🔍 Read the full analysis: A Guide To 50 Signs Of Superintelligence And AI’s Path To AGI And ASI 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 examined 18 claims in researcher Alan D. Thompson’s 50-indicator superintelligence tracker. It judged seven verified, five vendor claims, three based on a single source and three disputed or contradicted, while noting that the tracker marks only two of its 50 indicators achieved.
ThorstenMeyerAI.com published a review on October 26 of 18 claims drawn from AI consultant Alan D. Thompson’s list of 50 signs of artificial superintelligence. The review classified seven claims as verified, five as vendor claims, three as based on a single source, and three as disputed or contradicted, underscoring that evidence for individual AI advances does not establish that superintelligence has arrived.
Thompson’s public tracker listed two indicators as achieved, 33 as partially achieved and 15 as not achieved in its October 26 status grid. The review says the “partially achieved” category includes funded programmes, research infrastructure and other enabling work—not just systems nearing the stated goal. In an August correction note, Thompson described orange status as “documented material progress toward an indicator.”
The checked claims included AI contributions to mathematics. The review cites mathematicians Terence Tao and Don Knuth describing work involving AI tools, while preserving caveats about human feedback, coaching and who completed the proofs. It also says an OpenAI claim about disproving a longstanding conjecture was verified by external mathematicians. By contrast, claims about a set of results produced by GPT-6 Astra were classified as vendor claims; public formal certificates, the review notes, do not by themselves amount to independent confirmation of every claim.
The review also examined claims about AI improving the research and development process. It reports that Anthropic’s best internal model scored 62.8% on a test for which the company said a full substitute for its researchers would require at least 85%. Anthropic estimated its research and development was faster with AI, but “not yet by a factor of 2.” These figures are company-reported assessments, not evidence that AI can independently replace research teams.
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 Labels Change the Claim
The review matters because “AGI” and “ASI” do not have one shared threshold across researchers and labs. Thompson defines artificial general intelligence as a machine performing at the level of an average human and artificial superintelligence as a machine performing at expert-human level in practically any field. Under that definition, evidence of expert performance in a particular area can appear to support an ASI label.
Other frameworks set a higher bar. Google DeepMind’s 2023 levels scale distinguishes “Expert” performance from “Superhuman” performance, while its June 2026 report describes superintelligence as exceeding the capabilities of large human organisations. OpenAI’s charter defines AGI around highly autonomous systems that outperform humans at most economically valuable work. Definitions affect what a milestone means: a strong result in mathematics is relevant evidence of capability, but it does not by itself show broad, autonomous performance across fields.
For readers, the distinction helps separate observable results from sweeping conclusions. The review’s verdicts suggest that some technical advances have independent support, while many broader claims rely on company reporting, limited sourcing or definitions that are not comparable. The tracker can serve as a list of developments to monitor, but its status colors should not be read as a standardized measure of proximity to superintelligence.
Products Worth Considering
Bose QuietComfort Wireless Bluetooth Headphones with Active Noise Cancellation, Mic, USB‑C Charging, Deep Bass, 24‑Hour Battery – White Smok
Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)
Logitech MX Master 4 for Mac, Ergonomic Bluetooth Mouse with Haptic Feedback, Ultra-Fast Scrolling, USB-C Charging, for MacOS, iPadOS, Windo
As an affiliate, we earn on qualifying purchases.
How Thompson Defines Superintelligence
Thompson, an Australian AI consultant who publishes The Memo newsletter, maintains a checklist of 50 indicators, ranging from code self-optimization to speculative milestones. His October video, “ASI: A better world – 2026,” presents developments from January through October. The source review says he describes Claude Mythos Preview and GPT-6 Astra as “early ASI.” Those are Thompson’s characterizations, not a consensus finding among AI researchers.
Thompson also maintains a separate “conservative countdown to AGI,” which he placed at 99%. According to the source, he says the final percentage depends on a robot completing multi-step fine assembly, such as assembling flat-pack furniture, with average-human reliability. That countdown and the 50-sign tracker are his own frameworks; neither is presented in the source as an independently validated scientific measurement.
The review says it traced selected claims to primary materials such as papers, lab posts, researchers’ accounts and regulatory filings, then sought critical perspectives. Its results are therefore a check of 18 selected claims—not a full independent audit of all 50 indicators or every system mentioned in the video.
“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 about the tracker
What the Tracker Cannot Establish
The source review does not establish that ASI has arrived. It evaluates 18 selected claims, and three were based on a single source while five were attributed to vendors rather than independent replication. The tracker’s broad “partially achieved” category also combines different kinds of progress, making it difficult to tell how close any one indicator is to completion.
The source reports that there are no household robots operating at scale, no verifiably billion-dollar AI-run company, and no visible break in the US unemployment rate linked to AI. Those observations do not settle AI’s future economic or social effects. The provided material does not supply detailed methods or data for each of these assessments, and it does not show that the absence of a visible change means no change is occurring.
Some claims remain dependent on how terms such as “autonomous,” “expert” and “superintelligence” are defined. The source material also ends partway through its table of 18 claims, so the full evidence and verdict for every item are not available here. The assessment should be treated as a review’s findings, not as a definitive field-wide scorecard.
What to Watch in Future Checks
The next useful test is whether claims on the tracker gain independent replication, complete technical documentation and clearer evidence of performance beyond narrow tasks. For systems described as capable of improving AI research, relevant evidence would include reproducible measures of how much work they complete, what human oversight remains necessary and whether gains persist across different projects.
Readers can also track whether Thompson updates the status grid and explains changes in its criteria, while comparing those updates with definitions used by research labs and independent experts. The October review provides a snapshot of selected claims; the supplied material does not specify a date for a follow-up audit or a new tracker update.
Key Questions
Does the review say superintelligence has arrived?
No. The review assesses selected claims from a tracker and distinguishes reported advances from evidence that would establish superintelligence. Thompson’s “early ASI” description is his view, not a consensus conclusion.
How many signs did Thompson’s tracker mark achieved?
The October 26 status grid listed two achieved, 33 partially achieved and 15 not achieved. Thompson’s partial category includes enabling work and funded programmes, so it does not mean a capability is almost complete.
What did the review find about AI and mathematics?
It cites documented AI contributions discussed by mathematicians including Terence Tao and Don Knuth, with caveats about feedback, coaching and human contributions. The review says some mathematical claims had external verification, while other claims remained vendor-reported.
Why do definitions of AGI and ASI matter?
Different definitions set different thresholds. Thompson’s ASI definition focuses on expert-human performance across fields; Google DeepMind’s 2026 description refers to capabilities beyond large human organisations. The same system may receive different labels under different frameworks.
What evidence would strengthen claims of AI-driven research progress?
Independent replication, transparent methods and clear measures of human oversight would help. The source reports Anthropic’s internal test result and estimate of faster research, but those remain company-reported figures rather than proof that AI can replace researchers.
Source: ThorstenMeyerAI.com
