🔍 Read the full analysis: A Reality Check On AI: 50 Signs Of Superintelligence And The AGI Question on ThorstenMeyerAI.com
Get smart everyday buys delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A review published Oct. 26 checked 18 prominent claims from researcher Alan D. Thompson’s 50-indicator superintelligence tracker. It found evidence for notable AI contributions in mathematics and science, but the checklist uses a broad definition of superintelligence and does not establish that artificial superintelligence has arrived.
A review published Oct. 26 of 18 prominent claims in AI consultant Alan D. Thompson’s 50-indicator tracker for artificial superintelligence found a mix of independently supported results, company-reported claims and disputed evidence. The review says the tracker records two indicators as achieved, 33 as partially achieved and 15 as not achieved—but its definition of superintelligence is broader than definitions used by some AI research organizations, making the status labels a poor measure of whether ASI has arrived.
Thompson, an Australian AI consultant who publishes The Memo newsletter, describes artificial general intelligence as a machine performing at the level of an average human. He defines artificial superintelligence as a machine performing at an expert-human level in practically any field. In his tracker, he characterizes Claude Mythos Preview and GPT-6 Astra as “early ASI.” Those characterizations are Thompson’s assessment, not an independently established consensus.
The review checked 18 selected claims against primary material, including research papers, lab statements, researchers’ notes and official filings. It classified seven as verified, five as vendor claims, three as based on a single source and three as disputed or contradicted. The review says supported examples tend to involve work that can be checked, such as mathematical proofs, a physics calculation and regulatory filings. A verified result, however, does not by itself show that a system can perform reliably across fields or act autonomously.
Among the mathematical examples, Terence Tao described an AI contribution to an Erdős problem as made “more or less autonomously,” while noting that it involved feedback, specific tools and a problem that had been misformulated, with similar results already existing. Don Knuth reported that Claude found a construction for an open problem after coached exploration; Knuth wrote the proof. Another claim—that an OpenAI model disproved a unit-distance conjecture—was supported by work involving external mathematicians, according to the review. These examples point to meaningful assistance, but their qualifications matter.
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 Label Changes the Story
The central issue is not whether AI systems have produced useful work. The cited examples include substantive contributions to mathematics, and the review identifies other claims tied to evidence that can be examined. The issue is what those achievements establish. Results in particular tasks do not, on their own, show expert performance across practically every field, dependable real-world autonomy or the ability to operate beyond human organizations.
Definitions shape how readers interpret claims about progress. Google DeepMind’s 2023 “Levels of AGI” framework distinguishes competent, expert, virtuoso and superhuman performance. Under that scale, Thompson’s expert-human threshold is closer to Expert AGI than to superhuman capability. DeepMind’s June 2026 report describes superintelligence as more capable than large organizations of humans. OpenAI’s charter, meanwhile, defines AGI around highly autonomous systems that outperform humans at most economically valuable work. These are different benchmarks, not interchangeable labels.
That distinction matters for public debate and decisions made by companies, policymakers and researchers. A checklist that counts funded programs or enabling research as partial progress may be useful for tracking activity, but it should not be mistaken for a direct measurement of capability. The review’s findings support a more limited conclusion: some AI abilities are advancing, while the evidence presented does not establish that ASI has arrived.
Products Worth Considering

Apple Mac mini Desktop Computer with M6 Chip (2026)

FLEXISPOT EN1 Electric Standing Desk, 48 x 24 in, Black

Canon SELPHY CP1500 Wireless Compact Photo Printer Bundle with RP-108 Ink and Paper Set, Black
As an affiliate, we earn on qualifying purchases.
How the 50-Sign Tracker Works
Thompson’s public checklist ranges from recursive code self-optimization to far more speculative milestones. Its October video, “ASI: A better world – 2026,” presents dozens of tracker entries from January through October. The status grid cited in the review was last updated on Oct. 26, with two indicators marked achieved, 33 partly achieved and 15 not achieved.
Thompson explained in an August correction note that “partially achieved” includes documented progress such as funded programs, research infrastructure, formal scientific missions and enabling work. That category does not mean a system is close to satisfying the end goal. The review gives an example: two exoplanet candidates reportedly found with Claude Code were placed under an indicator for the first planet other than Earth being colonized. Such a classification shows how broad the tracker’s partial-progress category can be.
Thompson also publishes a separate “conservative countdown to AGI,” which the source material says stood at 99%. He says the final percentage depends on a robot completing multi-step fine assembly, such as building IKEA furniture, with average-human reliability. The countdown is another personal tracking framework, not a validated measurement accepted across the field. No household robots operating at scale or AI-run company verifiably worth $1 billion were identified in the review, and it found no visible break in the U.S. unemployment rate tied to AI.
“A machine that performs at the level of an expert human in practically any field.”
— Alan D. Thompson, describing his definition of ASI
What the Evidence Cannot Settle
The review’s sample is limited to 18 selected claims from a 50-indicator tracker, so its verdicts do not amount to a full independent audit of every entry. The source material does not provide the complete evidence and verdict for every claim in the montage. It also does not establish a shared definition of ASI that could settle the disagreement over whether current systems qualify.
Several claims remain dependent on companies’ own reports, single sources or contested interpretations. Anthropic’s internal-model score, for example, was reported as 62.8% on a test for which the company said a full replacement for its researchers would need at least 85%. Anthropic also estimated that AI had sped up its research and development, but not by a factor of two. Those figures are company assessments; the source material does not describe an independent reproduction of them. The review also notes that physical-world capabilities, employment effects and the economic value of AI-run firms remain unproven or not evident in the cited examples.
It is therefore unclear how well the tracker’s ratings predict when, or whether, systems will meet its stated goals. The evidence supports some task-specific AI achievements, but not a broad conclusion that systems now match expert humans across practically every field or surpass large human organizations.
The Next Tests for AI Claims
The next useful evidence will come from reproducible results: public methods, independent checks and clear records of how much human guidance a system received. For claims about mathematics, that includes inspectable proofs and confirmation by specialists. For claims about scientific discovery or business capability, readers will need enough detail to distinguish an AI’s contribution from human selection, correction and follow-up.
Claims of autonomy will require tests beyond one-off demonstrations. The source material identifies reliable multi-step physical work—such as fine assembly at average-human reliability—as a remaining milestone in Thompson’s separate AGI countdown. No timetable for reaching that standard is given. Future updates to the tracker and independent evaluations may clarify whether progress is broadening, but the current review does not establish an arrival date for AGI or ASI.
Key Questions
Has artificial superintelligence been confirmed?
No. The review identifies specific AI results, including mathematical contributions, but says the tracker’s “early ASI” label reflects Thompson’s definition and assessment. It does not establish that AI systems outperform large human organizations or perform at expert level across practically every field.
What does Thompson’s tracker say about its 50 indicators?
As of Oct. 26, 2026, the tracker listed two indicators as achieved, 33 as partially achieved and 15 as not achieved. Thompson’s partial category includes enabling work and funded programs, so it does not necessarily mean a system is close to meeting an indicator.
How many claims did the review verify?
It checked 18 selected claims: seven were classified as verified, five as vendor claims, three as based on a single source and three as disputed or contradicted. The review says its findings are not a full audit of all 50 indicators.
Why do definitions of AGI and ASI matter?
Different definitions set different thresholds. Thompson’s ASI definition is expert-level performance in practically any field, while Google DeepMind’s 2026 description places superintelligence above the capabilities of large human organizations. A system can meet a narrower benchmark without meeting a broader one.
Source: ThorstenMeyerAI.com
