🔍 Read the full analysis: How To Make Sense Of 50 Superintelligence Signs In AI on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A review of 18 prominent claims in researcher Alan D. Thompson’s 50-sign AI superintelligence tracker found seven verified, five based on vendor claims, three tied to a single source and three disputed or contradicted. The tracker lists two signs as achieved, 33 as partial and 15 as not achieved, but its definition of superintelligence is less demanding than some major AI labs’ definitions.
A review of 18 claims in researcher Alan D. Thompson’s 50-sign AI superintelligence tracker found seven verified, five supported by vendor claims, three resting on a single source and three disputed or contradicted. Thompson’s tracker, last updated October 26, labels two of its 50 indicators achieved, 33 partially achieved and 15 not achieved, but its definition of superintelligence is not the same as the one used by Google DeepMind.
The review, published by ThorstenMeyerAI.com, traced selected claims to research papers, company posts, public statements from researchers and regulatory documents. The verified examples cluster in areas where evidence can be checked, including mathematical results, a physics calculation and regulatory filings. The review says some mathematical contributions attributed to AI have been recognized by mathematicians, while stressing the limits of each result: systems received coaching, some work involved existing literature, and problem formulation mattered.
Among the examples, mathematician Terence Tao described GPT-5.2’s work on an Erdős problem as “more or less autonomously,” but said it followed feedback and involved tools; the problem had also been misformulated, and similar results existed. Donald Knuth documented Claude’s construction for an open problem after coached exploration, and wrote the proof himself. A separate OpenAI claim that a model disproved a unit-distance conjecture was reported as checked by external mathematicians.
The review distinguishes those cases from company-reported performance that has not been independently reproduced. It describes claims about GPT-6 Astra producing ten results on decade-old problems for about $2,000 as a vendor claim, while noting that Lean certificates were made public and humans prepared them. Its broader assessment is that progress is real but uneven; a tracker’s orange “partial” status should not be read as meaning a system is nearly superintelligent.
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
Evidence Behind the Superintelligence Claims
The tracker brings together developments that can otherwise be hard to compare, from mathematical proofs to robotics and scientific research. The review’s breakdown offers readers a way to distinguish results checked by outside experts from company announcements and isolated reports. That distinction matters because a striking demonstration does not by itself establish that a system can perform reliably across fields or act independently.
Definitions also change the headline. Thompson defines artificial superintelligence as a machine performing at expert-human level in practically any field. Google DeepMind’s 2023 framework calls expert-level performance “Expert AGI”; its June 2026 report describes superintelligence as exceeding the capabilities of large organisations of humans. A system may meet a narrower, domain-specific threshold without meeting that broader standard.
The review also challenges the impression that progress in models has already brought widespread economic or physical-world effects. It reports no household robots at scale, no verifiable AI-run company worth a billion dollars and no visible break in the US unemployment rate. Those observations do not settle what may come next, but they show why technical achievements and social or economic transformation should be assessed separately.
Popular right now
Breville Bambino Espresso Machine, Brushed Stainless Steel
De'Longhi Stilosa Compact Manual Espresso Machine with Milk Frother
Bose QuietComfort Headphones (2nd Gen), Wireless Noise-Cancelling Over-Ear Headphones, Eucalyptus Green
As an affiliate, we earn on qualifying purchases.
How the Tracker Defines Progress
Thompson, an Australian AI consultant who publishes The Memo newsletter, maintains a public checklist of 50 indicators, ranging from recursive code improvement to much more speculative milestones. His October video, “ASI: A better world – 2026,” presents a montage of developments from January through October. In the tracker, he describes Claude Mythos Preview and GPT-6 Astra as “early ASI.” That is his characterization, not a shared industry finding.
The status categories need care. In an August correction note, Thompson said “partially achieved” includes documented progress such as funded programmes, research infrastructure, formal scientific missions and enabling work. The source review points to two exoplanet candidates found with Claude Code as an example of how an item can be marked partial even when the milestone named by the indicator—colonising another planet—has not occurred. The tracker is best read as a list of developments to monitor, not a standardised scientific measurement.
Other organisations use different thresholds. Google DeepMind’s 2023 “Levels of AGI” framework distinguishes competent, expert, virtuoso and superhuman performance; its later report sets a higher bar for superintelligence. OpenAI’s charter, meanwhile, defines AGI around highly autonomous systems that outperform people at most economically valuable work. As the review notes, autonomy and breadth matter alongside performance on a particular task.
“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 on his tracker
Where the Evidence Falls Short
The review covers 18 selected claims, not all 50 indicators, so its findings do not amount to a full independent audit of the tracker. The supplied material does not give the names and supporting details for every claim in the 18-item review, and several verdicts rely on one source or on figures published by the companies being assessed. Publicly available certificates or posts can help scrutiny, but do not necessarily establish independent reproduction of a result.
It is also unclear how consistently the tracker applies its broad categories across very different milestones. A funded programme, a mathematical contribution and a functioning robot are not equivalent kinds of evidence. The meaning of “early ASI” remains a matter of Thompson’s definition and judgment; the source material does not show a consensus among researchers that the models he names meet a common superintelligence threshold.
Milestones Still Awaiting Evidence
The tracker will remain a running list as researchers, companies and officials report new results. Readers can watch for independent replication of model-generated proofs, clearer disclosures around company benchmarks, and evidence that AI systems can carry out complex tasks reliably outside controlled demonstrations. Those developments would strengthen or weaken specific indicators without automatically settling the broader question of whether a system qualifies as superintelligent.
Thompson’s separate conservative AGI countdown stands at 99%, according to the source review. He says the final step depends on a robot completing multi-step fine assembly, such as putting together flat-pack furniture, at average-human reliability. The review says no household robots are operating at scale. Whether that milestone is reached—and how it relates to expert-level performance across fields—remains open.
Key Questions
What does the 50-sign tracker report?
As of October 26, 2026, it lists two indicators as achieved, 33 as partially achieved and 15 as not achieved. Its author, Alan D. Thompson, says partial status can include enabling work and funded programmes, not just near-completion.
Did the review confirm that AI is superintelligent?
No. It assessed evidence behind selected claims and found mixed results. Whether the systems count as superintelligent depends partly on the definition used, and the review does not establish a consensus that they meet a shared threshold.
Which claims had the strongest support?
The review says the clearest verified examples were in areas where results could be checked, including mathematics. It describes contributions discussed by Terence Tao and Donald Knuth, while retaining caveats about coaching, tools, literature and who completed the proof.
What does “partially achieved” mean in the tracker?
Thompson says it includes documented progress such as research infrastructure, funded programmes, formal scientific missions and enabling work. It does not necessarily mean the indicator’s stated milestone is close to being achieved.
What remains unproven about AI’s broader impact?
The source review reports no household robots at scale, no verifiable billion-dollar AI-run company and no visible break in the US unemployment rate. Those observations describe the evidence reviewed as of October 26, 2026; they do not predict what will happen later.
Source: ThorstenMeyerAI.com
