 report describes the latter as exceeding the cognitive capability of large human organisations. OpenAI’s charter, by contrast, emphasizes highly autonomous systems that outperform people at most economically valuable work.
For readers, that means a headline claim that a model is “early ASI” cannot be evaluated without asking which definition, task and evidence standard are being used. Mathematical results may show notable performance in a defined domain; they do not by themselves establish broad, autonomous competence across work and everyday settings.
Products Worth Considering
Apple iPad Air 11-inch (M2), 512GB, Wi-Fi, Space Gray
Amazon Kindle Scribe Colorsoft 64GB (1st Generation)
Apple iPad Air 13-inch (M4), 128GB, Wi-Fi 7, Space Gray
As an affiliate, we earn on qualifying purchases.
How the 50 Indicators Are Scored
Thompson, an Australian AI consultant who publishes The Memo newsletter, maintains a public list of 50 possible signs of ASI. His October video presents developments from January through October 2026. The October 26 tracker records two indicators as achieved, 33 as partial and 15 as not achieved. Those figures describe the tracker’s classifications, not a consensus among researchers.
The review examined 18 headline claims and checked them against primary material where available, including papers, researchers’ statements, company posts and official filings. Its categories separate independently checked claims from company-reported results, single-source accounts and claims that were disputed or contradicted. The review does not make every item on the 50-point list independently verified.
One separate measure in Thompson’s work places his “conservative countdown to AGI” at 99%, with the remaining step tied to a robot completing multi-step fine assembly at average-human reliability. That is Thompson’s own estimate and chosen threshold, not a measured probability or an agreed field-wide milestone.
“documented material progress toward an indicator, including funded programs, research infrastructure, formal scientific missions, and enabling work”
— Alan D. Thompson, describing his tracker’s partial category
What the Scores Cannot Establish
The checklist is a researcher’s framework, and the source material does not show that its categories have been adopted as a common scientific measure. The review’s 18-claim sample also does not settle the status of all 50 indicators. For some claims, independent replication or fuller documentation is lacking; others remain disputed or rely on a single source.
It is also unclear how much capability demonstrated on selected research problems carries over to less structured tasks, sustained autonomous work or ordinary workplaces. The source reports no household robots operating at scale, no verifiably billion-dollar AI-run company and no visible break in the U.S. unemployment rate tied to AI. Those observations do not prove that AI has had no economic effects; they mark limits on the evidence cited for broad real-world transformation.
Several specific claims in the source material are presented with caveats about coaching, human involvement and problem formulation. The available account does not provide enough detail to independently resolve every claim or determine whether the systems could reproduce the results reliably without assistance.
Evidence to Watch From Here
The next useful milestones are independent replication of reported research results, clearer reporting of human assistance, and tests of whether models can perform long, varied tasks reliably outside demonstrations. Robotics provides a distinct test: the checklist’s remaining countdown milestone depends on a system assembling furniture through multiple steps at average-human reliability.
Readers can also track changes to Thompson’s public grid, while treating new partial marks as progress under his stated rules rather than proof that a system has reached AGI. Until researchers and labs settle on comparable definitions and stronger evidence standards, claims about how close AI is to AGI will remain dependent on which benchmark and threshold each source chooses.
Key Questions
How many signs does Thompson’s checklist mark as achieved?
The tracker, last updated October 26, 2026, marks 2 of 50 indicators achieved, 33 partially achieved and 15 not achieved.
Does the checklist show that AI has reached AGI?
No. It records Thompson’s assessment under his definitions. The source review found varying levels of support among 18 selected claims, and the checklist is not presented as a shared industry measurement.
What does “partially achieved” mean on the tracker?
Thompson says partial status can include funded programmes, research infrastructure, formal missions and enabling work. It does not necessarily mean a capability is close to completion.
Why do definitions of AGI and superintelligence matter?
Organizations use different thresholds. Thompson’s ASI definition centers on expert-human performance across fields, while other frameworks distinguish expert performance from systems that exceed human organizations or outperform people across economically valuable work. The label depends on the definition being applied.
Source: ThorstenMeyerAI.com
