🔍 Read the full analysis: The Superintelligence Checklist: 50 Signs And AI’s Distance To AGI on ThorstenMeyerAI.com
Get hardware and tech essentials 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 of researcher Alan D. Thompson’s 50-sign artificial superintelligence tracker found that its October 26 status grid marks two indicators achieved, 33 partially achieved and 15 not achieved. The review checked 18 prominent claims: seven were verified, five were vendor claims, three relied on a single source and three were disputed or contradicted. The findings point to real AI progress in mathematics and research, but do not establish that superintelligence has arrived.
A review published October 26 examined 18 prominent claims in researcher Alan D. Thompson’s 50-sign artificial superintelligence checklist, finding seven verified, five supported only by company claims, three based on a single source and three disputed or contradicted. Thompson’s tracker marks two indicators achieved, 33 partially achieved and 15 not achieved; those counts describe his framework, not an independently established measure of how close AI is to superintelligence.
The review says the strongest evidence appears in areas where results can be checked, including mathematical proofs, a validated physics calculation and regulatory filings. Among the examples it classifies as verified are AI contributions to problems associated with mathematicians Terence Tao and Don Knuth, as well as an OpenAI model’s reported disproof of a long-standing conjecture. The review’s descriptions also include important qualifications: Tao’s example involved feedback and tools, while Knuth said he coached Claude’s exploration and wrote the proof of the construction it found.
Other prominent entries are less firmly established. The review labels OpenAI’s claim that GPT-6 Astra produced ten results on decade-old problems for about $2,000 a vendor claim, even though Lean certificates were made public; it says human pre-approval complicates what the result demonstrates. Overall, the 18-claim sample was classified as seven verified, five vendor claims, three single-source claims and three disputed or contradicted. Those findings apply to the selected claims, not necessarily every item on the list.
Thompson’s tracker uses “partially achieved” broadly. In an August correction note, he defined the category to include funded programs, research infrastructure, formal scientific missions and enabling work. The review cautions that orange status therefore does not mean an indicator is nearly complete. For example, two exoplanet candidates reportedly identified using Claude Code are listed against an indicator about colonizing a planet beyond Earth; that classification does not establish that 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
Evidence Falls Short of an ASI Verdict
The checklist brings together developments that can otherwise be difficult to compare, but its status labels depend on a definition of superintelligence that differs from those used by some AI labs. That matters because the same system may count as expert-level in one field without meeting a broader standard for general or superhuman capability. Readers should treat the tracker as a researcher’s framework for monitoring developments, rather than a definitive test of whether ASI has arrived.
The review also highlights a gap between progress on specialized tasks and claims about systems operating autonomously across the economy or physical world. It reports no household robots deployed at scale, no verifiably billion-dollar AI-run company and no visible break in the U.S. unemployment rate among the indicators it considered. Those observations do not settle what future systems can do, but they show why mathematical achievements and research assistance alone cannot establish broad social or economic transformation.
Products Worth Considering
LG 24G414B-B 24-Inch UltraGear Full HD IPS Gaming Monitor
Apple MacBook Pro 2023 14 with M3 Pro Chip, 18 GB RAM, 2 TB SSD, Space Black
Apple MacBook Pro 14‑inch (2021) with M1 Pro, 16 GB RAM, 512 GB SSD – Space Gray
As an affiliate, we earn on qualifying purchases.
Different Definitions, Different Thresholds
Thompson defines AGI as a machine performing at the level of an average, or median, human and ASI as a machine performing at an expert-human level in practically any field. Google DeepMind’s 2023 “Levels of AGI” framework distinguishes levels including “Competent,” “Expert,” “Virtuoso” and “Superhuman.” Under that comparison, the review says Thompson’s ASI definition roughly aligns with DeepMind’s “Expert” level, not its “Superhuman” level.
DeepMind’s June 2026 report describes superintelligence as a system more capable than large organisations of humans. OpenAI’s Charter, meanwhile, defines AGI around highly autonomous systems that outperform humans at most economically valuable work. These distinctions affect how labels such as “early ASI” should be read. Thompson has described Claude Mythos Preview and GPT-6 Astra as “early ASI,” but that is his assessment under his own definitions, not a consensus finding across the field.
Thompson also publishes a separate “conservative countdown to AGI,” which the source says stood at 99% on October 26. He identifies the remaining task as a robot completing multi-step fine assembly, such as building IKEA furniture, with average-human reliability. This is a separate tracker and should not be confused with the 50-indicator ASI grid.
“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
What the Checklist Cannot Establish
The review is a source check of 18 selected claims, not a full independent audit of all 50 indicators. Its findings do not establish that every tracker entry has been checked, nor do the supplied details identify all 18 claims and sources. Some evidence is explicitly company-reported, some rests on a single source, and some is disputed or contradicted.
There is also no shared definition in the material for when a system should be called AGI or ASI. Thompson’s status categories combine different kinds of progress, including enabling work, while lab definitions emphasize other thresholds such as autonomy, performance across fields or capability beyond human organisations. The source material does not provide an independently validated method for converting the tracker’s counts into a measure of distance to AGI or ASI.
Claims that named models are “early ASI” remain Thompson’s characterization. The evidence described in the review includes substantial but bounded contributions, such as solving or formalizing mathematical results, and does not by itself show reliable performance across practically any field. The scale, repeatability and real-world autonomy of the reported capabilities remain unclear from the material provided.
Further Claims Need Independent Checks
The immediate next step is further scrutiny of the individual claims: whether researchers outside the companies can reproduce reported results, what human guidance or pre-approval was involved, and whether systems can perform comparable work reliably across tasks. Public certificates and papers can help with verification, but the review indicates that independent reproduction and clear accounting of human input remain important tests.
Thompson’s tracker is dated October 26, 2026, and the source material does not give a later update or a scheduled next milestone. Readers can track changes to its status grid and compare them with evolving lab definitions, while keeping partial progress distinct from completed capability. The evidence reviewed here supports continued advances in specific AI tasks; it does not settle whether or when AGI or ASI will be achieved.
Key Questions
How many indicators does Thompson’s tracker mark as achieved?
Its status grid, last updated October 26, 2026, marks two of 50 indicators achieved, 33 partially achieved and 15 not achieved. These are the tracker’s classifications, not an independently validated measurement of AI capability.
What did the review find in its 18-claim sample?
The review classified seven claims as verified, five as vendor claims, three as based on a single source and three as disputed or contradicted. The figures describe the 18 claims selected for review, not the full checklist.
Does the review conclude that superintelligence has arrived?
No. It reports evidence of progress in areas such as mathematics, but says that “early ASI” is Thompson’s characterization and that definitions of AGI and ASI vary. The reviewed evidence does not establish broad, autonomous superintelligence.
Why is the tracker’s partial status not the same as being almost done?
Thompson says partial status can include funded programs, research infrastructure, formal missions and other enabling work. The category can therefore record progress without showing that an indicator’s end goal is close to completion.
Source: ThorstenMeyerAI.com
